This blog has converted to XHTML, instead of HTML. Correctly served XHTML requires that all elements in the markup are well formed. Web browsers should interpret the contents with an XML parser, rather than a tag soup parser. This forces me not to make markup errors. Any XML markup error requires the browser to stop processing, and report the error. Instant feedback, very unlike HTML handled by a traditional tag soup browser.

As an aside, WordPress uses XHTML and explains how to use XHTML correctly. They also note the need for the server to send the correct MIME type. John Kilroy offers a XHTML to HTML Wordpress plugin, to give you a fighting chance.

Readers of this blog (if I have any) should note that Internet Explorer will not work. This is due to an IE bug I am no longer willing to work around. To view my web pages, use any standards based web browser. Alternate web browsers include Opera, Google Chrome, Apple Safari, Mozilla Firefox, plus numerous others. As far as I know, Internet Explorer is the only browser unable to display correctly served, totally valid XHTML.

Microsoft, this is your problem, not mine. Internet Explorer may be the 800 pound gorilla in the web access room, but IE is not my gorilla, and Microsoft's broken version of web access is not in my room.

The above was my last blog post of 2008. I repeat it to welcome any readers to the XHTML version of my blog. I am now working on converting all of my active web pages to XHTML.

Social networks like Facebook seem to be getting a rush of new members over recent months. So why not join? I find that I am not very sociable in person, so why waste my time with a social network?

On the other hand, Jean has found heaps of our old friends on such networks. She is catching up with people I have not heard from in years.

We had around 70 mm of rain yesterday and that was (mostly) overnight. Seems a reasonable amount during the wet season, which we are now really starting to see. The town is full of visitors during the New Year holiday. There were not so many during the lead up to Xmas.

I suspect that advertising the tropics in the wet season is a waste. You can not sucker people into visiting while their chances of good weather are reduced.

Black coal-fired power station fly ash is predominantly cementitious ash of around 60% SiO2 and 30% Al2O3. Brown coal fly ash, say from Victorian power stations, is highly variable. At least 13 million tonnes are produced annually by Australia.

Very little fly ash was recycled in Australia in the 1990's, but fly ash can be used in cement production, as at Gladstone. A tonne of Portland cement uses about 110 litres of fuel oil, mostly to calcine limestone, and produces around a tonne of carbon dioxide. As concrete from pure Portland cement, production of a cubic metre of concrete makes about 0.31 of a tonne of carbon dioxide. At least 16 million tonnes of concrete is produced in Australia annually. High levels of fly ash in concrete could substantially reduce the carbon dioxide from cement production. This is particularly the case if waste heat from the power plant were used as part of the energy input for making cement.

Fly ash is usually disposed of in Australia as an aqueous slurry, despite being produced dry. Dry disposal introduces dust problems, but is used in South Africa. Fly ash pools sometimes burst their dam walls, with disastrous results for surrounding countryside.

See also pages 164 and 175 of Environmental Management in the Australian Minerals and Energy Industries: Principles and Practices By David Robert Mulligan, Australian Minerals & Energy Environment Foundation Published by UNSW Press, 1997 ISBN 0868403830, 9780868403830 793 pages

Michael Durack and Ultimate Masonry Australia Limited (UMAL) used fly ash for lightweight bricks, following research at University of Southern Queensland in Toowoomba. The Smartbloc bricks were used by Abigroup on the Domain on Gregory twin tower medium-rise residential development on the fringes of the Brisbane CBD. Smartblocs also reduced construction time, by avoiding scaffold intensive rendering. Dr Obada Kayali and Mr Karl Shaw of the University of New South Wales have developed fly ash building materials called FlashBricks that can be manufactured entirely from waste fly ash brick.

Interesting article by physicist Steven Weinberg in The New York Review of Books. Without God looks at the history of conflict and accommodation between region, and the increasingly acceptance of science. In particular, after an enlightened early start, the rejection of science by Islam in the 11th Century was followed by a continuous slow decline of its influence. Weinberg looks at how Western Europe generally avoided the appeal to authority that is so incompatible with scientific enquiry. I find myself wondering how the slow rejection of the divine right of kings figured in this.

Mark Bauerlein writes on Online Literacy Is a Lesser Kind, subtitled Slow reading counterbalances Web skimming, in The Chronicle Review. Argues from Jakob Nielsen studies that people online skim rather than reading. Well, doh! Argues that this is bad. Nielsen has long said people do not read on the web. The problem is that some problems can not be solved by skimming.

For the past several decades there has been only one true superpower, the sort that had the economy to go with its armaments. Russia fell after the USA out spent it in Reagan's day. However it increasingly seems to me that the USA may no longer have the economy to support its armaments. You can not keep spending trillions on foreign wars and bank bailouts, and keep the same position in the world. The US Mint can keep running the presses, but will the world keep the almighty dollar as the world trade currency? Already trade in the Euro exceeds trade in the US dollar. I think that is a sign.

There are lots of thoughtful articles from the web community suggesting that you avoid XHTML. Spartanicus says No to XHTML, which concisely covers why using XHTML is mostly not viable. A most thorough coverage of why you should Beware of XHTML comes from David Hammond of WebDevout. Anne van Kesteren Quick Guide to XHTML says do not use XHTML, and lists traps to avoid. Ian Hickson says sending XHTML as text/html considered harmful, and rightly so. More comments on why xhtml is a bad idea. Tommy Olsson at The Autistic Cuckoo says XHTML is dead.

Dive into Mark has a thought experiment on XHTML that shows you really should not consider XHTML on any forum or blog style site that allows web hosted content editing. You also can not risk accept Google (or any other) externally generated advertising. Lachlan Hunt says XHTML is not for beginners, and rightly so.

Most people concerned with web standards suggest that using the Strict doctype is far more important than attempting to use XHTML. David Dorward says HTML is dead, long live HTML leaves nothing except HTML. The Surfin Safari folks explain html, xml and xhtml and say to use HTML. A bunch of web designers use HTML and why they changed back from XHTML.

There are a few article explaining XHTML myths, that help avoid some of the problems. This SitePoint forum post by Tommy Olsson covers differences between HTML and XHTML, with a long list of things you need to know. This is an excellent checklist of how to use XHTML correctly. Not that anyone much seems to pay any attention.

Some nice points from the discussion.

IE doesn't support XHTML. Period. What it does support is HTML, which is what you use if you serve anything as text/html.

This is important: IE does not support XHTML served as text/html. It supports HTML with certain syntax errors in it.

Actually, browsers don't have SGML parsers at all. It would be more helpful if you referred to the parsers browsers use for text/html as "tag soup parsers" rather than "SGML parsers with some bugs", because they are by their nature not SGML parsers. For starters, SGML parsers read SGML declarations and DTDs to determinate how to parse documents. Tag soup parsers don't use SGML declarations nor DTDs, even internally.

Tina Holmboe writes on XHTML – myths and reality. Well formedness rules are checked in XML, however validity is not. This article points out that some SGML DTD rules (like a p tag must only contain inline elements, or a blockquote must only contain block elements) can not be specified in XML. So, an HTML doucument is HTML only when it is valid. Whereas an XML document must be well formed, but may not be syntactically valid.

The rain is absolutely bucketing down tonight. I am glad that Jean managed to get two loads of washing done in the morning, while I was enjoying myself failing to find anything much at the local markets. Well, I did manage to get rid of two packages of surplus stuff, a cassette recorder and a few years of Analog magazines. The weather radar shows heavy rain all over the coast locally. We had only 5 mm overnight Saturday, but I suspect we will have much more tonight. And we did. It rained another 60 mm overnight.

Definition: Webalati is the collective term for pseudo computer geeks, especially those involved in writing, coding or looking after web sites. The term is often used disrespectfully. Based on literati.

Speculation that a future 17 inch MacBook Pro will use a recyclable rechargeable ZPower brand 40% longer lasting silver zinc battery in a new unibody notebook.

Speculation was wrong. Apple used a custom cell lithium polymer battery. It is not a user replaceable battery.

Pizza for lunch. I went at midday to collect a pizza (our first of the year) from Dominos for lunch. Jean is not usually the one to succumb to the lure of fast food, so for her to suggest pizza was a treat. My position is usually that of villain, spoiling the monastic dieting of our lives.

Walking with the pizza back up the twelve flight of stairs in the Whitsunday Terraces from the main street, I noticed the resort gardener and the handyman were removing lush tropical vegetation from the gutters of the Florin Terrace building. Since this is probably the first full working day since our Whitsunday Terraces body corporate budget meeting in December, I am pleased to see our suggestions had some effect. Especially as we have had a few rainy days recently. If we get real rain, the gutters will overflow anyhow, but clearing them at least reduces the still waters for mosquitos. It also reduces the chances of leaks into the rooms.

Since my Carlyle Gardens Gnome retirement resort web site now consisted of .xhtml files being correctly served as application/xhtml+xml, I decided to look at the problem of empty elements or self closing web tags within the body element. I find this a difficult subject, since most tutorials seem to get it wrong.

If we consider HTML an SGML generated markup (it is not really), the DTD permits elements that do not actually have a tag (omittag). These were specified in SGML Declaration for HTML by Daniel W Connolly in 1995. For example, you do not need to have tags for html, head, or body. Not having them is stupid, but it is not invalid. You do not believe it? Look at the DTD pointed to at the top of most web pages.

Also, HTML being SGML, some closing tags in HTML can be made optional. For example the p tag for paragraphs. This does not lead to many parsing problems. The paragraph element is only permitted to contain inline elements, so if you encounter another block element, then obviously the paragraph ended just prior to the new block element.

In addition, to allow HTML to be easier to write, shorttag is allowed. This permits unquoted attributes within a tag, something not permitted in XHTML.

XHTML, being a simplified XML markup, could not provide optional end tags in its DTD. Every tag you open must be closed, and closed in the correct order as well. However some elements in HTML never did have any sort of closing tag. For example, the break <br>, the horizontal rule <hr>, and the image tag <img> for displaying photos and graphics. As I had early misgivings about these tags, I have never been a heavy user of <br> or <hr>.

The general advice seems to be that in XHTML, you simply use a self closing tag. So an HTML break <br> becomes an XHTML self closing break <br /> The extra space before the slash is because some browsers had issues if there were no space.

So, problem solved? Not exactly. In fact, it ensures virtually no XHTML web page can ever be valid.

Okay, you have your brand new fancy XHTML web pages. Just what is an XHTML web page? Two things make something an XHTML page. The first is that the web server has to say it is an XHTML page. That means it is served as application/xhtml+xml. The second thing it needs is to contain a line like <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> the first part of which says it uses an XHTML name space. If either are missing, your web page is not XHTML.

Then you find Internet Explorer can not handle XHTML at all. So you get your server to send your pages as text/html, like many people recommend to keep Internet Explorer happy. However that means the pages are not XHTML. They are broken HTML. You can not use things like self closing tags in HTML. True, you can send the fake XHTML web page to an XHTML validator, and they can validate it as XHTML, but your server is claiming it is actually HTML. So at this point you have neither XHTML nor HTML. You are just hoping that the web browser error handling will cope with whatever it is you are sending from your web host. The 30% of web sites putting an XHTML Doctype in their web pages are basically admitting they don't know how to write a web page.

My images are marked up as <a href="images.jpg"><img src="thumb.jpg" alt="photo" height="240" width="320" /></a> at present. However I really want to get rid of self closing tags like <img>, because in XHTML you are forced to use /> to close the tag. This in turn means you run afoul of shorttag in SGML.

Using the object tag is the way to go to replace img. Complications are mostly because the version of object that Internet Explorer uses comes from a different heritage. I need not worry about object differences, because the IE XHTML bug means IE can not display my site.

Another problem is that you really should tell your browser what sort of image file you are serving. It looks to me as if a lot of people think image/jpg is what is used for photos. It is not.

The Internet Assigned Numbers Authority (IANA) maintains the MIME Media Type list. RFC2046 on Multipurpose Internet Mail Extensions lists media type image/jpeg. Despite the common .jpg file extension for photos, there is no such media type as image/jpg, so it must not be used. You must use image/jpeg.

Although Safari does not accept the incorrect media type image/jpg, both Opera and Firefox do accept image/jpg as part of their error correction (as well as the correct image/jpeg). You do not want to rely upon web browser error correction, as there are no standards for fixing errors.

The SGML NET (null end tag) shorttag means when you use XHTML things like <br /> in SGML (a web page served as text/html) then SGML rules apply. Under SGML, the <br /> element ends with the slash, not with the >. So wherever you have a shorttag, you should get a surplus > littering your text, because what you actually asked for with <br /> is <br >>. Most people do not notice this is a problem because most people use browsers that do not actually follow the rules.

Why do you get away with it on an <a tag followed by a URL with a slash in it? Because the URL is invariably quoted. If you quote all your attribute values, then any surplus slash in them will be hidden. The tags that will cause problems will be in files written as XHTML, served as HTML, and that contain self closing tags like break, horizontal rule, and img.

The W3C HTML validator used to give an error message saying NET-enabling start-tag requires SHORTTAG YES This generally means you used a shorttag in HTML.

Apple co-founder and CEO Steve Jobs reluctantly published a press release setting out the reasons for his weight loss. He confirms he has been losing weight in 2008. Neither he nor his doctors knew why, but have now found the cause. He is having treatment, but weight recovery will take some time. Apple Board of Directors back Steve Jobs.

Steve Jobs spent the holiday at home with his family, not preparing for and delivering one of his astonishing sales performances at Macworld Expo. The timing and vagueness of the Apple press release seems typical of any Apple product release, as one dumb analyst has noted. Next post, why is Steve Wozniak not thin?

I really do wish the blogosphere and the press would give stock bouncing Steve Jobs health scares a break. If there is something that affects the long term working ability of a CEO, the board are legally required to make it known to the market. If you know nothing, why speculate? If you know something, and trade in Apple shares, is that not insider trading? If you speculate online or in print, and speculate on the stock, is that not stock manipulation?

Post by Andy Ihnatko I really liked, nominally about his Touchdown, San Francisco. Part of it is because he praises the thousand buck bed at the Marriott, while noting that he is unwilling to spend that much at home. He further notes he realises he is an ass.

I like the comment on the merits of a good bed, because when we move house, we will be spending even more getting a decent bed. Maybe one of those Sleepnumber jobs, with comfort levels you can dial. Why spend so much on a bed? Because you spend a third of your life in bed, and you might as well be comfortable as uncomfortable. It is the same principle that keeps me looking at new technology, at science fiction, and at futurists. You will spend the rest of your life in the future. You might as well know how to cope with it.

The other bit is that towards the end of the post, he says he has heard wonderful things for Apple fans, but off the record. Does not say what. He even had to get permission to say he was told something. This is just the sort of leak Apple likes. Nothing at all, except sell the expectation. I will probably have to read his confounded blog over the next few days, to see if there are hints.

Adam Gopnik writes in The New Yorker on the passions of John Stuart Mill on 6 October 2008. This is a lengthy and interesting discussion and review of a new biography by Richard Reeves called John Stuart Mill: Victorian Firebrand. Alas, the ponderous prose of Mill would probably be disregarded in an age of sound bites. However, from the first time I read On Liberty, I thought Mill did more than any other individual to set the political course of the modern West towards universal citizenship and individual rights.

BBC Space and technology reporter Jason Palmer reports on Getting into space by broomstick. Age-Raymond Riise of the European Space Agency demonstrated the device at a space elevator conference in December. The demonstration is a cute idea to show transferring climbing power mechanically. Bounce around the broomstick with a sander. Arrange three brushes around it, bristles down. Friction is higher going down.

This is a party trick, and probably only intended as a party trick. The idea is fundamentally flawed. The BBC reporter should be able to clearly see this.

First, a broomstick is a rigid compressive load, so you can vibrate it up and down. A cable is a load in tension. Jerking it up and down just flaps one end of it. You might as well try pushing a length of string.

Second, even if it were a solid body, you would need to move the entire mass of the cable. This way exceeds the power needed for a capsule riding on it. Just look up estimated mass for space elevators.

Which browsers are used by most people? Well, Internet Explorer, since it comes with Windows, which comes with most computers. Pity that Internet Explorer is the most bug ridden browser around these days. IE also benefits from everyone changing user agent strings to fake being IE. Opera is probably underestimated, since it makes it easy to fake the user agent string, and because it has excellent caching.

W3 Schools browser statistics probably overestimate beginning web site developers. However W3 Schools content is suspect, so experienced web site developers probably avoid them. Both factors will modify browser statistics.

The actual accounting for which browsers visit are also unlikely to be accurate. Browser Statistics shows results from three different free tracers. Notice the figures are all over the place from each other.

Browser News Statistics takes figures from six different statistics sites. It also has a nice account of why statistics lie.

Caught the end of a Commonwealth Bank TV advertisement when I flipped back to the movie a little too early. I seemed to be saying something about (possibly free) music downloads in MP3. So I went to the Commonwealth Bank /music directory (seen at the end of the TV ad). So this is a good setup for someone advertising. A voluntary eyes on view of at least part of their web site. A potential sell through opportunity.

Right! An invalid, obsolete frameset site that showed a completely empty screen. They probably did this shit with Flash or something else that I always block (because of the number of advertisements done with Flash).

When will companies learn to make their web sites easy for viewers rather than hard? When will they start testing it with a variety of browsers and browser settings? At the very least, have some alternate content to encourage someone to continue. I left the site, and will probably never return.

Up at 4:30 this morning so I could find out whether Apple had released anything new and interesting at Macworld Expo. So I sat listening to the gentle rain as the sky gradually lightened behind the clouds.

Apple senior vice president of worldwide product marketing, Phil Schiller, presented the keynote talk at the last Macworld Expo Apple will attend. He started with I can't tell you how much I appreciate you all showing up. That got a laugh.

Phil Schiller started in on Apple Stores, mentioning Beijing and Sydney. I have to say I am enormously impressed with Sydney Apple Store. I schedule the Apple Store for a visit whenever I make the 2000 km trip to Sydney. Punchline, 3.4 million visitors to Apple Store each week, 100 Macworld Expos, each week. That is why Apple does not need Macworld Expo. Needing Macworld Expo is like needing a floppy drive.

Neither the iMac nor the Mac mini got an update, despite rumours and potential hardware being available for a update. I can see not rushing the iMac to a mobile Core2 quad that has only just been announced at a good price but at a low speed. Having a new model with a lower speed chip will just confuse people. Also, decent support probably arrives with Snow Leopard.

The Mac mini is looking increasingly outdated. Now that the iMac and most of the Macintosh notebooks are only available with a hated glossy display, the Mac mini was looking to me as the only way I could get a replacement for my iMac G5 ALS. I would prefer a large iMac, but I will not buy a glossy display. That sort of left a Mac mini, despite not being all that happy about it having insufficient disk space.

Keynote presenter Phil Schiller noted Apple had their best year in 2008, with Macintosh sales of 9.7 million. At the end of the keynote presentation Tony Bennett comes on stage to sing The Best is Yet To Come. Talk about Apple being on message.

Apple iLife 09 is a welcome update, if there are improvements. I don't mind paying an Apple tax for a family pack every few years for worthwhile increases in ability. I guess I will buy in March, when the local Post Office address changes stabilise.

iPhoto 09 has Faces, automatic face identification across multiple photos. If it is consistently correct, that would be handy. Events worked better than I expected, so this has potential. Places, tagging photos by location for GPS Geotagging. You can use Google Maps or type the location in, for cameras that do not yet have GPS built in. Maybe that will help kick the camera makers along a little towards finally putting GPS in their cameras. I have wanted that GPS support for a long time.

The concepts of Events, Faces and Places are an example of Domain language. They change the way Apple think about iPhoto. As a result, they change the way users think about photos, and particularly about metadata.

iPhoto 09 has Facebook and Flickr support. I use neither. However it notes a photo sent to Facebook can have face identification returned to iPhone. I do have a bunch of photos of folks whose names I can not recall. Incorporate maps into travel books. It is sounding a nice upgrade.

iMovie 09 gets an updated editor, an expanded timeline, and custom control of drag and drop. Project Library for film snippets is updated. Video stabilising available. Randy Ubillos was the engineer responsible for iMovie 08, and he does the demo. Nice to see the engineer credited with a neat product.

GarageBand 09 gets a learn to play feature, with HD videos. Nine guitar and piano lessons free, you can buy others via Garage Band at US$4.99.

Nothing noted about the changes to iDVD or iWeb. I don't use either. DVDs are so 20th Century. iWeb doesn't seem to do fluid designs (it does however do valid pages).

About time iTunes goes DRM free is all I can say. I have never bought any DRM music from iTunes (and never would), so it makes no difference to me. It does make a difference to the bottom line of music producers. Now I might buy something from them. The iTunes music store is really convenient. Several times some friend mentioned a piece of music (mostly nostalgic stuff), and if it were DRM free iTunes Plus, I bought it on the spot. Phil Schiller says 6 billion songs sold, 10 million songs in store, 75 million accounts. Impressive, for a standing start from 2003.

I am a little surprised at the .69, .99 and $1.29 price points to be introduced on 1 April. I did not think Apple needed to run different prices. The music producers would have eventually given in on that. However I do not object. It is easy to read a price in the store, and decide how much you are willing to pay.

3G downloads to the iPhone will be welcomed by my younger friends, or so I gather. Of course, since their parents won't let them have an iPhone, and they don't have a credit card, I am not sure how much of a sales difference it will make. You could already use WiFi.

iWork 09 update. Plus public beta of iwork.com for collaborative viewing of iWork documents online via iWork. So, not exactly viewing in a browser? I guess I will need to see a demonstration to understand that.

Keynote gets object transitions, and more text transitions. Magic move is tweening using core annimation. iPhone remote available.

Pages gets full screen. Not sure how this is different to just filling the screen with it, like I always do. Outlining and listing, for larger projects. 40 new templates. Mail merging with Numbers and Address Book. Great, except our post office just closed. Does anyone actually still use physical mail? MathType. My mathematician friends will like that, if it is any good. I gather there is also EndNote support, for academics.

Numbers gets MathType. Drag and drop formulas, over 250 function, new function view, new charts, error bars (at last), trend lines. Lots of new templates (good idea).

Newly revised 17 inch MacBook Pro notebook computer with the unibody aluminium shell is released, as expected. Glossy 1920 x 1200 LED backlit display, driven by Nvidia GeForce 9400M and Nvidia GeForce 9600MT video. 2.66 GHz Core2 Duo, up to 8 GB 1066 MHz DDR3 RAM, 4 GB standard. 320 GB Serial ATA hard drive at 5400 rpm standard, optional 256 GB SSD. Slot loaded 8X DVD R/W drive. New glass multitouch trackpad. Backlit keyboard, with magnetic latch.

iSight camera, WiFi 802.11n, Bluetooth 2.1 EDR, Firewire 800, 3 USB, ExpressCard 34 slot, Gigabit Ethernet, mini DisplayPort. Audio 3.5 mm line in and line out, with optical digital as well as analog. MagSafe power connector on the 85 Watt power brick, of course.

I have a 15 inch 1.25 GHz Powerbook G4, from early 2004, and would like an excuse to upgrade. The problem is size. I choose the 15 inch because it was an excellent compromise between desktop replacement, and travel notebook. However the 15 inch MacBook Pro has a glossy display, and I will not buy a glossy display. That means the 17 inch matte display MacBook Pro, which is much more a desktop replacement.

The new 17 inch MacBook Pro weighs 2.99 kilogram, and is just 2.5 cm thick. The 17 inch is 39.3 cm wide (compared to 36.4 cm for the 15 inch MacBook Pro). It is 26.7 cm deep (compared to 24.9 cm for the 15 inch). However my five year old Powerbook weighs 2.5 kg, is 34.8 cm wide, and 24.3 cm deep. In short, the ancient PowerBook takes less space in a bag than the new MacBook Pro 15 inch. The only dimension in which the older machine is not smaller is thickness, of 2.8 cm. Not to sound critical, but 3 mm thickness is rarely the limiting factor in carrying a computer.

Battery is 95 WHr lithium polymer with custom shaped cells. Claimed to have three times standard life, up to five years, adaptive charging on a cell by cell basis, retaining 80% of charge up to 1000 charges. The battery is not removable. Apple presented a whole video about making batteries. Claims 7 hours use with discrete video card, 8 hours on integrated graphics.

Display worries me. Viewing angles given as 140 degree height, 120 degree width. I thought the previous model claimed something like 170 degrees? Does this indicate major cost shaving on the display? 700:1 contrast ration. There is a US$50 anti glare option. However if this is just a coating over a glossy piece of glass, while the glossy LCD is behind it, I am not at all sure that is much use. I will have to wait until I can see it in a store.

Compare this Apple photo of 2009 MacBook Pro 17 inch glossy display with this Apple photo of 2009 MacBook Pro 17 inch matte display. Notice the glossy display has glass over a black background to the edge of the case. The matte display is showing what appears to be an aluminium surround. Note that it does not appear to have the glass cover over the LCD display. I am guessing it is using something like the plastic feeling display of my old Powerbook G4.

Cover glass for LCD displays needs special treatment. It used to use barium for the expansion phase. To eliminate oxygen bubbles, either antimony or arsenic are used as fining during the melt. Arsenic is favoured because it is one of the highest temperature fining agents available. Alkaline free glass is needed for LCD substrates, because alkaline ions contaminate liquid crystal structures. The arsenic free glasses are relatively new.

My question is what green credentials do you give up in demanding a matte display? I am not buying a glossy display, no matter what. However I would like to know whether a protective plastic LCD cover happens to contain arsenic or other toxic material.

In the great tradition of the 1945 Ern Malley affair that bit Max Harris, Quadrant editor Keith Winschuttle has been tricked into publishing a scientific hoax. The hoax article was Scare Campaigns and Science Reporting by Shanon Gould. Windschuttle critic Crikey is happy about the hoax, but say they didn't know about it beforehand.

The silliest watches in the world? Adjustable length telescoping minute hands, revolving hour hands, powered by compressed air turbines. Urwerk UR-202 Or for that retro kick, Cathode Corner Nixie Tube Watch shows the time on 1969 vintage nixie tubes. You also can get the PIC CPU firmware for the watch. Really cute.

As usual, the Australian Bureau of Meteorology has an excellent historical summary of floods in Queensland, by decade. It commences pre 1860, and continues until the past year or two. I was using it to check flood patterns where we live, and also at Townsville. Jean suggested I note it in my blog. There is also an excellent set of Detailed Reports on Notable Queensland Floods.

The W3C have some tests for HTML 4, based on tests contributed by Microsoft. I was particularly interested in tests for the Object element. Browser support does not seem startlingly good.

On Tantek Celik wrote on a solution to the Human vs. ISO8601 dates problem in MicroFormats such as hCalendar. His first thought was to use the object element, with the ISO8601 date as the data, and let browsers fall back to the human readable data as the content. As I have done with the date above (which looks fine in Firefox, Opera and Safari). Except when it fails in Safari.

However back then the Apple Safari browser did not work correctly even with simple objects. So instead Tantek used the abbr element, with the human readable date as the abbreviation, and the ISO date in a title attribute. The problem is that this works for visual user agents (the ISO date may show as a tooltip under the cursor). The abbr tends to fall over with user agents for sight impaired readers, that may speak the content of title attributes to assist in explaining the meaning of a particular abbreviation.

This use of difficult to read numeric data in a title attribute in turn led the BBC to drop the use of hCalendar. This despite the BBC supporting Semantic structure in HTML originally.

I am increasingly coming to the opinion that you should never work around browser bugs. Sooner or later the browser bug may be fixed. Meanwhile, the work around is just as likely to produce its own set of problems.

Boston icon, the historic Harvard Square news stand Out of Town News, may close. It sold, well, expensive and somewhat dated out of town newspapers. The former owner would like the lease back, perhaps regretting the 1994 sale. Manifest Magazine says print on demand magazines and newspapers are the answer.

Mark Fletcher did an article in 2005, Restructuring the representation of newsagents in Australia, available on United Newsagents in PDF, although it is marked confidential. Newsagent representation sounds somewhat chaotic and suffering from neglect. Mark Fletcher runs Australian Newsagency Blog. Skip to 2008, and the Australian Newsagents Federation has a new CEO. The newsagency system in Australia has served publishers and consumers very well and has played a very big part in making magazine sales in Australia among the highest per capita in the world. Very true.

Meanwhile, NewsAgency Blog has complaints about newsagents giving up home delivery of newspapers, due to excessive costs. Complaints Joe Hockey misrepresented competition policy for newsagents by the Liberals. Mind you, Hockey is a politician, and his mouth was open, so did you expect anything except spin?

Some sort of little local cloud over Airlie Beach has added 40 mm to my rain gauge on the balcony at the Whitsunday Terraces by 10 a.m. The vast areas to the north of us are flooded, with roads cut all over the north. However I thought we were getting away with just quick little dust settling showers. Meanwhile, Coconut Grove is starting to flood (only very shallow). Plus it is high tide. I wonder how the main street of Airlie Beach is going?

The Christian Science Monitor stops printing a daily newspaper, and goes all digital online. They will have a weekly printed magazine, and hope to move 80% of their subscribers over.

Clay Shirky reports in Boing Boing on the newspaper industry and the arrival of the glaciers. That is, that everyone could see the internet was a threat to print long ago. However many newspapers hoped the glacier would somehow move aside at the last minute.

Clay Shirky recalls his 1995 article Help, The Price of Information Has Fallen and It Can't Get Up. It basically disputes that users will eventually have to pay for content. Simple pricing says cost plus profit. But price rises occur when demand exceeds supply. However digital distribution means supply is infinite. Paper distribution is costly, real costly. Competitors have the same problem. Not so digital. Anyone with a ten dollar web site can compete.

The internet has a global reach, and is not limited by geography. It has an unlimited capacity to perfectly copy any digital material. The margin cost of making another copy is approximately zero. Basically, the printing press is stuffed. Any business that relies upon restricting access to a printing press is stuffed. Newspapers, classified advertising, book publishers, telephone directories, magazines, you name it. Totally stuffed.

In a 2001 article called The Real Wireless Innovators, Clay Skirky points out that wireless phones are not really useful. In particular, the walled garden approach of WAP just was not something you had to have. In contrast, the web was of use even back when it was only text. I used Archie and Gopher and the rest, and the first Web pages beat them every which way. Everyone I knew was installing web servers and writing web pages (badly) within hours of understanding what was involved. In the phone business, RIM understood they needed a killer application. The biggest killer application on the early Internet was Mail, and that is what the RIM Blackberry got right.

Skirky points to NTT DoCoMo, which started using Compact HTML. Anyone could write content for their i-mode phones, and they took off in Japan. The wireless phone companies just did not get it. They want walled gardens, and that went out with AOL. The real breakout wireless services are being launched not by the telcos but by innovative device and service companies who think of wireless as a feature, not as an end in itself. That was an invitation for companies like Apple to provide a device so attractive that it works (see iPod Touch) even if you can not make phone calls, even if you can not connect to the cell phone system. Wireless is a feature (iPhone), not an end in itself.

Australia would probably have a steady state (static) economy, except for two factors. Immigration, and mining. One major problem raised by a non growing population is an increase in the average age of the population - more retirees relative to workers. To cope, either higher taxes, older retirement age, or reduced retirement pensions. Japan is already going down the drain with stagflation due to their ageing population. Their culture can not cope with massive foreign immigration. Australian population increase (before deaths) is 350,000 a year, on a 21 million base, which is way high for a first world economy. The population increase over deaths is entirely immigration. Despite Costello's plea to have a third kid for the country, the Australian population would be in decline if not for immigration.

So Windows 7 has been released in beta to 2.5 million suckers. Vista was Microsoft's great white hope. Umpteen years in preparation, all manner of features dropped before release. Vista was almost universally panned.

Now Microsoft is talking to customers about what they want. Give it a break. Customers have not got a clue what they want. If it was up to customers, Ford would have delivered a faster horse pulling a fancier carriage. Apple don't give a shit what customers want. Sure, sometimes the customers jack up, and don't buy. But the people who write operating systems have a way better idea of what is possible than the customers do. The real problem is deciding what to leave out.

If Microsoft do not have better ideas than keeping it compatible with a 1982 IBM PC, and asking customers, they might as well give up. Meanwhile, for most people, the important thing is not whether they have operating system U, V, W, X (if they happen to know), it is whether they have access to the internet (which they will call The Web, because they don't know better).

It took a long time, but now the U.K. MacWorld Magazine has come out and said the glossy display is not as good as a matte display for photographic work. MacWorld used PressMatch reference print and colour calibration equipment to test a number of monitors next to the Apple LED Cinema Display.

I am not a professional photo user. However I do have eyes, and I do live in the bright light of the tropics. I am not buying a glossy display from Apple. I am not buying a glossy display in an Apple monitor. I am not buying a glossy display in an Apple notebook. I am not buying a glossy display in the replacement iMac I devoutly desire.

I do not care if a display has arsenic free glass covering it. What I care about is whether I can tolerate it. I can not tolerate a glossy display. This problem is not going to go away. It is the way my eyes work. In a glossy display, I continually notice reflections of the room around me. My MacBook Air is close to unusable because of glare (and as a result, I rarely use it). My iPhone suffers from glare, and is often impossible to read outside.

Dell or someone else can have my monitor business (I already have an older Dell 24 inch display). Also, the current Dell Ultrasharp has a wide and convenient range of USB hub facilities, a camera card reader, and at least five different input facilities, from low TV quality through VGA, DVI, and I believe now also DisplayPort.

During the Sydney to Hobart yacht race, local Whitsunday Sailing Club boat Telcoinabox Merit rescued 14 sailors from stricken Georgia, minutes before Georgia sank. They will also bring home the 2008 Performance Handicap Championship trophy. Merit is owned by local Leo Rodriguez and Ian Bishop. Merit was sponsored by Telcoinabox.

Leo Rodriguez runs local Airlie Beach ISP Internet Whitsunday, a part of Voice2Net telecom. The only time I had to deal with their business (on behalf of a third party), the level of service was above and beyond usual ISP levels. Including a personal visit with a couple of extra ADSL modems to solve a nasty Internet connection problem.

Much threat of light rain cause most of the sales tents to fold early. The wind didn't help either. At least this time I could get all the fruit and vegetables that Jean wanted at the Airlie Beach markets. Afterwards I had to carry it all up the stairs at the Whitsunday Terraces.

The old busker told me of his childhood near Boulia, and his one encounter with the Min Min light when he was around 17. That is a long time ago to think UFOs. Professor Jack Pettigrew thinks Min Min lights are Fata Morgana.

Not very happy to notice that the garbage bins at the Whitsunday Terraces were not put out this morning. They seemed very full, which made me wonder whether the bins went out the previous day. I am still not sure why it is so hard to get the bins put out every day, like the contract says.

Since we are moving, and I will not be on the Body Corporate committee after the end of the month, I guess I can stop worrying about it pretty soon.

Mohamed El-Erian of Pimco has a great example of what trust means in the market. At a drive-in fast food joint like McDonalds, you trust that paying at the cashier's window will result in collecting a meal when you get to the pick up window. Without that trust, you may have a peanut butter sandwich at home instead.

Bill Gross of Pimco continues that capitalism is inherently unstable. This he blames on the need for long life capital assets to function. However capitalism uses short term buying and selling of equity and debt to finance long term assets.

Jean disconnected her telephone before the weekend started. This left the internet connection intact, since it was only the phone unplugged. Neither of us actually make phone calls in any case, unless it is a last resort, say to someone who does not have an email connection. We are not expecting any business phone call until the working week commences on Monday. Why bother risking some annoying phone call from some sales person?

The phone should be on the Do Not Call list. However politicians and surveys and charities do not seem excluded by DoNotCall. This lack of inclusiveness vastly reduces the value of any DoNotCall list.

What I would really prefer is a naked connection, for DSL only. The concept of talking over a telephone is so 20th Century.

Jean came for a walk with me to the local newsagent, when I got the morning newspapers. Alas, on the way up the twelve flight of stairs to get back, it became apparent that her left leg was again giving problems. Jean had to lever herself up many of the steps via the handrail or by using my arm as a handrail. This walking problem is making our future move to someplace flat seem better timed than ever.

Girl, 80 kg, heart attack when only ten years old. Brittany Lindsay from Ipswich was the unfortunate victim. Brittany is one of a group of 30 overweight children now taking part in the University of Queensland and Royal Children's Hospital Eat Smart program. You have to worry about what we are doing to our eating habits.

In Australia, the two best known indicators of inflation are the Consumer Price Index (CPI), and the Gross Domestic Product Deflator, which measures inflation over the whole of the domestic economy.

The quarterly headline inflation rates in Australia in 2006 were 3%, 4%, 3.9% and 3.3%. In 2007 they were 2.4%, 2.1%, 1.9% and 3%. In 2008 they were 4.2%, 4.5% and 5% (I will not have the last quarter until late January).

CPI tracks changes in the price of a basket of goods and services that is typically bought by a metropolitan Australian households. In the USA, arithmetic weighing of items in the basket were replaced at some stage by geometric weighing. This gave lower weighing to components increasing in price, and higher weighing to components dropping in price. In Australia, changes in the weighting pattern have been made at approximately five-yearly intervals to take account of changes in household spending patterns in Australia. So if we all change from Wagyu beef to hamburger to dog food, it will take a while to show up in the CPI. The decline in living standards is not taken into account. In short, cost of living is actually more like the cost of survival. Hence any pensions or other income based on CPI adjustments are generally underestimated.

Cost of Living Index (COLI) takes into account the cost to maintain a constant standard of living. In other words, what would it cost you from year to year to live exactly the same. That is harder to find. Actually, I failed to find it.

US banks changed from utilitarian financial institutions into casinos. Part of the problem was the financial ratings agencies like Moody's and Standards and Poor turned out to be asleep, instead of evaluating risk. I guess being paid for by the people they were rating was a factor. I mean, Enron was rated at AAA four days before it failed. So Basal II wasn't worth worrying about, and the credit rating agencies should never have been government approved, especially after they started rating collateralized debt obligation. Sounds to me like a bunch of born losers.

The idea of an efficient market seems another fiction. Almost an item of religious faith. Markets could never fundamentally mis-price assets, and so regulation was never justified.

Well, markets get asset values wrong all the time. Someone will always think someone else knows less about the value, and that gives a margin. The market ends up essentially either a con game with inside traders, or at best a zero sum game. So bubbles will always happen. The money ends up with speculators who jump in early, and sell at the middle. The suckers are the momentum traders, who buy when things are going up, fail to sell until it is going down. The bubble bursts when the momentum traders run out of money to push prices higher. Not much to do with the quality of the assets. Just like buying property, you make your money when you buy cheap, not when you sell.

Making the moral imperative to be growth at all cost means economic efficiency gets priority. Leverage becomes a moral duty, automatically leading to an excess of feral financial flimflam.

What makes an XHTML document actually be XHTML?

The server must send a Content-Type HTTP header with a MIME type of application/xhtml+xml, application/xml or text/xml.

The <html> tag must have this exact attribute: xmlns="http://www.w3.org/1999/xhtml".

Nothing else matters. Not the XML declaration, not the doctype declaration, not any meta element.

What stops XHTML being used? What is wrong with Internet Explorer? IE does not accept XHTML. At all.

Suggestion for pricing model, based on time wasted attempting to make web pages work in various web browsers:

• Website for Standard browsers. Cost X
• Add looks good in IE8. Cost 2X
• Add also looks good in IE7. Cost 4X
• Add also looks good in IE6. Cost 8X
• Add also looks good in IE5 ...

... and so on to taste, with variations on the exponent.

I posted that in CIWAS newsgroup, in response to an aside on the need to earn a living from writing web pages. Several people seemed to appreciate the pricing structure.

Graphic thanks to Google Charts API (generated while online only).

WebKit has for some time included experimental CSS features along the lines proposed for HTML5. Opera and Firefox may also support some of these features. Internet Explorer does not, however the bug that prevents Internet Explorer displaying XHTML pages at all, also means I do not have to worry about more minor display problems. So I intend to experiment.

In this blog, I have been using WebKit text-shadow for some time. So far I have not used CSS text stroke (fills the interior of letters). I have been using WebKit CSS box shadow, and also CSS rounded corners, called border radius.

WebKit includes support for CSS gradients, both linear and radial. WebKit supports CSS reflections. WebKit supports CSS Alpha Masks. WebKit also supports CSS Canvas Drawing, allowing Javascript to draw on the canvas.

Best description of CSS browser support, helps me decide which browsers can handle what. Rewriting my CSS is shaping up to be a difficult task.

Multiple column layout (at least in Safari), using the CSS3 support for multiple columns. Namely the browser specific experimental CSS column-count and column-width. When the iPhone did not understand either, it provided a single column layout. However now the iPhone also renders multiple column layouts, and they are too small to read. This leads to a lot of scrolling in two dimensions. This would be reduced if I could force the iPhone to go back to a single column display.

The media type handheld is intended to cope with exactly this problem. It should let you feed a custom CSS style sheet to handheld browsers. Unfortunately, Apple decided Safari on iPhone was not a handheld browser. So that blows that idea.

Even better would be if the iPhone could handle text zooming. My web sites are all fluid designs. My web sites cope just fine with text zooming, so someone with vision problems can expand the text to any degree they need. But there is no way I can find to text zoom on an iPhone. Zooming the whole screen is not helpful, because that ends up requiring horizontal scrolling.

A potential starting point seemed to be the CSS Media Queries. I should be able to test for display width, and feed a single column layout to anything that has less than, say, 500 pixels of display available.

@media only screen and (max-width: 480px) { div.article { -webkit-column-count: 1; } }

Unfortunately this seemed too inconsistent to use at all. Safari only makes static responses to media queries, you can not test by dynamically changing the browser width. At least, in testing with Safari on a Macintosh I could not. Once or twice I got the results I wanted, but mostly I did not. Seems maybe the caching is too aggressive, and I didn't clear the cache enough. I think my testing problem relates only to the caching.

Apple once had a really great web page on how to write web pages to suit an iPhone. Basically using a viewport meta in the head section of your HTML pages. However the original detailed page, called something like designing content for iPhone, disappeared from the Apple web site. I had reached the tearing the hair out stage before I finally found Safari HTML Extensions, which included details of using the viewport meta. This page is not nearly as helpful as the original page, but at least covers the attributes of the viewport meta.

<meta name="viewport" content="width=device-width">

<meta name="viewport" content="height=device-height">

I realise these two meta tags are usually shown with width equal to 320 and height equal to 480, rather than device-width and device-height. Using the more generic term means that if an iPhone like device with a much larger (or smaller) display appears, the device will add the correct pixel values. For these pages, this would leave an iPhone showing my page content with a reasonably sized font.

The sales folks at Carlyle Gardens phoned. Builders are still at the house. Landscaping is not getting anywhere (not that this surprises any of us). Meanwhile, the rain from Tropical Cyclone Charlotte has been drenching the north. There was flooding in Townsville CBD, among many other areas, but much of this is related to the very high King tides at this time of year, combined with 200 mm or so of rain. We will see if we hear further reports over the next few days.

My Apple wireless Bluetooth Mighty Mouse batteries failed today, 12 January 2009. This blog entry is to help me keep track of the battery life of an Apple Might Mouse that is used every day, and never switched off. The replacement batteries were Lithium, as usual.

The rain depression left over the north after Tropical Cyclone Charlotte dispersed over land near the Gulf dropped 25 mm overnight in my rain gauge. Airlie Beach is under a cloud, but the weather radar does not really indicate very heavy rain. The winds were more active than I expected before dawn. They made enough noise that I couldn't sleep, so I started checking the news around 5 a.m.

By early evening we had another 25 mm of rain on our balcony. We also had a heap of heavy winds all day. This seemed the major complaint in towns all along the east coast. Winds heavier than expected.

Updating my old blog posts to XHTML standard comes next. These make a easy test case since their format is already highly standard, and all of them are valid HTML. In addition, I do not use images in my blog. This means the number of problems should be minimal.

First steps were to look for potential problem tags. To me this means any tags that do not require a closing tag. I basically did not want any self closing tags left within the body element of my web pages. <hr> and <br> are obvious ones. A simple grep through the directory revealed 46 <br> tags, mostly in early blog pages.

grep "<br>" *.html | wc -l

With all the 46 <br> and 46 <hr> removed, I searched for images. The only images were two Google charts. I decided to turn the two images into objects, thus removing the img tag.

Another potential problem area is the anchor tag <a, normally used to indicate links. You can use the name attribute, often in some form like <a name="marker"></a> to indicate a unique location named marker to which to link. Having an empty tag seems to me to be asking for potential differences in how a browser handles it. In addition, it encourages people to make erroneous links with block elements like header inside an anchor. Finally, when styling the anchor element (often done for links), you may accidently style more than you intended.

The obvious solution is to use the id attribute to assign a unique marker to an element. The id attribute can be applied to any element within the body of a document. The id attribute identifier must commence with a letter, but can include letters or numbers thereafter. Apart from being unique within a document, an id attribute can not contain character references, say to international characters. The id attribute and the name attribute share the same identifier list. Another reason to avoid name (but watch for any in a meta in the head). These are not serious restrictions, if you take care.

Since id can also be the target for styles, I believe you should think cross document, or project wide in the use of any id name. If you will be using an id for style purposes, consider working out a pattern of id for the entire project. Use that id consistently, always applied to the same type of element, always styled with the same css. For instance, an id like header, footer, sidebar, navigation, menu, will almost certainly be used in multiple pages. Think about such identifiers, and keep their use consistent. You need the equivalent of a writers style guide for their use. Or at least a cheat sheet.

It turned out I had only three name attributes in my blog, in two early pages. One left over from an earlier mechanical cleanup, two were errors where abbreviation was the correct tag to use.

Organic Light Emitting Diode displays emit light from the display, not from a backlight behind the display. They are potentially capable of brighter and more brilliant displays, unlimited viewing angles, and low power consumption (no backlight). As well, they should have a fast response time. At present efficiency and lifetime (especially in blue) are problems. In the further future, video wallpaper is not impossible. OLED can be flexible, and thin (under .05 mm). Samsung demonstrated a 4 inch 480 x 272 flexible display, with 100,000:1 contrast ratio. At the moment most OLED devices are used in phones, although there are a few, very expensive, larger displays.

Samsung are getting out of (relatively) cheap to build passive matrix Organic Light Emitting Diode (PMOLED - used in small screens like cell phones) in favour of more difficult Active Matrix Organic Light Emitting Diode (AMOLED). This leaves one PMOLED manufacturer. This move is despite higher power drain from AMOLED. Samsung recently displayed 14 inch and a 31 inch full HD panel. LG displayed a 19 inch panel.

At CES, it seemed an LG 15 inch display to be produced in June. The LG widescreen panel has a 1,366 pixel by 768 pixel resolution and has a lifetime of 30,000 hours. LG sales manager said not a good time for us to launch this type of luxurious product because the market is in recession. LG are waiting for customers to put this in laptops. Interestingly, LG Display recently disclosed to the market that Apple were going to prepay a half billion dollars in a five year deal. LG Display already supply 70% of LCD displays used by Apple.

I thought there might be a connection with a revised MacBook Air midyear. The current model 13 inch MacBook and MacBook Air have an 1280 by 800 (native) display. However despite the large display surround in the MacBook Air, I just can not see how a 15 inch display can fit in there. The whole diagonal is barely 15 inches. Now if the LG display had been 14 inches, who knows.

On the other hand, a 15 inch MacBook Air? I could go for that, if the battery life could equal a work day. I already know a 15 inch Powerbook will fit in my normal air travel bag. The MacBook Air is more powerful than a Powerbook. I don't recall ever using my DVD drive or Firewire for a camera while travelling. I think I could like a 15 inch MacBook Air as my travel computer (especially if a matte display were available as a glossy display). The weight saving would be really welcome, as airlines here have very strict limits on carry on weight.

The A$1.2 billion water desalination plant at Tugun in the Gold Coast not only missed its November start date, but parts intended to last 50 years are already rusting. So 125 megalitres a day of water will not be desalinated. This at least has the advantage that the 4kWh per kilolitre energy cost is not wasted. Especially when all this energy comes from carbon dioxide emitting fossil fuels, which in Queensland means coal.

Desalination plants are futile, idiotic white elephants. If you can't find enough places to put dams, then you need to stop population increases in the area. Running out of water means the universe is saying there are too many people.

There was another 35 mm of rain overnight according to our gauge. Several heavy falls around 3 a.m. woke me up. News reports say 6 metre tides at Townsville added to flooding in some areas.

I started looking more carefully at converting old blog pages from HTML to XHTML. I would need a .htaccess file to do redirects from all the old HTML pages to new XHTML pages. First ls *.html > htaccess so I have a list of affected files for the real .htaccess file. I can leave all the old HTML pages in place while organising this change, but needed to make sure the files on the server were otherwise up to date.

Change 46 instances of the Doctype from HTML 4.01 to XHTML 1 Strict using Smultron (or any multi-file editor) to change all files at once. This tells a validator which Doctype to validate against. It also tells a user agent (web browser) to use Strict mode, not Quirks mode. Note the Doctype tag must NOT be self closing.

Change 46 instances of an HTML tag with a language attribute to an HTML tag with a pointer to the XML namespace. I also include the language attribute here. An XHTML file must have the pointer to the XML name space. It is one of the two requirements for a file to be XHTML. The other requirement is to be served as an XHTML file, not as an HTML file. Note the HTML tag must NOT be self closing.

Delete the content-type meta that says it is text/html, and that the character set is ISO-8859-1. The server is what tells a browser whether the file is HTML or XHTML, so the meta would be ignored in any case. If it is XHTML, then the default character set is UTF-8, not ISO-8859-1. You do not need to say this. You do however need to make sure all the characters you use in the page are correct.

Check there are no self closing tags anywhere in the body of any page. Convert all the meta tags into self closing tags. Convert the link tags in the head into self closing tags. Check that the Doctype and the html tag and head tag are not self closing, and that every other tag in the head element is self closing. If not, consider whether that tag should exist.

Now comes the nasty work. Change the extension of all the web pages to .xhtml using any tool that changes multiple file names easily. Since a Macintosh has all the standard Unix tools, I used Terminal.

for i in *.html; do mv "$i" "${i%.html}.xhtml" ;done

Use a standard web browser to check how many web pages you broke. This will probably actually be every one of them. Firefox has excellent web developer tools, but I use Safari for this first pass. This is because Safari shows the web page up to where it detects the error. You get better feedback that way.

I opened the ten pages I had for 2005. Pages sizes were between 44 and 60 kB, with an 80 kB and a 120 kB. My earliest blog page had a dozen or two errors, all of them paragraphs that did not have a matching closing tag. Not required in HTML, so it is easy to miss, even on validated pages. As I found patterns to my errors, I did search and replaces over all 46 files. By the time I reached the last file for 2005, I was down to two or three errors a page. That was just after lunch, around 1:30 p.m. I am fairly pleased with the error rate so far.

Going through the 2006 blogs. August 2006 was already well formed, with no errors. That is the first page I did not need to correct. I completed 2006 at around 4:30 p.m. Wasting a lot of time reading my posts, wondering which pieces can be repurposed.

Started on 2007. The May 2007, July 2007, and November 2007 pages had no errors. I had 2007 completed before 7:30 p.m., despite a complete turkey dinner, with sparkling wine (France will not let us call it champagne), and starting the washing up.

The last year was 2008. October 2008 had no errors. The corrections for XHTML in the month entries were all complete at 9 p.m.

Needed to correct all the links between pages, to allow for the change from the .html file extension to .xhtml file extension. Why make that .xhtml change? Because my browsers (Safari, Opera, Firefox) then all identified the files locally as XHTML and used an XML parser on them.

While doing this, I discovered many links formerly made to anchors with a name attribute still commenced with a numeric identifier. However the anchors had all been changed to tagging a heading using an id attribute. These must commence with a letter, not a number. For h2 headers indicating dates, I had used shortened ISO date (20090114) for that day, preceded by the letter d (d20090114). For h3 headers indicating an article, I had used the same ISO date, preceded by the letter a. The first article of the day ended with the letter a, the second with the letter b, and so on. This article, being the fourth of the day, is a20090114d.

More rain overnight, but not a real lot. Maybe 5 mm or so. Enough rain sprinkled in the morning that I decided to wait until well after 7 a.m. to go out to the newsagent for the local papers, and to get breakfast in town. When I emptied the gauge in the evening at 7 p.m., there was 40 mm.

The late start in the morning means the garbage truck passed before I left. The rubbish bins had all too obviously not been put out at all today. It seems I was not the only one to think there was enough rain to postpone tasks. The gardener and the handyman at the Whitsunday Terraces had a cement mixer under the shelter of the car park. They started on a neat little concrete block edge to the garden bed adjoining the Florin Terrace car park, along the footpath. Nice to see that started.

Tried to mark a term as an acronym in XHTML 1.0. The W3C Validator accepted this in a file upload. The RELAX NG validator does not. So I uploaded my file to the W3C validator as if it were XHTML 1.1. This time the validator objected to my html tag including not just the correct xml:lang="en" but also the older HTML attribute lang="en" (I left that for possible compatibility with HTML). No objection to acronym.

In contrast, Henri Sivonen's RELAX NG validator reports Error: The acronym element is obsolete.

The XHTML 1.0 Strict Doctype still includes acronym. For that matter, so does the XHTML 1.1 Doctype. How can an element that is in the Doctype be obsolete?

So I tried the Validome but it timed out. Relaxed passed my file. Page Valet was not able to accept my file upload. XML Schema Validator passed my file as well-formed and valid.

Apple CEO Steve Jobs to take medical leave until the end of June. Says his health-related issues are more complex. Apple is doomed! At least, that is how this will get reported.

Health issues involving recovering from weight loss remind me that a vegetarian has a much tougher task than does an omnivore. Especially after a complicated and invasive Whipple procedure that removes so much of the stomach and intestinal structure. Obsessive workaholics probably make really bad patients as well.

I doubt that Steve Jobs needs to be involved in every day to day decision at Apple. It is no secret that Tim Cook is an exceptional executive, and probably just as obsessive about getting things right. Phil Schiller did a fine Keynote. Jonathan Ive will still be doing neat designs. Look at the depth of Apple executive talent. What Apple needs Steve Jobs for is to stop committees going forward with products that do not make sufficient sense. It is not that hard to come up with ideas that might be great. Apple need a NO man.

My expectation is that Apple will release several excellent products before Jobs returns. If one of those is a big iMac with a matte display, I will certainly buy despite the recession. Or I might be tempted by the new 17 inch MacBook Pro with matte display, although with a MacBook Air, I really do not need another notebook. I doubt I will be alone in buying something.

Having Steve Jobs less involved for a while may be better for me. I want Apple to make several products that I would really like, as a technophile. A MacBook Touch, midway in size between MacBook Air and an iPod Touch. Something to compete with Acer and Asus, a cheap computer with razor thin margins. An xMac mid size desktop with desktop class processors and plug in video card, just like everyone else makes, with hardly any margin at all. With a committee maybe not being able to say no, I have a chance at both, whether they make long term financial sense or not.

There were once heaps of complicated ideas for doing multiple column layouts in HTML and CSS. My general impression was most broke in liquid layouts. It seems some new faces have appeared in the web layout field. Matthew James Taylor has several examples. This equal height CSS page is explained. No hacks, and it has a liquid layout. Same idea can be used for any number of columns. He also presents some nice clean centred horizontal menus that are liquid. While I have not tried any of them, on a quick look they seem to have nice clean HTML and CSS.

Reports of SubMiniature Notebook Computers (incorrectly called Netbooks) having a high rate of return. An early post (March 2008) was about unreliable solid state drives, probably in small notebook computers, perhaps from Dell, perhaps using Sandisk SSD. Just this month, One Laptop Per Child have halved their staff with sales a fraction of their original expectation.

A January 2009 posting claims Mini-Notebook Computers Misunderstood, and this is retailer presenting mini-notebook computers wrong. Not getting users to understand a mini-notebook is intended as a second or third computer, a supplement to your desktop or full size notebook computer. In particular, not explaining that Linux is not Windows. There are a lot of people out there who do not understand that computers can run something other than Windows.

Customers return Linux mini-notebook computers to retailers at four times the return rate for Linux as for Windows XP. Ubuntu confirm high Linux return rates when asked. Up to Linux return rate 20%, vs 5% for Windows XP. Of course, this still means 80% of buyers kept a Linux min-notebook computer.

Asus will offer Windows 7 on future EeePC mini-notebook models. Will move straight from Linux and Windows XP to Windows 7, and not use Windows Vista.

Issues with mini-notebooks has advice on what min-notebook computers need. The boot sequence to disappear. Batteries to last longer. Persistent wireless internet connectivity (not available most areas). Touchscreen display. Lower prices, as mini-notebook computer can not cost more than a regular full function notebook computer.

Psion netBook is a trademarked term for the 1999 Psion netBook. This had a 7 inch colour display, was instant on with no boot time, a seven hour battery life (10 hours with modern batteries), accepted a PCMCIA wireless connection, had a touch screen, and ran the Psion Epoc operating system. Netbooks are not new. Computer makers just tried to hijack the trademarked term. This is why I keep writing mini-notebook computers.

Huge rush to download Windows 7 beta, during this second anniversary of Windows Vista (released in January 2007). I note also a certain continuing demand for Windows XP. Does this say the long awaited, and much delayed, Windows Vista has not been universally acclaimed? Sounds a bit like it to me. Microsoft seem to be saying release data January 2010, three years from the release of Windows Vista.

I have no idea what Vista is really like. It may be great. Windows XP on an IBM ThinkPad (a robust notebook, back when IBM still did PCs) was enough to convince me first to find a way to install Windows 98 instead. At least I knew how to fix the bugs in Windows 98 (mostly by removing cruft with Win98 Lite). My Linux using friends told me Linux was not really ready for the desktop. Helping a friend with a data file conversion problem on a Macintosh (Unix tools solved the problem) made me look at Macintosh owned by other friends. That put me on the path to buying an Apple Macintosh PowerBook at Las Vegas in February 2004 (yes, I had a good visit in Las Vegas). I have not used a Microsoft operating system since then. I can not say I have missed the fiddling and mucking around.

Because CSS Selectors have not been well supported, I make little use of them. However they are really very handy. I have been using attribute selectors for some time in my browser user style sheet to suppress the display of annoying advertising. I have failed to use attribute selectors to improve the appearance of my own web sites.

A:link[HREF*="advertising"] IMG, { display: none !important; visibility: hidden !important; }

One typical use is for styling links. See link styling using MIME types for an example of adding a different image to links to PDF, MS Word and Quicktime files.

/* specific styles for the different link types */ a[type="application/pdf"] { padding: 0.5em 0 0.5em 20px; background: url(pdf.png) no-repeat center left; } a[type="application/msword"] { padding: 0.5em 0 0.5em 20px; background: url(doc.png) no-repeat center left; } a[type="video/quicktime"] { padding: 0.5em 0 0.5em 20px; background: url(qt.png) no-repeat center left; }

The major problem with this approach is that not every link actually contains a proper MIME type. This is especially the case in blogs, where the URL of a web page is probably just pasted in without any thought of including the type attribute.

Attribute selectors can target the existence of any attribute. For examples of attribute selectors, targeting an anchor that has a title attribute: a[title]

Target only the entire content of the attribute. a[title="whatever"]

Target the first item of the content of the attribute. a[title^="whatever"]

Target the last item of the content of the attribute. a[title$="whatever"]

Target an item anywhere in the content of the attribute. a[title*="whatever"]

The use of ^ for start, $ for end, and * for a universal selector will be familiar to Unix users.

So to target an external URL (as distinct from say an internal link), you might be able to use a[href^="http://"]

To target a PDF file, perhaps to add a PDF icon, you may be able to use a[href$=".pdf"]

Daniel Lyons says Imagine what it might be like if the Church of Scientology went into the consumer electronics business, and you'd have a pretty good picture of how Apple operates. That is not a bad sentence. Lyons seems to hate Linux and Apple, hence most of what he writes is editorial. Not sure why he hates. Don't care. Lyon was far more entertaining when he was being Fake Steve Jobs.

The rubbish bins did not go out at the Whitsunday Terraces this morning, despite it not actually raining. At least, they were not out as at 7 a.m.

I tried watching the TV news at 6 p.m. No signal. Not just on Free to Air. The cable channels were also dead. Either a cable break, or a distribution amplifier fault. A rescan of the channels via my tuner did not change anything. If I remember, I will try to scan the channels on the tuner in the morning, in case it got fixed. If not fixed, I must try to draw it to the attention of whoever is in reception.

On the other hand, it is hard to argue that being able to watch TV is worthwhile, or an advance in civilisation or anything. But since I did want to catch up on the news, I downloaded the Channel 9 news videocast into iTunes.

10:40 p.m. Jean told me that the water supply was now out, and that this failure was very recent. Doing really well tonight.

Apple paid dividends from 15 June 1987 to 1995, but has not paid dividends since. Apple did stock splits. Apple is sitting on a pile of cash (around US$25 billion), as is prudent for tech companies in fast changing areas.

If you are investing in a company that does not pay dividends from earnings, you are essentially gambling on the stock exchange, not investing. Your only hope of a return is a better stock price.

Airline baggage limits in Australia may be lower than international visitors expect. All airlines charge for all checked baggage, including the first bag, except Qantas. There is no free checked baggage allowance. Limits on weight and size are a matter of airline policy. Your bags need to be within the limits for the airline who permits the least baggage. Just because one airline accepts a bag does not mean the next will do so.

In general, a carry-on not exceeding 7 kg (about 15 pound). A computer in a slip case can generally be carried separate.

Typical Australian domestic carry-on baggage size limits are 48x34x23cm, or 105 linear centimetres. This is closer to 19x13x9 inches, smaller than the common USA size limit of 22x14x9 (45 linear inches).

In general, one single item of checked baggage, not exceeding 20 kg (about 44 pound), however the limit on Tiger is 15 kg. All airlines charge for checked baggage, except Qantas. Some codeshare Qantas flights may be on Jetstar, which charges for all baggage. Charges may be considerably less if pre-paid. Airport charges range from $5 (Tiger) through $20 (Virgin Blue) to $40 (Jetstar). Be warned.

Additional charges apply for excess weight or a second bag. These charges can be very high.

Qantas Carry On luggage. As light as possible, but not exceeding 7 kg (15 pound) in total. Total dimensions 105 cm (41 inches), but some variations allowed. On Dash aircraft, one item only, 4 kg (9 pound) weight limit.

Economy Qantas checked baggage is reasonably high for international, although some island flights have 14 kg limits. Additional baggage on Qantas domestic may be as low as A$20.

Virgin domestic luggage limits is 23 kg. Checked baggage is charged for at $20 unless pre-paid at a cheaper rate of $8 per sector.

Opera complains about Internet Explorer, asks for Internet Explorer to be removed, or Microsoft to include other browsers as well. Microsoft abusing monopoly power, says antitrust European Commission. Microsoft responds to European Commission.

The FTC in the USA says the courts found Microsoft has an operating system monopoly, characterised by having market power, and doing exclusionary or predatory acts. In settlement, Microsoft agreed to clean up its act. The USA findings would probably influence the European Commission decision to act on the complaint by Opera.

Microsoft have available a number of articles regarding their level of HTML and CSS support. Understanding CSS Selectors explains and lists which versions of IE support which selector. In particular, IE7 supports attribute selectors.

Internet Explorer 8 attempts to support Standards by Default: What Does It Mean?. The article explains additional browser modes for backward compatibility with earlier versions of IE. There is an extensive list of HTML and CSS support in IE8. An extensive table of CSS support in various versions of IE. You should also see What changed in Internet Explorer 8. Additional pseudo-classes are among much extra CSS support in IE8.

If you want to kluge support in to your web pages with conditional style sheets, there is a list of conditions for various versions of the IE browser. Anyone attempting to do server side browser detection should see What Will Windows Internet Explorer Report as the User-Agent String?

Jean told me this blog did not have sufficient indication of each blog post being a separate two column article. Needed more to indicate there was something. The problem was a lack of background colour on the h2 and h3 headers. In Opera and Safari, the drop shadows made the nature of the article fairly clear. However Firefox will not show the drop shadows until version 3.1 appears.

The style sheet used in this blog was a patch of one patched from something else when I first started writing the blog several years ago. Very messy. Plus I was in a rush, and colours like red and blue were easy to type ... I have made a few minor changes, and done some cleanup, but there is a need for a heap more. Colour choice is still terrible, and I have never had a good eye for colour. Luckily Opera have a tutorial on colour theory that I can work through sometime. Rich in Style has a colorizer tester to see what colours seem suitable. There also seems to be Color Lovers, a website about colour choices.

Colour choice on my blog is still terrible, and I have never had a good eye for colour. Luckily Opera have a tutorial on colour theory that I can work through sometime.

Rich in Style has a colorizer tester to see what colors seem suitable. The Well Styled color scheme generator seems pretty extensive. Elsewhere there is a contrast analyser to usability testing.

For every percent by which unemployment exceeds the (estimated) natural unemployment rate, real gross domestic product is reduced by 2 or 3 percent. The difference between potential and actual GDP is called the GDP gap. This rule of thumb was proposed by Arthur Okun in 1962 (Potential GDP: its measurement and significance, American Statistical Society, Proceedings). Okun was primarily concerned with a policy outcome of full employment. Later economists would certainly put a higher priority on low inflation. Although a Keynesian, Okun would be unlikely to be happy with the implication that full employment implied high inflation.

Okum's Law seemed appropriate to many developed economies over the thirty years since Okun proposed (but did not name) it. However several factors in it (including GDP at full employment) can only be estimated, as measurement is too difficult. It does however help explain periods of jobless growth, with increased productivity, increased workplace participation, increased GDP, but no decrease in the unemployment rate. In more recent decades, the concept of the GDP gap has morphed into a more Friedman oriented monetarist view, according to writers like Tim Congdon in Keynes, the Keynesians and Monetarism. See also Harberger's Triangle for a diagram, deadweight loss for a summary, and the fundamentals of rent seeking by Gordon Tullock for a brief history.

Apple risks green makeover with corporate social responsibility snub, as the Apple board recommends rejecting a shareholder request from an environmental group for more green reporting. I first noticed that in Exception magazine.

At least one financial analyst said long ago in November that Apple will Beat 2009 Revenue Consensus. The article points to the nature of Apple's deferred revenue from iPhone. It speculates that some analysts may miss some of the subtle nuances of Apple accounting. It rather kindly suggests these may be analysts who do not specialise in Apple. I would suggest there are a goodly handful of analysts who comment about Apple from time to time, who probably could not thoughtfully read a set of accounts. But perhaps I am too cynical.

Added a little PDF icon to any link to a PDF file. This is to help viewers decide whether they want to click a link to a PDF file.

Technically, I added the image to any link that has href as its attribute, where the last content of the attribute is .pdf The icon is placed right and centre, and not repeated. I also had to pad the link a little to allow space for the little icon. Since my pages are always liquid designs, spacing of text is usually best done in em, not in pixels. In the case of an image, we do know the size, but I continued to use em. It is fairly safe, unless the browser is set to show very very small type.

a[href$=".pdf"] { background: url(pdf_icon.png) right center no-repeat; padding-right: 1em; }

I will gradually add a variety of icons to my file links, on links that do not go to standard web pages. Obvious targets would be those annoying pages that open on MS Word or Excel documents, or similar non HTML content.

Web browsers that support experimental use of the CSS feature column-count have received this blog in a two column layout. Column count is supported by Apple Safari and Mozilla Firefox, and presumably by Google Chrome. Column count is not yet supported by Opera or IE, however IE can not display this blog.

Two columns does not suit an iPhone, whereas one column looks nicer. Plus if someone has a large display, and keeps their browser occupying most of the display, they could probably use more columns, say three or four.

Media Query is supported statically by Safari. It is also supported dynamically by Opera (which does not yet support column count). It will be supported by Firefox in 3.1, when that is out of beta. I used two media queries. One checks for a max-width of 480, and changes column count to one, to suit the iPhone. The other checks for a min-width above about 1200 (still fine tuning that figure), and changes column count to three, to suit wide displays. In addition, it changes the max-width to 70 em (still fine tuning this) to make the total width of the text wider.

Actually, not in hot water. Jean's hot water heater broke last Wednesday or Thursday. It is summer in the tropics, so hot water is mainly something you use for washing dishes. Plus the apartment has two bathrooms. Redundancy is good. When we opened the panel hiding the hot water heater, it was a 1993 model, and the tank was leaking. Those little Whitsunday Terraces apartment hot water heaters typically only last 15 years anyhow, so a straight replacement seemed more sensible than looking for repairs.

On Monday, Jean finally phoned a plumbing company that works in the Whitsunday Terraces resort regularly. They said they would arrive 7 a.m. (and they did). Gave a quote over the phone for replacement. Jean asked if they wanted to be paid by cheque, credit card, or cash. For cash they would take around 10% off. I had been to the money machine a few days previous, so a discount for cash was fine by me.

My CSS lacks class. Actually, it styles lots of things that do not exist, because the original CSS file was converted from a much older CSS file, and never cleaned up. So I needed to know what classes I was really using, and which could be discarded.

A quick grep for all the class entries, then excluding those I already knew would be common. Because attribute content must be quoted in XHTML, I had to escape the quotes with a backslash.

grep "class=" *.xhtml | grep -v "class=\"date" | grep -v "class=\"article" | grep -v "class=\"totop"

ACMA takes SMS Spammers to court in Brisbane. The Australian Communications and Media Authority took action in the federal court in Brisbane against SMS spam companies. ACMA allege Mobilegate Ltd, Winning Bid Pty Ltd and International Machinery Parts Pty Ltd send unsolicited SMS messages.

Some people think you should not sign enforceable undertakings just to get ACMA off your back.

On the computer side, an interesting interview with an adware author tells about some of the nasty stuff he wrote. Some of the techniques used to stuff up PCs really are tricky.

Removed over 1400 links to the top of each web page. They looked a bit messy. They were not part of the content of the page, just navigation that seemed to be in the way. My reasoning is that everyone who uses a browser probably knows by now the keystrokes used to go to the top and the bottom of a web page.

The fussy old fashioned navigation menu at the top of the page had to go. I added the Eric Lindsay's Blog site name to the level one heading. To get it on two lines I used a div, since that is a block element. No way will I use a <br> tag in an XHTML site. Since many people are used to it on many web sites, I figured the site name could have a link to the home page. I wanted previous and next navigation in an arrow next to the month name.

I decided to try the left-pointing double angle quotation mark «, also known as the left pointing guillemet, to indicate the previous page. Also the right-pointing double angle quotation mark », also known as a right pointing guillemet, to indicate the following page. I am not particularly happy with this choice, but it seemed slightly better than a less than or greater than sign. Still a disaster next to the month heading.

Eventually I used the character entities for some mathematical arrows, &#8592; ← and &#8594; →. I was not willing to allow underlines in a h1 heading, but to make the links slightly more obvious I did use CSS to bold them when the cursor hovered over them.

All my recent blog entries have been in a h3 heading with a topic. However some early entries went straight after the h2 date element. I wanted to find and alter all of these. I figured a temporary set of adjacent selectors for h2+p in the CSS would make them obvious, if I coded the text to red. Opps! I had an article div around any entry, for the multi column layout. So I had to search for h2+div. That worked. Sure is a lot of them to fix. I had to do them manually, since each needed a new heading, and category tags. I had not completed these fixes and it was approaching midnight. I gave up for the night.

Apple iWork '09: Serial number not required for installation of retail boxes. If you have downloaded the trial version, and you buy online, you will get an emailed serial number that you can use to unlock the trial. However if you reinstall from a retail box, you do not need a serial number.

While I have never had any serial number trouble with the iWork 08 family pack, or the previous versions, this is more convenient. I downloaded the trial iWork, and hope to use it on my iMac G5 shortly. However in February I will be travelling a lot, so then I will install the trial on my MacBook Air, to continue with whatever I have started. In March, I will be able to order the family pack of iLife, iWork and OS X Leopard, when I am able to take delivery of the actual box.

That will give me a legitimate original DVD copy of Leopard for each of my three Macintosh. I already have the OS X Leopard family pack, so I am already legally able to use it. However at some stage someone will inherit my PowerBook, and someone else would inherit my iMac G5. I would not have a legitimate copy of Leopard for one of these computers otherwise. They will have to make do with the older versions of iLife and iWork.

In the lead up to Apple announcing their most recent quarter earnings, some articles have appeared in the press. Fortune puts a Spotlight on Apple’s hidden revenue stream. Like, deferred revenue was not known previously? It was announced at the time. Apple even drew conspicuous attention to it last earnings call. Bullish Cross goes with How the iPhone and Poor Apple Management have contributed to the Downfall of Apple. The shares are down, more than they need to be. So there are a lot of Dumb Analysts out there.

Personally I am pleased to see a company treating short term shareholders and day traders with the contempt they most certainly deserve.

The current trend to replace web browser text zooming with page zooming is likely to be an utter disaster. I realise vast numbers of web pages break when users text zoom, or set their default minimum font size to something they can read. Having a web page break because of increased text size means the web author got things wrong. So I can see why browser programmers decided to give full page zooming a chance. After all, it works for Opera.

The point to note is that Opera changes the text layout intelligently when a web page allows a fluid layout. This means text columns reflow to suit the available space. Opera does not produce a horizontal scroll bar until some unbreakable picture or similar is encountered.

Firefox also avoids horizontal scrolling of fluid layouts (although it is slower), by changing the text layout to suit the browser viewpoint. In short, it scales the site. In addition, Firefox still includes the text only zoom.

Apple Safari full page zoom was mentioned in WebKit in March 2008. So far it has not shown up in Safari menus, which still contain text resizing. The question will be, will the Safari page zoom be optional (like Firefox) and smart (like Opera).

Microsoft Internet Explorer has long had issues with resizing some text, but has had text resize for a long while. I regret that reports of effects of page zoom on layout in IE7 were not encouraging. It zoomed the site. IE8 has adaptive zoom, which changes layouts and text wrapping to avoid horizontal scrolling, where possible. It both zooms and scales the site. Microsoft advise web page designers to allow for this.

The disaster is web designers thinking they can avoid fluid layouts because the user can zoom. Any web designer that produces a horizontal scroll bar is doing things wrong. A display is not a sheet of paper. Learn to design for any size display, any size typeface.

Apple developer have details of CSS support in the Safari web browser. CSS Property Functions are visual effects. Supported CSS rules downloadable fonts, @media properties specific to a particular media. Supported CSS properties from CSS 2.1. Includes -webkit experimental. These are all handy resources if wondering what CSS you can use with Safari and other webkit browsers.

Tried checking Apple financial results this morning. Not available as yet. Started running the laundry. Ideal Plumbing truck parked below. They were replacing the hot water service for our neighbour, just two days after ours was replaced. His was dated 1994, the one Jean had replaced two days ago was 1993. I took the opportunity to give Jim the manuals and connecting cable for the Garmin III GPS I had given him. Better for a boat owner to have an extra GPS.

Took out the rubbish. Once again the Whitsunday Terraces Resort had not put the garbage bins out for collection.

Took my morning walk. Dumped bottles in the recycling bins on the main street. Stop for conversation with a local about which developer is going to pull out of which building project. It seems I was correct that the garbage truck is not willing to enter some of the roads in new developments around Airlie Beach. They have an arrangement to collect garbage bins from the bottom of the hill. I said in my Airlie Beach Bum web site that was going to happen on some of these ridiculously steep private roads. I simply can not see why some of these places ever got council development approval. They are only suited to mountain goats.

Collected the two local papers, plus The Australian, at the news agency. It had what appeared a helpful two person analysis of Barack Obama's inauguration speech. Interesting that someone who is that good an orator did not give in to the impulse to throw in at least one unifying phrase intended for the history books.

McDonalds for breakfast. I know it is popular to sneer at McDonalds. However it is hard to really wreck bacon and eggs, which are my once a week treat. McDonalds are open, which is more than you can say for most cafe in Airlie Beach before 7 a.m. (McDonalds are 24 hours). The service is reasonably quick. They have the Courier Mail newspaper available for guests to read (I like reading it, but not enough to buy it). They have WiFi, which is free (after login) to all Telstra iPhone users. They do not hassle you to move on (I can space out my drink long enough to quickly read the paper).

Collected the mail, for the few remaining weeks while we still have an official Post Office. Collected a loaf of bread at Brumby's. Walked up the twelve flights of stairs at the Whitsunday Terraces resort. Then I had to put out the laundry that Jean had ignored. It was way too hot in the sun out on the balcony, even not much after 8 a.m. At this time of year in the tropics I probably have a time to sunburn of around 15 minutes.

Then it was shopping time, by car, which neither of us like. Rivers had been advertising cheap backpacks, so we got one each. Jean's leg went bad again. Not so much she could not walk, but not a good thing before a trip. Coles for food shopping, our last large shopping visit. I do not know why our bill always exceeds $100 these days. We were mostly buying sensible food, lots of vegetables. Got some last meals at Kitsons. We want the freezers stocked for ten days, but not more.

The sales person at Carlyle Gardens phoned soon after we returned home. They were expecting to have the new home handed over by the builders this afternoon. That is a little later than the earlier estimate of 19 December. It is good that it is not hanging on into February or March. However the timing was pretty bad, unless it could be done as a walk in handover. Next week is a holiday weekend. We will be away all February, and I have commitments towards the end of January. Looked like that gave us three working days, which did not look viable.

I had already lined up a van and (child) labour for a January move (during the school holidays), and had to cancel when the December date would not hold. I did not want to try to organise moving with even less notice, so that means March at the earliest.

While we do not know exactly how everything relating to paperwork proceeds from here, the two week cool off period is already impossible. We could see all manner of paperwork problems piling up over a short time while we had an inflexible deadline for leaving Australia. Our solicitor here is also not exactly flexible, which could complicate matters even more.

Jean said she didn't think the end of January was going to work for us. We would be back at the end of February, with only one commitment at the end of the month. Try early March for the handover. That way we have probably a whole month clear to work around potential problems. Odds are there would be no problems if there are no hard deadlines in the way.

Bullish Cross on analysts estimates of Apple results. Fortune reported the Apple Q1 2009 earnings smackdown. Mind you, the quarter ended in December 2008, January is just when it was reported. The actual Apple Reports First Quarter Results show record revenue, record profit, despite the economic downturn. Results are closer to flat than Apple probably like, but not down. Nine percent up for Mac, 34% up for notebooks, unit sales 25% down for desktops. Apple spent more space in their report explaining their accounting principles, presumably to help analysts avoid shooting their feet off again. Other than that, it was the usual Apple minimalism. The the best tables and easiest to read presentation Apple reports record profit for first quarter comes from Macworld magazine.

My initial change of link colouring using attribute selectors worked well, however after a link was visited, the attribute selector for the original link no longer applied. While you can live with it, this was not what I wanted.

You have to position CSS pseudo classes in a particular order, namely link, visited, hover, active (mnemonic LoVe-HAte), so that link actions work correctly. On Usenet I was told you may combine attribute selectors with pseudo classes.

a[href^="http://"]:visited {color: #4488ff; }

Normally an earlier style is replaced by a later style. However the pseudo classes of hover and active that followed visited did not function. This is because of CSS specificity. The more specific the style, the greater importance is placed on it. By having both an attribute selector and a pseudo class, I made this style more important than a simple selector style. So I had to repeat the combined attribute selector and pseudo class for both hover and active.

Normally the stylesheet cascade decides what style is applied. Whatever is last wins. However specificity overrides the cascade.

An id selector (#selector) has a specificity of 1,0,0, an attribute, a pseudo class or class selector (.selector) has a specificity or 1,0, and a regular HTML element or pseudo element has a specificity of 1. A string of selectors has the value of all the selectors included with it (link specificity is more complex than this). So the attribute selector plus pseudo class code in the previous article has higher specificity than a regular pseudo class.

Assassins of the Mind by Christopher Hitchens, in Vanity Fair, is a rift on Salman Rushdie, and self censorship in the Western press. Hitchens celebrates Rushdie's facility with the English language (his first language was Urdu). He continues with the history of fanatical Muslims attacking any criticism of their ridiculous barbaric religion, and its illiterate prophet. He points to the political advantages to Ayatollah Khomeini of Iran of making an attack on Rushdie. It was a distraction.

The result of this murderous fatwa and the bloody deaths that followed from it has been increased self censorship by the media in the west. Al-Jazeera understand the merit of showing violence in their coverage of Gaza. Welcome to the Al Jazeera Creative Commons Repository of selected broadcast quality footage, released under Creative Commons. Al Jazeera also understand the merit of not locking all their best footage away from potential viewers.

Western media seem to mostly understand only product placements and how to advertise consumer crap. What happened to quality newspapers and quality TV?

The showers that had been apparent to the north the previous afternoon reached us at the Whitsunday Terraces. At least we did not have the massive power failures reported from Townsville and Cairns, where it seemed most of the area lost its mail feed line. We had 23 mm of rain overnight, and now have cloud cover moderating the summer temperature.

During the global financial crisis, we see governments rush to preserve banks, with vast handouts, guarantees and buyouts. There seems far less of a rush to preserve the real economy. The mining and manufacturing industries to a much larger extent have to fend for themselves. However seeing finance as greasing the wheels of industry is not all that unreasonable.

What is unreasonable is the preservation of monopoly via intellectual property laws. The very limited time span of the patent system means the world can (eventually) benefit from invention. Some companies like IBM release patents for public open source use. True, there are rorts in patents, especially in software patents (how many are not obvious to a competent practitioner?) Even here, there are legal attacks on software patents. The ever extending life span of copyright is the most obvious example of companies pressing to have a monopoly forever.

Dean Baker, co-director of the Center for Economic and Policy Research in Washington, wrote a book about reforming intellectual property. He even has a section Efficient Mechanisms for Supporting Innovation and Creative Work, promoting methods that may be more efficient. The book is The Conservative Nanny State - How the Wealthy Use the Government to Stay Rich and Get Richer. Available in paperback, online, or as a free PDF download. Be warned, much of it is polemic.

Even large companies like Google open source some stuff. Open Sourcing Google Desktop Gadgets for example. More specialised, CSS Selector Shell, a simple Javascript tool for testing how a given browser parses CSS text. If I were trying to support IE6, equivalent tools would be essential. Luckily all modern web browsers do support selectors, even IE8 supports them.

Lots of companies are using open source. The Darwin core of Apple's OS X. Webkit is open source, and forms the basis for Google's Chrome and Apple's Safari web browsers.

Cloud Interop looked at customer lock in, and resistance to it. Comments noted interoperability often drifted to discussing standards. Noted that Cloud API often means Amazon Web Services.

Personally I am avoiding cloud computing. Cloud computing is not reliable. In the past year my internet connection was down for a week. I was out of range of the phone network for most of three weeks. This year I will be overseas mostly without network access for four week. I will be moving to a new home that lacks internet connectivity. Who knows how long it will take to get a phone line in - one person in the area has been trying to get a phone since last September. The basic fact is, that outside major cities, the internet is stuffed.

Recession hits economy, Microsoft to drop about 5000 employees (over 18 months), and all of a sudden Microsoft is dead? Sounds more like wish fulfilment to me. OK, Microsoft failed to destroy Apache, but it is not like Linux threatens Microsoft on the desktop. Actually Linux doesn't threaten anything on the desktop, as Linux is still almost invisible. Even on sub-notebooks, it was displaced by Windows XP fairly quickly. Not that XP helped Microsoft all that much. It costs around half what Windows Vista does. Microsoft will probably continue selling Windows XP for so called Netbooks until June 2010.

Microsoft have not been able to persuade everyone that their cash cow Windows Vista is not a dog. Service Pack One did not convince everyone either. Some people are saying that Windows always gets better after the second Service Pack. True of much software. It is interesting to note that Windows Vista is now expected to get a very reduced SP2 before Windows 7 is released. Internally Windows 7 still lists as 6.1, meaning this is the SP2 update to Windows Vista. By now, Microsoft must be willing to do anything to bury the Windows Vista name.

Much more likely that hardware makers will be buried. Except for memory, Windows 7 will run on lighter hardware than the previous version. What then is the incentive to upgrade your hardware? I see lots of businesses holding on to old hardware.

Meanwhile Apple are working on Snow Leopard. Also lighter weight, more finely tuned, faster (about time), and able to handle multiple cores. However it will not work on the old Freescale G4 and IBM G5 CPUs (like my 5 year old PowerBook and almost 4 year old iMac). Incentive to buy new hardware, higher. Plus Apple have the option of a miniature touch computer using ARM CPU running their OS X phone version. This will hardly touch their real computer sales.

Is it just since Barack Obama arrived? The White House is valid XHTML, Transitional alas, rather than Strict. Served incorrectly as text/html from Microsoft-IIS/. Much good web stuff here, but there are critics. Making Site Architecture Search-Friendly: Lessons From whitehouse.gov. Perhaps the Obama team are not as good as it first appears.

In contrast, Kevin Rudd, the Prime Minister of Australia does NOT have a valid XHTML web site. It is also Transitional, not Strict. Served incorrectly as text/html from Microsoft-IIS/. The character set is wrong. The server says UTF-8, the actual web page says charset=iso-8859-1. Despite declaring an XHTML Doctype, there is no xmlns attribute on the HTML element, so it can not actually be XHTML. Then they have filled up the page with scripting, hidden with HTML comments. However you can not get away with that in XHTML. It does not look like they understand XHTML. Better to declare as HTML. Probably a bunch of the 20 errors would go away then. No, expands to 44 errors, mostly due to all those self closing tags they added because they thought they were doing XHTML. You can now clearly see heaps of characters are wrong, probably because someone cut and pasted text from Microsoft Word without converting the entity types. Not much leadership here.

Mac Trojan Horse OSX.Trojan.iServices.A Found in Pirated Apple iWork 09, Intego breathlessly announce. So, an actual in the wild exploit for the Macintosh OS X, with 20,000 downloads according to the above report. Lots of reference to it in various news sites.

The usual story. It is not a worm. It is not a virus. It does not automatically infect a Macintosh. It is a trojan, a malicious file. Needs social engineering to get installed. You have to give it permission to infect you. Get the original iWork from Apple and there is no threat. That is a pretty easy solution.

Anyone who torrents commercial software probably deserves everything bad that happens. Unfortunately, people stealing via torrents are not the only victim. Taking over computers for inclusion in a botnet is big business. The botnets can be used for all sort of nasty activity, including distributed denial of service attacks on web sites.

Although not the fault of ISPs, I think eventually ISPs will need to step up to the bar and accept some responsibility for the cleanup. If someone has an infected computer, I believe their ISP should be responsible for disconnecting them from the Internet. I believe ISPs should be protected, by law, if they do so.

I decided to use attribute selectors in an attempt to show an Australian flag next to links that pointed to Australian web sites. Since I live in Australia, such links may have particular relevance to material I am writing, at least for any of my audience in Australia.

Cascading style sheets do not allow complex analysis of text. I can only really test for the presence of the Australian country code .au within a link. Such a simple test will also pick up any link that has .au anywhere within it, so there will be some false positives. In addition, any Australian site not using an Australian domain name (like my own sites) will not be detected, so there will be false negatives.

a[href*=".au"] { background: url(flag.png) right center no-repeat; padding-right: 20px; }

First annoyance of the morning was finding that the garbage bins had not been put out at the Whitsunday Terraces. I took out one Florin Terrace bin.

It was a cloudy morning, fairly typical for a market day, and rain was threatening. I didn't take my hat, which for me is very unusual. The markets were very sparsely occupied, with many regular stall holders away. No place to get breakfast, so after talking with a bunch of people I went to McDonalds. It rained heavily while I was there. Back at the markets, the sun came out. Without a hat I managed to get a sunburnt face.

Decided I needed to make some favicons. These are the little icons that sometimes appear in the address bar of a browser when you view a web site. This was the easiest favicon generator I found. Unfortunately ease of use does not compensate for lack of artistic talent and poor colour choices on the part of the designer.

For my own Eric Lindsay blog I did the letters , which shows a certain lack of imagination.

For the new Jason Sonnet web site for a young local Whitsundays singer and dancer I did much the same simple letters. Since I expect that brand new web site to need a bunch of updates soon, the favicon can be changed to something more appropriate later. Perhaps some musical notes.

My satirical Airlie Beach bum and Whitsundays web site got a favicon. I wanted something less nautical, perhaps involving bums on beaches, or footprints on sand, but on a 16x16 array, I do not think I could draw it.

My recent Carlyle Gardens Gnome retirement resort website got a favicon, after I failed to do a good background instead. The gnome looks more like an alien with bad colour sense, but that was all I could manage.

Once again the rubbish bins were not put out at the Whitsunday Terraces. Again I put out the Florin Terrace rubbish bins, at 7 a.m. when I went to the news agent. I was back well before 8 a.m., and the bins had been emptied, so I brought them back in.

The rain during the weekend had put 22 mm into my rain gauge, so I emptied it at 9 a.m. There was a threatening looking cloud approaching from the north east, so I brought the laundry stand in just in time to miss the rain. The weather radar was showing lots of scattered shows along the coast from Hinchinbrook to Mackay.

For some reason I can not get at my Dreamhost web hosting statistics. There are instructions on accessing Dreamhost statistics at WhoIsHostingThis, but my password is always rejected. I can get at the files by using FTP (the password works there). I have no formatting nor graphs when I download via FTP. It is very annoying. If it were a general failure, I am sure there would be many complaints, but that does not help me.It was not until I carefully read about users at my domains that I realised there was a third user name involved. Not using the password from the ftp access either.

Dreamhost have an interesting (and generally easy) home made web site control panel (not that horrible CPanel), but every now and then it bites me. Especially on things I hardly ever do, like checking web site visitor statistics.

This rather dated article explains why you can not trust web site statistics. That is part of the reason I rarely bother to check my web site statistics. The main things I check are whether I am getting an increase in requests for pages that do not exist. That may indicate one of my cross links or menu items is wrong. Or that someone else has a link to me, but got part of the web page filename wrong.

In case you are wondering what sort of statistics reports I get to see about visitors, here is an explanation from the people who made the Analog web site statistics package. I think they are basically pretty boring.

Web robots, spiders or crawlers are programs that look though web sites, and generally are intended to index web pages. Your web site statistics will show robots from Yahoo and Google, and possibly many other web robots from search engines and elsewhere. These blind, dumb robots can be your most important site visitor. They show humans how to find your web site. How do I get the best listing in search engines? It helps explain why I write my web pages the old fashioned way I do.

What I had failed to do is add a robots.txt file to all my more recent web sites. It is not required, but it helps avoid spurious entries in your web site statistics showing failed requests for robot.txt.

A momentary power outage at 9:15 p.m.

Australia Day. I had stuck an Australian flag from the news agent on a box last night. Pacific Sun, a large tourist liner, was anchored out in Pioneer Bay. I saw it when I was awoken by the rumble of thunder before 6 a.m. Starting a proper Australia Day with WeetBix for breakfast. Vegemite? There are limits to how far I am prepared to go. The BBC tried to explain BBQ.

I was just seconds too late to check that the garbage bins at the Whitsunday Terraces had been put out for collection. They had not. That is three days in a row the rubbish has not been put out. True, it is a holiday weekend, and maybe regular staff were absent, but this is just shoddy. One of the bins was already full, despite me putting it out on Sunday. Holiday weekends generally mean more guests, thus more rubbish.

The was a good collection of stall holders at the markets for the tourists from the ocean liner. I chatted with a few I know, but threatening clouds made me check the BOM weather radar. I phoned Jean at home, but she had already spotted the clouds and brought in the nearly dry laundry rack I had put out earlier.

Collected a six pack of family owned Coopers Sparkling Ale at the pub. Then to Brumby's for traditional Australian food like lamingtons and a couple of meat pies. The rain came pouring down before I could leave Brumby's. It was about five minutes before I could walk up the twelve flights of stairs to our apartment.

The next version of iPhoto has Places, taking advantage of GPS readings in digital photos. Previously iPhoto would let you see GPS readings, in Info about photos, but adding that GPS information to existing photos was not terribly practical.

Hardly any digital cameras include GPS information. I have been looking for this GPS facility in cameras since last century. Camera shops all basically regarded me as nuts.

However cameras in phones regularly include GPS facilities. The problem with the Apple iPhone is that although GPS relies upon satellites available anywhere not under cover, the iPhone also expects a cell phone connection (for the Google maps). 95% of Australia (and 100% of the rest of the world) are out of reach of a phone system due to geography (and vendor lock in and roaming charges when outside Australia). So I can not use my iPhone as a GPS camera in New Zealand.

I started seeking a standalone GPS app for the iPhone. Something that used standard GPS, and did not require aGPS (except to speed up the location process). Something that would set the GPS location for use by the camera. Something that did not rely upon access to Google maps.

InstaMapper allows real time tracking a GPS equipped phone. They store 100,000 points, which can be exported as KML or CSV. You could do things like embed on a map in Facebook. You can set the location update time from the defaults (in seconds) to minutes. The forums have posts from people using it outside the USA. The crucial point for me was while you are out of coverage, up to 100 positions can be buffered in phone's memory so that your track doesn't get interrupted. Unfortunately, if you turn off the application, you lose the data, so InstaMapper will not work for me. They are contemplating an offline mode.

Here I Am sends your location via email. No details of operation on the web site. No link to App Store. Their web site address includes illegal characters (spaces). Sigh!

Navizon offers software only cell phone location via WiFi and cell tower, using a collaborative entry system. GPS is used to allow participants to provide increased accuracy when mapping WiFi. Does not fit my problem space at all.

TrailGuru forums report it will (eventually) operate without cell phone connection. Photos from TrainGuru are not stored on CameraRoll (Apple restriction).

Google Earth is reported to be able to operate offline. For small values of offline.

GMap for iPhone claims to be standalone. Uses a massive map. Looks like it covers USA (in two map sets), with Europe coming. Not suitable for my locations.

GPS Kit runs Topo maps. Seems to be stand alone, uses data connection for map imagery, the info page, and to email your tracks. I wonder whether that will be enough to keep the iPhone camera happy about locations? I sent them an email to ask. So far that sounds the best bet.

Books Gone Wild says Time. Potted history of the vulgar rise of the novel, and some self publishing successes. But publishers can no longer afford risk, so their editorial filtering function is poorer. While electronic media has made self publishing affordable.

Meanwhile the old media of the book has moved online. The New York Review of Books has Robert Darnton on Google & the Future of Books. Google has digitised a vast array of books. A recent settlement with authors and publishers will permit Google to continue to digitise millions more books. This article points to the massive exchanges by correspondence between intellectuals of the 18th Century. Also, to how such delights were generally restricted to the well off, or those subsidised in some way. This is followed by a piece on the place of copyright in the US Constitution, essentially the congressional neglect of the limited time clause. It is a nicely informative article to this point.

However the main trust is the effects of the Google class action settlement with publishers and writers. Google have created the largest library ever seen. Digital it is true, but nevertheless enormous. Thanks to the class action settlement, Google also have a monopoly over all these extended copyright books. Who else can afford to scan and digitise over 7 million books? Google simply makes out of copyright works available online. The terms under which Google make access available to in copyright but out of print works will be the interesting topic.

The quality of the morning newspaper I collected at the news agent remains low. I was hoping the Christmas slump was over.

The Post Office agreed to note our PO Box that we would not be collecting mail in February. I hope that makes a difference.

Avis tell me they do not have trucks to rent.

Perhaps computer companies do better than the one out of two valid politician web sites? I looked at the list of World Wide Web Consortium members. These are the guys that establish the recommendations for web page markup. Then I picked the companies that actually made web browsers.

Apple make the Safari web browser. The main Apple web page is invalid, although there is only one single error. They also serve HTML, which is a wiser move than XHTML. It is however Transitional, not Strict.

Google make the Chrome web browser, like Safari, based on Webkit. The Google web page is invalid, and a real mess. With no Doctype, the Validator decided to check as legacy HTML 4 Transitional. This produced 65 errors, 8 warnings. The Validator is unable to check HTML 2 or HTML 3.2. Google have cut the white space out of their page source, probably to reduce load times. Google also has a bunch of scripting. The page source is hard to read.

The most used web browser, Internet Explorer. This is from the largest computer company of them all, Microsoft web page is invalid, with 177 errors and 37 warnings. A lot of these errors are probably cascading errors from the head element, and unescaped entities like & in URLs that should be encoded. The page is written in XHTML 1.0 Transitional. The page source is full of scripting, and very messy to read.

Mozilla make the Firefox web browser (available at a valid XHTML 1.0 Strict web page). The Mozilla org home page is valid HTML 4.01 Strict, with one shorttag warning.

HTML elements with optional closing tags are risky. Elements without closing tags seem to me really dangerous, and prone to producing errors when you convert between XHTML and HTML. Personally, I close all optional tags these days. I am so uncomfortable with self closing tags within the body element that I removed all br, hr and img elements from my work. The only place I use self closing tags is within the head element, and these will mostly be metas.

The Opera web browser home page is invalid XHTML 1.0 Strict, with two errors. They used an HTML element called string, that does not seem to exist.

Maybe computer hardware companies are more likely than politicians to have valid web sites?

The Dell web page is invalid HTML 4.0 Transitional, with 297 errors, 46 warnings. We have a winner for number of errors in a home page. OK, Dell is not a member of the W3C, so perhaps we had better stick to W3C members.

The IBM web page is valid XHTML 1.0 Strict. Well done, IBM.

The largest CPU manufacturer. Intel web page is invalid, with 31 errors and 2 warnings. It is in XHTML 1.0 Strict, which would be good. However what gave Intel the idea that showing an almost blank page is a good idea? The page relies on a (slow) proprietary Flash plugin to display any content. Landing pages went out with the ark. Using proprietary plugins to display something web standards can handle is just silly.

Try another computer hardware company. SUN Microsystems web page is invalid, with 31 errors and 18 warnings. It is written in HTML 4.01 Transitional, but for some reason heaps of Meta tags have XHTML style self closing slashes, producing shorttag errors. I suspect at some stage SUN realised XHTML can not be used on the web, and tried changing back to HTML, but had trouble with self closing tags. If SUN got rid of the partial XHTML markup, most of their errors would go away.

Sugar cane is a big industry here. However it is blamed for environmental destruction and nutrient run-off. The spectacular burning of sugar cane waste has changed. While nothing is a cheap as burning coal, sugar cane waste can produce power at around twice coal costs.

The sugar cane is harvested. It is shredded and crushed, extracting the juice, around 13% of the material. The waste (bagasse) can be burnt for co-generation power for the mill, and for sale as electricity.

Take a tonne of sugar cane. You get maybe 400 kg of bagasse. The remaining 600 kg probably contains 60 kg of dirt and similar stuff you can not use. About 400 kg is water. You get at least 100 kg of raw sugar, and 40 kg of molasses.

Previous and next links that look like arrows. The site needs to be fluid so I did not want to use a graphic. As they will be in h1 elements, I do not want search engines to be indexing weird characters. So I thought I would use attribute selectors on the rel="prev" and rel="next" attributes of the links, and create the arrows in CSS as before and after generated content.

However I am a bit uncertain about whether to put the character content in CSS like I would do in HTML. e.g. &#8592;

At present I am using attr(title) in my CSS but that feels somewhat strange. Works in Firefox, Opera and Safari.

h1 a[rel="prev"]:before { content: attr(title) ; }

h1 a[rel="next"]:after { content: attr(title) ; }

/* Pad our arrows away from the year and month in heading one. */

h1 div a { padding-right: 5%; }

h1 div a+a { padding-left: 5%; padding-right: 0; }

Luckily there is a good tutorial on generated content at Maxine Sherrin and John Allsopp's Western Civilisation. They also make the StyleMaster CSS editor for Macintosh and Windows.

Managed to totally forget that you can not put a block level element like a div inside a heading. Just like you can not put a block level element into a paragraph. It is still well formed, just the wrong element. The error handling in browsers is really excellent, so it rendered the way I expected. It was not until I asked about the unrelated problem above that people on Usenet mentioned (politely) that I needed to check that I noticed it.

OK, replace each div in h1 with a span. Style the span as display: block; so that the name of the blog, the year (in one span) and the month (in another span) all go on different lines.

I had made my h1 and h2 content a serif face, with a little additional letter spacing. The generated arrows do not look good in Georgia font family. I styled a h1 span a as san-serif so the arrows looked better. I also made them bold on hover, as well as change colour.

This is starting to get there. If only my colour sense was better. I do not feel confident enough to do a lot with the colours so far.

Next thing to change will probably be the links to each day. This is just a paragraph full of links, with white space from the line feed between each. The semantic web folks will say that it is a list, and should be styled as a list.

Dropped caps and small-caps added. Correct selectors for the first letter and first line of only the first paragraph in a div confused me, but I finally got it.

h3 + div > p:first-child:first-letter targets the first letter. I kept finding examples that wanted me to use a class in a span to identify the character, something I think is the wrong approach. I also made the first letter a serif, rather than the san-serif of the body text, to add interest. I keep contemplating a more ornate font, but the default fantasy fonts are usually dreadful.

h3 + div > p:first-child:first-line targets the first line. I am not at all sure that I like small-caps.

New Zealand money is what we need. The local bank branch had none the last two times I asked. Despite threatening rain, we went seeking some New Zealand money this afternoon. The place that was formerly a currency exchange was closed. Found a currency exchange in a travel shop. Too crowded to ask, but we can try again tomorrow. There is also a currency exchange hidden in the shop next to Neptunes Seafoods.

About 33 mm of rainfall on the balcony at the Whitsunday Terraces between Saturday and noon.

Jean is on Facebook. She said she found heaps of people we both know on Facebook. While I am suspicious (I already have a life, and it is already full), I set up another email alias, and started signing up.

Second step was to give them my email address and password. Not a Facebook password. My email password. That is not going to happen. The security implications of that are ... Well, it is totally unacceptable to me. I certainly hope that most people will realise this is an optional thing, used with your address book data for helping find people you may know on Facebook. Still unacceptable of course, despite privacy being dead on the internet.

Twitter signup was next. First attempt brought up a server failure, but it worked the second time. Twitter also tried to get mail passwords. No chance. We can find people the slow way.

Until I know more about their security setup, I do not plan to give them any more information about anything.

Not at all sure there is anything new here, except for using HTTP, dodging firewalls and overloading Port 80. I was using Unix mesg in 1987 to chat, and it was old then. The more commercial Instant Messaging companies are also old.

So I joined Facebook last night. This morning, I can not log on. Facebook is down for maintenance. Good start.

So I went to Twitter, thinking I could do a tweet about Facebook being down. I can see my account. I can sign on. However I do not seem to be able to add a comment. This is also less than impressive.

OK, so the Twitter web box for inward messages is on their home page, not accessed by signing in. That does not seem very intuitive to me, although it is easier to type. I wonder what happens when you remove the persistent cookie, or try posting from a different computer?

Oh my. I came back from my morning walk and breakfast, read three papers and added material to various of my web sites. Then I opened my email. 29 messages. I never get 29 messages. Excluding spam and lists, I get a message every day or so. 29 emails from Facebook! I guess I better learn how to navigate the Facebook system.

Facebook was too overwhelming. Especially the stream of consciousness from Gary. Accepted a bunch of friends requests, many from people I have not contacted in way too long. However many of the features, like messages, seem to be broken for the moment. When originally researching Facebook I did find some suggestions that they would have problems scaling. Logged out soon after so I could get some work done.

Article by professional photographer Rob Galbraith pointing to the good quality of first MacBook Pro LCD backlit matte display from mid 2007, with results of colorimeter tests. This was especially by comparison with earlier CCFL backlit display MacBook Pro.

A look at the evolving laptop display updates four laptops, including the late 2008 15 inch MacBook Pro. It gets good marks in a lot of areas. However it is marred by an omnipresent glare from its mirror-like front.

The top five laptops for displaying color gamut does not include any from Apple. Not cheap models, and they are all 17 inch or larger.

You get away with lousy web pages because browser error handling is so very good.

Ever since HTML2, the HTML markup language has been declared as being an SGML derived product. This means the Doctype points to a machine readable DTD which actually specifies exactly how you need to parse the HTML. That is technically what a validator does. You could write a browser that did the same, except most pages would break. So all the browsers ignore the DTD, in favour of using their own internal representation of how a web page is constructed.

In Internet Explorer 5 for Mac, Tantek Celik first started using Doctype sniffing to decide whether a web site was an old site (and probably broken). A Strict Doctype would mean a new site that should be written to W3C specifications. IE6 for Windows did the same. In particular, a Strict Doctype was intended to mean it followed the standards, and a Transitional Doctype indicated an existing web page that was not converted to Standard.

Unfortunately, people doing web sites started putting Doctypes in them, mostly because other people were putting Doctypes in them. And people writing web editors started adding Doctypes automatically.

Since browser error handling is so good, and so standard, why not just take advantage of it?

There are W3C specifications for how to write web pages, and these have the authority of a standard (according to the RFC that passed control to W3C). All the major browser makers (including Microsoft, which were first to use CSS, for example) are members of W3C. All have agreed to make their web browsers use W3C specifications. It is pretty hard to do this right.

For handling errors, each web browser tries to imitate what previous versions of browsers (mostly Netscape Navigator and IE3 through 5) did. Then IE6 tried to imitate IE5. Then IE7 tried to imitate IE6. Then IE8 got a special little switch that would let it imitate IE7 imitating IE6, and so on. And every other browser tried to imitate most of this.

Writing each browser to handle all the errors exactly as they were handled by previous browsers is proving an impossible task. However if this is not done, then most web pages break. Plus the better browsers got at handling errors, the less attention was paid to making web pages correct.

The Internet Explorer Layout Engine Sucks according to these demonstrations by CSS website developer Ruthsarian. It comes down to triggering Has Layout on one of multiple elements.

Ruthsarian mostly concentrates on some very nice example of CSS layouts. These can be downloaded and freely used. Some do use Javascript patches to get things working on old versions of Internet Explorer.

I had forgotten about many of the excellent examples of HTML and CSS out there for use by web page writers. I had seen many of these previously. I was reminded of them by helpful folks on Usenet.

The rain gauge on our balcony at the Whitsunday Terraces had 35 mm since yesterday. Constant cloud cover has dropped temperatures in the tropics rather nicely. Pity about the humidity.

I phoned Carlyle Gardens to talk to the sales folks. The reason they had not contacted us a week or so back was that the house builders were not in fact ready for inspection and hand over mid January. It seemed likely they are not exactly ready even at the moment.

We arranged to phone again at the end of February. We plan a trip to Carlyle Gardens in the first week of March. I hope to be able to take a car full of stuff from our shelves during each trip. My room will be a real pain to pack without getting rid of some of the full boxes first.

Whitsunday Terraces body corporate meeting this afternoon. Looked as if we would not get a quorum for a while. Once we did the meeting went through quickly. I did not nominate for the committee. I will be leaving here soon, and travelling a lot. Not sure I could contribute under those circumstances.

Marlene phoned to suggest I attend the committee dinner this evening. It was entertaining as usual. Some people seemed impressed by the capabilities of an iPhone, especially weather radar. I had an enjoyable time at dinner, but attending means I am even further behind in doing other things.

Jean's Taming OpenOffice.org website has books and PDF available about using OpenOffice for word processing and similar office tasks, instead of paying for Microsoft Office.

If you have an older Macintosh with the PowerPC CPU, you may have problems. The official OpenOffice.org download page correctly detects that you are using an older Macintosh. It even offers to download OOo 3. However the actual official OOo download page for Linux and Mac only offers OOo 2.4 for Mac PPc. This page explains why OOo PPc is at 2.4. Most important, it points to where to get the actual OOo PPc 3 version. Same as the Mac Intel version, just has not been tested by QA.

Scharffen Berger, Schmidt chocolate plants to be closed by Hershey, reports this San Francisco Chronicle story Spike emailed to me. I had the pleasure of a conducted tour of the Scharffen Berger chocolate factory several years ago (Tuesday 24 February 2004), in the company of local fans. It was obvious the staff loved making fine chocolate. I had not realised that Hershey had taken over in 2005. My encounters with Hershey chocolates had generally not been satisfactory. I note however that large chocolate manufacturers here now often also have a quality variety or associated brand.

Strong winds reduced the stall holder numbers at the markets to less than a dozen. Then in the afternoon it started raining, with a continuous drizzle sufficient to discourage going outside for another walk.

The iPhone has a tiny screen (3.5 inch, 480x320 or 153,600 pixels at 163 ppi) compared to a decent desktop computer (24 inch, 1920x1200 or 2,304,000 pixels at around 90 ppi). The pixel count ratio may be 15 to 1. You simply do not have much screen real estate.

Safari browser on iPhone's browser does not support :hover pseudo class or mouseover effects. This makes an enormous difference to how web sites work. Links and tooltips are not indicated by a change in state. Except for underlines (not always appropriate or even possible), it is harder to indicate a link is available.

Your finger is a far fatter pointing device than a mouse cursor. Closely grouped links (such as a list of dates) are impossible to navigate.

Horizontal zoom is unacceptable. However iPhone can not handle fluid layouts at all, except when they first load. If the text size is unreadable, and you magnify the entire display to read the text, you may have to scroll horizontally. iPhone can not handle text zoom, although this is the traditional way to handle tiny fonts, especially in fluid layouts.

iPhone can not handle web pages elements located in a fixed position. For example, menus or logos stuck to the top or bottom of the display.

The iPhone has no proprietary browser plug-ins like Adobe's Flash or Microsoft's Silverlight. Programming languages like Java are also not available. I think it is unlikely that Apple will permit any of these in the iPhone. Given Adobe's Flash runs like a dog and chews all available CPU on a Macintosh with a PowerPC chip, I am not surprised Apple do not want it anywhere near their phone.

Apple hid their iPhone website guide somewhere behimd their developer wall. However Evotech writing on Website Development for the iPhone have Targeting the iPhone Safari browser. Seems a handy guide to what to watch for if iPhone is one of your target web browsers.

Which characters can you display? Alan Wood's Unicode resources web site covers Unicode and Multilingual Support in HTML, Fonts, Web Browsers and Other Applications. It is a fine resource.

Although there are not a lot of fonts in an iPhone, there are resources to make the most of them. The Elements of Typographic Style Applied to the Web covers a heap of typographic background relevant to web pages. For instance, I am using Georgia serif font for my dates, which naturally include a lot of numerals. The Georgia font has numerals as text figures, or lowercase figures. Georgia is the only common web font I have heard of that has text figures. If you look at the numerals in this blog you will see ascenders and descenders in the numerals. The Microsoft C series fonts are also reputed to be very good.