Phlogiston

Monday saw the final stage of the ratification of OpenDocument 1.0, the Open Document Format for Office Applications by OASIS, the international e-business standards consortium.

OpenDocument provides a royalty-free, XML-based file format that covers features required by text, spreadsheets, charts, and graphical documents.

“XML doesn’t always mean open. You can hide a lot in a file format. OpenDocument represents an opportunity to ensure truly open file formats for productivity applications, which is why it will receive the enthusiastic support of public sector steering organizations on a global basis,” commented James Governor, principal analyst at RedMonk. “The participation of enterprises in vertical industries, such as aerospace, will also ensure adoption in the private sector. One key to success will be the royalty free status of the spec; there are no financial penalties associated with developing to it.”

“Office productivity applications and the documents they create are key to today’s knowledge economy. Information critical to the long term functioning of any organization is stored in the spreadsheets, presentations, and text documents its employees create,” said Michael Brauer of Sun Microsystems, chair of the OASIS OpenDocument Technical Committee. “Today, for the first time in the 25-year history of office applications, such documents can be stored in an open, standardized, and vendor-independent format.”

OpenDocument provides a single XML schema for text, spreadsheets, charts, and graphical documents. It makes use of existing standards, such as HTML, SVG, XSL, SMIL, XLink, XForms, MathML, and the Dublin Core, wherever possible. OpenDocument has been designed as a package concept, enabling it to be used as a default file format for office applications with no increase in file size or loss of data integrity.

“OpenDocument is a fine example of an OASIS Standard that originated in and continues to be endorsed by the open source community,” noted Patrick Gannon, president and CEO of OASIS.

Erwin Tenhumberg’s blog has further details.

ComputerWeekly is reporting further developments: Bristol ready to put open source on 3,500 desktops. The report not only mentions Bristol City councils rollout of StarOffice (a version of OpenOffice.org) across 3,500 desktops but also Birmingham City Councils announcement of a 1,500 seat pilot of a completely open source desktop based on Linux.

Well, maybe not famous …

But Ivan pointed out that I am mentioned in the new, third edition of O’Reilly’s Learning GNU Emacs:

“Serious web developers may want to investigate some of the cutting edge development going on to make Emacs even more powerful. Check out HTMLModeDeluxe and the Emacs WebDev environment by Darren Brierton. […] Both are excellent tools for building complex web pages.” (p. 220)

Okay, now who wants to touch me? I said who wants to touch me?

As reported in FT.com , The Register, and ZDNet, the EU are seeking legal advice on how to re-draft the controversial software patents directive so that patents on pure software can be excluded. This is a very, very good thing.

I blogged earlier about the problems raised for the FOSS community by the fact that a slew of new features in OpenOffice.org had been implemented in Java.

This NewsForge article, Free Software Foundation and OpenOffice.org team up to escape Java trap, suggests that a resolution of the problem might be on the horizon. The article is worth reading both for its objectivity and balance, and because it explains the political background of the dispute and some of the differences between different factions within the FOSS community. Thankfully, it would appear that despite the differences, the community has the ability to heal these kinds of rifts in mostly the right way in the end, largely in the same chaotic, anarchic ways it gets everything else done.

I blogged previously about the Acid2 challenge, and it seems that Safari is the first browser to pass it. (Well, strictly speaking, the next version of Safari will be.) Kudos to Apple and David Hyatt, the Safari lead developer.

With Opera dominating the handheld and mobile sector, and Firefox gaining in usage on the PC/Windows platform, this bit of news completes a very encouraging view of the browser space. Now if those of you still using Internet Explorer would get with the program …

Great news in The Register: Firefox doubles market share as IE slips.

Janco Associates found IE had 83.7 per cent of the market for this month, down from 84.85 per cent, while Firefox grew from 4.23 per cent to 10.28 per cent. Janco believes Firefox could take 25 per cent market share in the next quarter.

Thanks to Kris for the pointer.

This article, Browsers Get Ready for Graphics Boost, claims that Scalable Vector Graphics (SVG), a W3C specification for XML based vector graphics (i.e. graphics composed out of vectors and arcs instead of pixels), will at long last be getting mainstream browser support.

Opera, who lead the market in browsers for handheld devices, plan to have native support in the forthcoming Opera 8. SVG is particularly significant in the handheld sector, where pixel based graphics formats such as JPEG and GIF are often unsuitable. Handheld devices have both limited bandwidth and small screens. A JPEG image which might appear quite small on a standard computer screen might be too large to fit completely on a handheld screen, and will often be a fairly hefty download, which becomes particularly significant if you are also paying for every single byte as many handheld users are. Vector graphics are scalable, and so can be rescaled to fit appropriately on small screens, and because they consist merely of statements like “draw a red circle with radius r and centre ⟨x,y⟩” they are much smaller than their pixmap equivalents. SVG is so important to the mobile sector that “the 3GPP (3rd Generation Partnership Project), a GSM network technical standards body, has begun mandating SVG support for next-generation mobile devices” and the W3C has tailored a specific “mobile profile” of the SVG specification.

Firefox will apparently also have SVG native support enabled with release 1.1. This is very cool indeed, because SVG isn’t just important for the mobile/handheld sector, but is a very exciting prospect for the web in general. Because SVG is just another XML application, it can be freely mixed and matched with XHTML using namespaces. By embedding XHTML in SVG and vice versa many things become possible: text inside graphics can be fully accessible to search engines; a whole document of mixed XHTML/SVG content can be scripted through the DOM and styled with CSS, so, for example, a particular part of a graph could glow or change colour when your mouse hovered over a piece of text referring to it, or the graph could change dynamically as you typed various “what-if?” values into a form.

Currently the only real way to do many of the things SVG is intended for is to use Flash. Now Flash has its place, and is certainly capable of many things which SVG isn’t and never will be. But Flash is also wholly unsuitable for the kinds of things it has been pressed into use for in order to fill the gap left by the lack of a viable alternative. SVG could be that viable alternative.

It is almost certain that Microsoft will not be providing support for SVG in the forthcoming Internet Explorer 7, just as they will not be supporting CSS 2 or 2.1. They simply don’t have time now given their intended release date for IE7 (some time this summer). But it is also unlikely that they will ever support those standards, given that Avalon actually includes competing, non-open, non-standard vector graphics and stylesheet languages. (Despite the fact that Microsoft are W3C members and helped to draft the SVG and CSS standards.)

Apple’s browser Safari is actually based on the KHTML rendering engine used in Konqueror, the KDE browser. KDE, like its competing GNU/Linux desktop environment GNOME, already makes heavy use of SVG, and it is surely only a matter of time before KHTML also has native support for SVG, and therefore only a matter of time before Safari does too.

In that case the situation will become really interesting: on one hand we will have native support for SVG in all mainstream browsers except IE, and a demand for SVG content driven by the mobile sector; on the other hand the dominant browser, IE, will not support such content and Microsoft will hold out for Avalon to hit the streets (which even if they release it before Longhorn is unlikely to be before 2006). Either Microsoft will leverage its market dominance to kill SVG in its infancy by refusing to support it, or users, already frustrated by IE’s security flaws and other problems, will switch to standards compliant browsers at an even greater rate than they already are, thereby undermining Avalon’s future.

Like many FOSS advocates, I have tried to encourage the uptake of FOSS software by suggesting users switch to free, better alternatives on an application by application basis: don’t use Internet Explorer, use Firefox instead; use Thunderbird instead of Outlook or Outlook Express; use The GIMP instead of Photoshop. Perhaps most important in this is encouraging users to move away from Microsoft Office and its proprietary file formats and to use OpenOffice.org. This is particularly relevant now that the forthcoming OpenOffice.org 2.0 will use the OASIS OpenDocument XML format as the default file format:

The OASIS OpenDocument format is a vendor and implementation independent file format, and thus guarantees freedom and independence […] The OASIS OpenDocument file format is […] one of the file formats recommended by the European Commision.

This is pretty significant. Finally users have an option to use a full-featured office productivity suite without having their data locked in to a vendor’s closed proprietary file format.

Sadly, as this NewsForge article, Java fallout: OpenOffice.org 2.0 and the FOSS community, points out, the OpenOffice.org developers have shot themselves in the foot by choosing to implement a raft of new features in a way which is unpalatable (at best—at worst unacceptable) to the FOSS community, and makes it harder for ordinary end users to switch to OpenOffice.org.

Although the substantial new parts of OpenOffice.org written in Java are themselves licensed under a FOSS license, most users will need a JRE installed in order to use them:

• Sun’s JRE is not open source, and whilst it may be free to download it is not “free” in the sense required by most FOSS licenses. Many GNU/Linux distributions cannot and will not include it. For many platforms it is not even available. This means that those distributions will be forced to either
  • refuse to distribute OpenOffice.org altogether;
  • distribute OpenOffice.org with the components written in Java disabled;
  • expend many many hours of valuable developer time attempting to compile the Java components into native code so that they can be run without a JRE.
• For end users wishing to switch to OpenOffice.org they now not only need to download and install it themselves, which in itself can be daunting for many users, but also download and install Sun’s JRE. OpenOffice.org is already a large application with a big memory footprint; Sun’s JRE is a notorious memory hog. The consequence of the former now requiring the latter is that OpenOffice.org will simply be unusable on many older low-end machines.

This has to be one of the worst decisions the OpenOffice.org’s project leaders could have made.

Wired have a nice article on Wikipedia here: The Book Stops Here (I hate the title, because it misrepresents both the article itself and Wikipedia). For those of you who don’t know what Wikipedia is (are you people living under a rock?), in short it’s a completely free online encyclopedia with over 1.3 million articles in 75 languages, over half a million of them being in English; compare that to Britannica’s 80,000 and Encarta’s 4,500 articles in English. However, that doesn’t really capture what’s truly interesting about Wikipedia, but the article does a much better job than I could here of introducing you to the world of which I’m very much a part: Wikipedia is produced in the same way that FOSS (Free Open Source Software) is, so those of you who’ve wondered what on Earth it is I’m banging on about when I talk about Open Source might learn a thing or two about that too.

Jef Raskin, the creator of the Macintosh computer, died aged 61: In Memory of Jef Raskin.

He was truly one of the first and greatest experts in the field of Human Computer Interaction. Although I am not a Mac user myself, and despite the behaviour of all the Mac-fanboys out there who drive me up the wall, there is no question that Apple are the leaders in usability and ergonomics by several orders of magnitude. That is due in large part to Raskin’s legacy. They are always my first suggestion to anyone who asks me for advice about buying a computer, and almost all of the reasons I hear for using the PC/Windows platform are inaccurate or erroneous in some way. Whilst my personal preference is for GNU/Linux, and in time I hope that platform will surpass the Mac in functionality and usability, I have no doubt that the right choice for most people today is a Mac. In fact, I really don’t know why people (excluding fanatical gamers) still buy those ugly beige boxes that constantly crash and become infested with viruses and malware within seconds of being used.

Raskin created something truly wonderful. He will be missed greatly.

I’ve just upgraded this blog to the new release of WordPress, the server-side software that runs it. It was a bit hairy, as the upgrade didn’t go quite according to plan, and the whole blog just returned an HTTP 500 error, which means internal server error. It turns out that my .htaccess with all the mod_rewrite rules in it got borked. I just deleted it and let WordPress recreate it and all was well again.

The announcement—WordPress > Development Blog > Announcing WordPress 1.5—explains the new features. There’s lots of good stuff in there, but most notable is the fact that WordPress has matured from being purely for blogging into more of a personal content management system. I actually think it would work well for small companies and organisations as well, and I think it will be one of the things I use as a platform for those kinds of clients.

The plugins architecture is particularly powerful, and I think I will be evaluating plugins for this blog for two of the most obvious features: threaded comments (i.e. the ability to reply to a comment) and the ability to preview a comment before posting. Also I don’t understand why the nice interface I get for quickly inserting hyperlinks and the most common kinds of markup is denied the posters of comments. There’s probably a plugin for that too.

One feature you lot might be interested in: supposedly there is an options somewhere for you to be notified by email if someone comments on a particular post. I’m not sure yet where that is to be found. It may be an option when you post a comment, or it may require registration and then be made available as an option (i.e. something like “Notify me by email of comments to posts I have commented on”, or something like that). Poke around for yourselves and see what’s on offer.

It looks like all the Ximian Monkeys (well, technically, they’re Novell Monkeys now) are blogging about this: Hula, which is basically an open-source Exchange killer, drawing heavily on Gmail. Nat Friedman seems to have the most detail, including screenshots (which are very pretty thanks to Garrett LeSage’s great graphic and UI design).

As Nat goes on to say:

We will build a JavaScript-based rich client for mail and calendaring, in the style of GMail. Those of you who follow my blog may have noticed me waxing optimistic about the power of web clients over the last few months; even before maps.google.com came out and blew everyone’s mind.

I searched for weeks to find an open source project that is working on implementing an open source version of GMail. I even posted to Google answers looking for one. It doesn’t exist. Well, today we’re starting one, and we’re inviting the world’s crack JavaScript/DHTML hackers to help us.

It looks like this is something I should get involved in when I have a little more time on my hands.

Kris: isn’t this exactly what you’ve been looking for?

My friend and sometime colleague/collaborator/client, who should probably remain nameless, rang me today to discuss a potential business opportunity. I hope you’ll forgive me if I can’t remember the precise details of the conversation. At some point very early in the conversation all of my synapses fused together. “I’ve got these neighbours,” he says, “they’re really good friends of mine, really lovely people. They’re a bit ‘out there’, but in a liberating way, you know? Anyway, they’re a gay couple and they want to set up a gay porn site.”

Me: **blink**

“They’ve got a really high end DV camera, and basically they’re going to go out and find young, good-looking guys and film them jerking themselves off.”

[Do you mind if we just pause a moment here? How does this work again? Can you really just casually walk up to people and ask them if they’d mind you filming them while they crack one off? At what point do you broach the tricky issue of financial incentive? And does the average person know what the going rate for such things is? I’ve half a mind to go out and buy such a motion picture recording device myself (or at least something that looks like one) and wander across to the student hall of residence opposite my flat and see if any of the young ladies over there would mind me filming them while they smacked the pony (for purely sociological reasons, of course). I wonder what my chances of emerging without the camera wrapped around my head—or shoved up my arse—would be?]

“It’s not like it’s heterosexual porn”, he continues. “That’s just exploitative and nasty. Gay porn is completely different.”

Me: **blink**

“And anyway, everyone knows we’re straight, so it’s not like anyone could accuse us of having some lascivious or exploitative interest in this.” Now I’m really worried: that last bit actually made sense to me.

Of course, what’s really odd is that after I recovered from the initial shock the two of us just started talking about the technical details like it was any other web project. What would we use as a streaming multimedia server? Flumotion, the completely free, open-source streaming multimedia server from Fluendo (based, I believe, around the GStreamer multimedia framework) would be the obvious choice. And I know the Flumotion hackers would love to see it stress tested on a high volume site. Would they mind if it was gay porn? Maybe I should ask them.

Then there are the legal complexities. Is this kind of material legal in Britain? Neither of us knew. To be on the safe side the site should probably be hosted abroad. Holland, probably. Everything is legal in Holland. But does it matter where the material is hosted? Perhaps what counts is where the company is registered. Who knows? More importantly, who is going to be the one to find out? I don’t think I’ll be the one calling at my local Citizens Advice Bureau for some free legal advice on gay porn websites.

How do you make things like this safe from children? I’ve heard of Net Nanny and Cyber Sitter but that is the full extent of my knowledge of them. Do ordinary credit card handlers deal with sites like this? Or do you have to go to some special “adult” handler? And what about the hosting service? Presumably most hosting service aren’t interested in hosting adult sites, if for no other reason than for the vast amount of bandwidth they take up.

I can just see this going in my portfolio. One thing is for certain: I’m not sourcing the graphic designer. “Hi, it’s Darren. Yeah, I was wondering if you’d be interested in a project I’m working on. We’re kind of looking for a sort of young, muscular, handsome man having a wank look.”

Excellent! jgwebber has gone to the trouble of dissecting how Google Maps works in this article: Mapping Google.

First with Gmail, and now with Google Maps, Google prove that really rich client-side web applications are possible.

Here’s a quick XHTML tip. Image captions sometimes present a bit of a puzzle to people trying to do proper semantic markup. The result is that either completely the wrong markup gets used (a caption is not a paragraph) or absolutely no semantic markup is used and the image and caption are simply wrapped in <div>s and <span>s, the generic block and inline elements.

In fact the proper way to attach a caption to an image is to use a definition list (<dl>). A definition list is a list of pairs of elements, a definition term (<dt>) and a definition description (<dd>); the first is an item about which the second gives further information, which is exactly the relation between an image and its caption. The “definition” in “definition list” is potentially misleading. Presumably it’s called that for want of anything better. The important thing to note is that each list item is a pair: a term followed by its description. (Actually, definition lists are a little more complicated than that. See comment 4.)

So, here’s how to markup an image and its caption:

<dl class="image">
  <dt>
    <img />
  </dt>
  <dd>
    caption
  </dd>
</dl>

All that’s needed is some CSS to give us the right visual presentation:

/* this first rule is probably unnecessary, */
/* but better safe than sorry               */
dl.image,
dl.image dt,
dl.image dd {
  display : block ;
  list-style-type : none ;
}
dl.image {
  margin : 1.2em auto ;
  border : none ;
  padding : 0 1.2em ;
  text-align : center ;
}
dl.image dt {
  margin : 0 ;
  border : none ;
  padding : 0 ;
}
dl.image dt img {
  display : block ;
  margin : 0 auto ;
  border : 1px dotted #38abc8 ;
  padding : 15px ;
  color : #38abc8 ;
  background : #eee ;
}
dl.image dd {
  margin : 0 ;
  border : none ;
  padding : 1ex 1ex 0 1ex ;
  font-size : 90% ;
  font-style : italic ;
  color : #38abc8 ;
}

and the result will look like this:

Mmmm … Scarlett Johansson, she purdy

Of course, that was just an excuse to put a picture of Scarlett Johansson on my blog.

The only semantic issue that might give one pause for thought is whether a list can contain only one item—or pair, in this case. I would suggest that any misgivings about that are misplaced. A list with no items is no list at all. But a list with only one item strikes me as a perfectly legitimate limit case of a list.

There is a nice article in The Economist on the economics of sharing:

The characteristics of information—be it software, text or even biotech research—make it an economically obvious thing to share. It is a “non-rival” good: i.e., your use of it does not interfere with my use. Better still, there are network effects: i.e., the more people who use it, the more useful it is to any individual user. Best of all, the existence of the internet means that the costs of sharing are remarkably low. The cost of distribution is negligible, and co-ordination is easy because people can easily find others with similar goals and can contribute when convenient.

Good stufff. It’s nice to see the mainstream media not only paying attention to FOSS, but actually beginning to understand it and recognise its value and advantages.

As I prepare to announce this blog properly, it occurs to me that a lot of my friends probably have no idea what a blog is. So this entry is for them. (And can I just say from the rest of us, welcome to the 21st cenury).

Years ago, after the web explosion, people started keeping a “log” of the cool and interesting sites that they’d found whilst browsing. They put this “web log” on the web itself so they could share the information and links with friends. Soon this took off and lots of people started doing it. After a while the “web log” started to evolve, as people not only “logged” the interesting sites they’d found but started to comment on them. After a while not only did people comment on what they’d been visiting, but they allowed their friends to add their own comments to their comments. After a while, web logs, by now abbreviated to “blogs”, became a peculiar mixture of online diary (in some cases very personal), micro-community (because of the blogger’s friends’ comments), and do-it-yourself journalism. Blogs sprang up that spanned the whole gamut from Samuel-Pepys-style personal reflection to political activism and/or commentary. Blogs were a major feature in last years American election.

Another striking feature of blogs is that they are at the forefront of the kinds of web technologies that I am most interested in: the semantic web and web ontology. Most blog software publishes your innermost thoughts not only as a web page but also as something called a “feed”: a feed is a compressed form of the headlines from your site. These feeds are aggregated (i.e. signed up to and published) all over the web. There are sites called “Planets” which aggregate the feeds from the blogs of people who are involved in some particular activity, be it political or social or to do with IT. These “Planets” form a virtual community.

Also, most blogs have something called a “blogroll” which is a list of other bloggers who you have something in common with: “blogrolls” themselves are also published nowadays in a semantic form, detailing how exactly you know the person in question, whether they are a friend, an acquaintance or a colleague, whether you have physically met or just know each other through the internet. These allow the automated mapping of social networks.

There is a longer, more detailed article at the excellent WikiPedia (a free, online encyclopedia).

The “blogosphere” is connected to the so-called “smart mobs” phenomenon … what do you mean you’ve never heard of smart mobs?

XML.com: Hacking Open Office is an excellent article on using XSLT to manipulate OpenOffice.org’s XML file format.

My computer died over a week ago. Last weekend I got up and discovered that my computer wouldn’t wake up from being in xscreensaver. Hmmm. X has crashed, or something, I thought. Moving the mouse and fiddling with the keyboard made no difference. So I held down the power button until the machine rebooted. It got past the BIOS screen and then just said: “Primary internal hard disk drive not found. Fixed optical drive not found.” Sure enough attempts to boot from a CD-ROM failed with the same message, but attempts to boot from a floppy worked (albeit with the same message again). So, the machine boots from a floppy, but neither the hard disk drive nor the fixed optical drive can be found. It must be the on-board IDE controller, I decided.

Thankfully everything is still under warranty (but only until June or July, I’m not sure which), but I had to wait until Monday to call Dell. The guy I spoke to in Ireland was actually very helpful and seemed to know what he was on about. He seemed to think it was a hard drive failure whilst I favoured the idea that it was a motherboard failure. As I was covered for next day on-site maintenance I suggested that an engineer come out with both a new drive and a new motherboard. To my astonishment I was told that Dell engineers are only allowed to come out with one part at a time! I feel mean complaining about that right now, as the guy (Declan, I think) was actually helpful and good at his job—he used Linux himself at home and didn’t blink when I said I did—and it turned out that he was right and I was wrong. Dell sent a new hard drive the next day, and sure enough that was the problem. However, based on my report (especially the bit about the machine refusing to boot or even recognise the CD drive) it seems to me that there was a pretty good chance that I might have been right, in which case sending out a new drive would have made no difference at all and I would have been left having to call Dell again, and waiting another day for the right part to come out.

Anyway, the bottom line is that hard drive failed. I have a farily good backup regime: I backup my home directory every night onto an external Firewire drive, and I backup /var, /etc and /usr/local as well. But there simply isn’t the backup space to backup everything. Sadly, however, everything includes a Windows XP partition, a Linux (Fedora Core 3) partition, and within the Linux partition four VMware Workstation virtual machines (all varieties of Windows with different versions of browsers installed). So all in all I had to restore the equivalent of six machines, each with radically different software on them. What a complete nightmare.