Danny Ayers has been working on turning XSPF (the XML playlist format) in to RDF/XML. 
Meg Pickard: "So I'm at FoWA London 2007, and so farŠmeh." Looks like another moderately dull conference, albeit without any politicians... yet. ";->"
Tantek has some good hypotheses on human interface design.
I'm glad that Danny Ayers is enjoying my recent output. He disagrees with me when I say "less 900 page specifications and less waffly academics". I agree with what he says though - the problem isn't the wafly academics or the large specificatons, but that both need better translation for everyone. We need to soften the learning curve. We also need less acronyms and more real names. Nobody calls their child "GRDDL" or "
One of the neat functions about the SemWeb is that you can quite easily recycle what has gone before. For the BibTeX project, I have found two ontologies which do the job reasonably well - a DAML+OIL ontology from 2001 and an OWL ontology from slightly more recently. The OWL one is slightly odd but very well-specified. Most of the work is done for me. I've defined some class names and now I just need to write up how we use the eRDF and non-eRDF versions, and put together a GRDDL stylesheet.
Announcing GetSemantic 
Wow, it's been an absolute mad panic of announcements. Firstly, "macroformats" is dead. It lasted all of a few days, but realism set in - assisted by some pissed off microformateers - and we ditched the name.
We've still got the domain names, but they will redirect and we aren't going to advertise them.
I'm just waiting for the Internet to catch up - specifically, DNS. Once the DNS machine has figured out what it's doing, then we can proceed to building the site.
I actually bought the licence for Snapz Pro X ($69!) because I feel that screencasts are going to be very important in what we are doing. Screencasts certainly helped with things like the Ruby on Rails project.
The plan is to help people understand the process of coming up with their own formats - which can be as simple as writing up a bunch of class names or as complex as coming up with a 3,000 item ontology. Of course, if they only want to do the first one, there'll be people who know how to do all the other steps and will do it for them.
I've sent out a sort of 'vision' statement to the people on the list, but I won't bore you with it here - my blog isn't the best place for it, after all. Once the site launches, something very much like it will be up there.
The first GetSemantic project I'm going to be pushing for is Embedded BibTeX. I use BibTeX a lot. The "citation" work at microformats.org is suffering because there's no clear cowpath to be paved. But we have a BibTeX ontology written in DAML+OIL and it wouldn't be too hard to use eRDF to turn that in to HTML. I'm already writing academic essays in XHTML with CSS and having the tools to embed and extract those citations would rule.
The other thing that I might do is "hRSS". hAtom is a great format, but not all web sites can be turned in to Atom - RSS 2.0 serves sites like mine better. I'll follow hAtom as closely as possible, but then move away when the RSS 2.0 specification differs from the Atom specification. Before I get flames, there are good reasons to choose RSS 2.0 if you have untitled blog entries. And, yes, there are good reasons for that too. You may not like the reasons, but they exist.
One of the key differences between GetSemantic and the more formalised microformats is that we're going to say "yes" more often. Think of them as science experiments - have fun, build something, see whether it works. We'll start herding cows down new paths and then if that works, then it might become a microformat. If it doesn't work, then we will learn why it doesn't work and try not to make that mistake in the future.
Tags: getsemantic
Why Code Assist is necessary for web developers
I really love the fact that in Eclipse, many of the plugins have Code Assist - whereby one gets autocomplete on function, variable, class and instance names as well as information on what arguments functions take.
The PHP extension has this, and it's extremely useful. Who can quite remember what order the arguments for array_multisort() go in? Having that information flash up as you type is helpful.
oXygen XML is a great example of this - if you provide a well-fleshed-out schema for an XML file, it'll make sure that the elements and attributes you insert validate. No putting div in head! It not only keeps your code wel-formed, but keeps it validating as you type it. This has helped me dramatically reduce the number of validation errors I have.
Finally, there's Aptana which does the same for JavaScript and CSS. (It installs as a separate application, but you can use it with Eclipse as a plugin).
While for a lot of non-web developers, they can learn one language like C++ or Java and spend their life doing it, the skills of web development seem to require a knowledge of hundreds of languages and tools. Which is why I'm so thankful for Code Assist. It's not that Code Assist makes you a bad developer (it doesn't). Developers are there to solve problems, not be walking, talking reference manuals.
JavaScript is especially annoying in this regard. Firebug goes some way to making JavaScript easier to learn, but I think that Eclipse and Code Assist will help even more.
If, like me, you aren't wild about JavaScript, take a look at this. It's really rather nice.
