Home › web › rdf

RDFa, sure, but now?

Sun, 2007-08-19 20:34 — Jeni

I think RDFa has a lot of things going for it. I’ve been a big Semantic Web sceptic for a long time, but getting semantic information embedded into web pages is cool, and RDFa is that much more rigorous than microformats. Sure, it’s a bit more complicated too, but I’m not afraid of namespaces!

The problem is that I’m supposed to be assessing the introduction of RDFa into a biggish, important, real-world website. It’s a website where every change has to go through a Process. There are project managers, development managers, product managers. There are functional specifications, technical specifications, and rollout documents. There are unit tests, peer reviews and user acceptance tests. The database is huge; the HTML is pre-generated.

In short, every change takes time and costs money. I’m just not sure it’s a good idea to introduce a technology that only exists in a Working Draft of a Primer now, when tracking that technology could be very costly.

And if we used RDFa, we wouldn’t be using XHTML 1.0 any more; we’d be using XHTML 2.0. If we want to have valid pages, we’d have to reference a non-existent DTD, or make our own. And tracking XHTML 2.0 is even more of a risk, it seems to me.

So my current thinking is that in this case, RDFa just isn’t practical. Better to use GRDDL with something less revolutionary, like eRDF or microformats or plain-old custom class names.

Anyone want to persuade me otherwise?

Comments

Mon, 2007-08-20 13:49 — Jim Fuller (not verified)

Re: RDFa, sure, but now?

Before you decide on meta data instantiation; ask yourself ‘whom is going to consume this data’….if u find that in your internal Process there are going to be internal consumers of this data, try to identify integration routes and correlate which approach would be best. Same thing goes for external consumers of metadata.

The next issue, is what will happen when this meta data is let out ‘into the wild’ e.g. with respect to search engines, other applications gleaning information from it….it will be relatively easy to reconfigure your meta data format (e.g. RDFa or GRDDL), the problem will be what to do with cached assumptions. I would propose you plan for multiple meta data formats.

gl, Jim Fuller

Mon, 2007-08-20 08:32 — Danny (not verified)

Re: RDFa, sure, but now?

My impression is that RDFa is close enough to done to be relatively low risk, although personally I’d like to see a little more documentation of the recognition/interpretation mechanism(s). If I wanted to use RDFa right now I’d go belts and braces and reference a GRDDL profile/transformation (there’s an RDFa2RDFXML.xsl linked from the GRDDL Primer) so even if there were radical changes with RDFa, the document would still be interpretable as RDF.

eRDF (with GRDDL) seems to be stable - at this point in time it’d be my own first choice where flexibility of the data embedded is a requirement.

It depends quite a lot on what’s going to be embedded. If it’s going to be the same kind of stuff throughout a lot of documents, using microformats and/or plain-old custom class names (like a locally-defined microformat) would probably be a good choice, though someone would have to write the GRDDL XSLT (I wonder who :-)

Multiple profiles can be applied to a doc with GRDDL, and generally the different approaches can be used together without side effects.

(heh, the captcha I just got seems appropriate “depth choice”)

Mon, 2007-08-20 08:57 — Michael Hausenblas (not verified)

Re: RDFa, sure, but now?

+1 to Danny. Definitely go for a profile.

eRDF/RDFa really depends on personal taste (Danny@Talis naturally opts for eRDF, myself as the ‘quirky- W3C-guy’ for RDFa - but hey that’s life).

Just one thing: Go for it. Start playing around with XSLTs, templates, SPARQL, extractors, etc. and enjoy the wi(l)de world of the Semantic Web ;)

Cheers, Michael

Mon, 2007-08-20 07:57 — Keith Alexander (not verified)

Re: RDFa, sure, but now?

The big thing that prevents me from using RDFa now is that the syntax is still being developed/finalised - so you might have to go back and change things if you try to use it right now (though the RDFa folks seem to be working hard to get something stable out the door soonish).

It would be interesting to know what ends your project hopes to achieve with embedded semantics. Sometimes an external document makes more sense.

Mon, 2007-08-20 08:23 — Jeni

Re: RDFa, sure, but now?

It is the cause of much amusement to me that the aim of this project is no more and no less than “to put RDFa on our web pages” rather than being something functional like “we want our web pages to be more easily searchable”. I don’t think the customer has a really good reason for using RDF or RDFa in terms of functionality, just a kind of “if we build it they will come” mind-set.

Which is not to say that it’s a bad idea: if you’re publishing a large amount of information that might potentially be useful to someone, and that information is auto-generated anyway, then why not embed semantics in it? We’d never get any semantics on the web if everyone waited until there was an immediate benefit to themselves.

Tue, 2007-08-21 12:58 — Keith Alexander (not verified)

Re: RDFa, sure, but now?

There is a danger though, if you don’t have any particular use for the embedded metadata, that you might not notice (especially if the syntax is unstable) if the markup doesn’t reflect the meaning you intend.

Mon, 2007-08-20 07:50 — Jeni

Re: RDFa, sure, but now?

Thanks to Tom, Got and Michael for the links and the reassurances about use with XHTML 1.1. That answers that drawback.

But what about maturity? Isn’t it a bit premature to base implementations on Editors Drafts of Working Drafts? Or is RDFa basically sound, and the whole W3C standardisation process likely to only last a short time with few changes? (Having been involved in a couple of W3C standards, I’m well aware that languages can be pretty much stable before they reach Recommendation, but equally that parts only get properly worked out at the final stages. And that getting to Recommendation always takes longer than you think.)

Or perhaps I should ask the question this way: what bits of RDFa are stable, and what bits are likely to change?

Mon, 2007-08-20 08:03 — Michael Hausenblas (not verified)

Re: RDFa, sure, but now?

Regarding your question: “what bits of RDFa are stable, and what bits are likely to change?”

All the base stuff (@href,@rel,@property, etc.) is stable. Some syntactic sugar (as @instanceof, @src, etc.) is currently being fixed. We plan to go rec track ASAP (target: REC for primer & syntax in the next months, cf. also [1] - RDFa Schedule). You can safely start implementing stuff, others do as well ;)

Just to give you an idea about the progress: in the last telecon [2] we resolved quite a lot …

Cheers, Michael

[1] http://www.w3.org/2006/07/SWD/wiki/RDFa

[2] http://www.w3.org/2007/08/16-rdfa-minutes.html

Mon, 2007-08-20 06:01 — Michael Hausenblas (not verified)

Re: RDFa, sure, but now?

Jeni,

As I really like what you do here (and as a matter of coincidence being responsible for the RDFa Test Cases) I wonder who told you that RDFa is XHTML 2.0-only …

Actually, the current spec [1] targets at XHTML 1, along with the Test Cases [2]. You may also wish to check the availability for other members of the HTML family [3] …

Please, in case you need any support, let me know :)

Cheers, Michael

[1] http://www.w3.org/2006/07/SWD/RDFa/syntax/ [2] http://www.w3.org/2006/07/SWD/RDFa/testsuite/ [3] http://www.w3.org/2006/07/SWD/wiki/RDFaInHTML

Mon, 2007-08-20 05:57 — Got (not verified)

Re: RDFa, sure, but now?

To complete the previous answer, the URL of the DTD XHTML 1.1 + RDFa is http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd and the doctype is <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN" "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">. Moreover, the new version of W3C validator offers a support of this Doctype (cf http://rdfa.info/2007/08/07/w3c-htmlrdfa-validator/ ). So don’t be worried, just have fun !

Mon, 2007-08-20 11:22 — Jeni

Re: RDFa, sure, but now?

I just tried this, and the big problem is that RDFa uses namespace declarations to interpret CURIEs. There’s no possible way for the XHTML+RDFa DTD to contain attribute declarations for every single possible namespace you might want to refer to from RDFa, so users either have to create a specialised DTD for their particular application, including the relevant namespace declarations, or add an ATTLIST at the top of every HTML page.

As I’ve argued before CURIEs really shouldn’t be reusing XML namespaces anyway, because they use namespaces in a completely different way from XML. It’d be far better for RDFa to extend the <base> element to cover providing the base URIs for CURIEs. That would solve the undeclared-namespace-attributes-in-DTD problem too.

Sun, 2007-08-19 22:14 — Tom Morris (not verified)

Re: RDFa, sure, but now?

There’s an XHTML 1.1-based RDFa DTD published by the W3C which uses the modular basis of XHTML 1.1 to add RDFa support.

Tue, 2007-08-21 13:55 — Mark Birbeck (not verified)

Re: RDFa, sure, but now?

That’s right…and in addition to the DTDs there are also XML Schemas that conform to the XHTML Modularisation architecture.

In terms of the main points—is RDFa stable enough?—I think for most uses that RDFa would be put to, it is certainly stable enough, and has been so for quite a while. I hold my hands up and say that we haven’t been very good at keeping people up-to-date with developments as they happen, but we’re closing in on final drafts now, and the kinds of things that might still change are the sorts of things that would rarely be used anyway.

It would be a shame if Jeni didn’t make use of RDFa, purely because she wasn’t sure how things were progressing—if it wasn’t suitable that would be fair enough—so please do keep me posted of your progress, and if there’s anything I can do to help your decision-making, I will.

All the best,

Mark

— Mark Birbeck, formsPlayer

mark.birbeck@formsPlayer.com | +44 (0) 20 7689 9232 http://www.formsPlayer.com | http://internet-apps.blogspot.com

standards. innovation.