Is the W3C failing us?
I’ve detected recently a note of frustration, and downright annoyance, amongst standardistas with the W3C. I receive the newsletter “W3C Weekly News” and see endless announcements from obscure working groups working on things that are bafflingly irrelevant to what I and most other web developers actually do, whilst the interminable process of proposals and working drafts on XHTML 2.0 and CSS 3 offers no hope of a final recommendation being anywhere in sight.
9 Ways to Misunderstand Web Standards is mostly what it says it is: tips for the web developer (most of which they should frankly know already) on how not to drink the Kool-Aid. But in that document, particularly “Misunderstanding #2: ‘We Need an Alternative Mobile Web on Top of the Existing Desktop Web’” and “Misunderstanding #8: ‘The Semantic Web is Just Around the Corner’”, there are problems picked out that are more the fault of the W3C than of any web developer. If anyone is advocating the use of the term “Mobile Web” it is the W3C. And the W3C’s work on the Semantic Web has so far issued in a bunch of standards that are so complicated and difficult to implement that I doubt anyone will. (Cory Doctorow’s wonderful, funny, and short, 2001 article Metacrap suggests it’s all a pipe dream anyway.) In the meantime Microformats are taking off and actually being used, a lot, and people who are more closely related to real world web development and are sick of how slow moving the W3C are are turning their attention to groups like WHATWG and WCAG Samurai (see next paragraph).
A case in point, Joe Clark’s article, To Hell with WCAG2, over on A List Apart makes for depressing reading. It’s taken the W3C’s Web Accessibility Initiiative (WAI) five years to draft the second version of the Web Content Accessibility Guidelines (WCAG), and in that time they’ve managed to come up with something almost totally incomprehensible:
In an effort to be all things to all web content, the fundamentals of WCAG 2 are nearly impossible for a working standards-compliant developer to understand.
What’s worse is that the document apparently contradicts widely accepted best practices of accessibility that standardistas have developed over the years. In fact it also contradicts some of the W3C’s other standards!
July 13th, 2006 at 11:22 am
This Joe Clark stuff looks useful, We are just about to ‘accessiblise’ our software (web sites and desktop apps). I. time for dinner.
July 13th, 2006 at 1:59 pm
Joe Clark is to web accessibility what Jakob Nielsen is to web usability. Clark\’s book Building Accessible Websites is something of a classic.
July 19th, 2006 at 2:41 am
[…] Hot on the heels of my recent “Is the W3C failing us?” post, or, more accurately, coincedentally within a similar time frame, Slashdot has an article voicing similar concerns: Problems at the W3C. Particulary disturbing is Björn Höhrmann’s email. That’s quite a littany of failures. […]
July 19th, 2006 at 3:26 am
As Jeffrey Zeldman linked to this post, I feel I ought to give at least a little love to the W3C: so, a big “yay!” for: HTML 4.01, XHTML, CSS, XML, XSLT, XPath, and SVG. These standards have been enthusiastically embraced by developers like myself and with good reason: they fill a need, and they’re reasonably clear and simple to deploy. So please, W3C, get a grip on why some things get embraced and why others wallow for years in committees that no one pays any attention to. Without us developers your members ain’t got nothing!
July 19th, 2006 at 3:42 pm
Björn Höhrmann’s email
July 27th, 2006 at 9:22 am
Darren,
Actually I’d be a bit more cautious about the “obscure committes discussing things of no use for us laymen” (I’m paraphrasing you for the sake of explaining what I understand of your point, sorry if I’m wrong).
I can understand your feeling, and Karl did a little survey a few months ago about a list of W3C projects and most of us french-speaking people could only recognize a few (namely, the ones you list in your ‘big yay’ comment).
Nonetheless, when the W3C first worked on CSS, Xpath, etc, I’m sure most of the real world have thought “what the heck, my job’s all about HTML”. And look where we are now.
I’m really looking forward to Karl’s speech at Paris Web 2006, on that account. (disclaimer: and I’m not saying that as a co-organizer but as a member of the audience)
July 27th, 2006 at 3:07 pm
Hi Stephane, thanks for the comment. Paris Web 2006 looks interesting, or at least I think it does—I believe the technical term for my proficency level in the French language is “rubbish”.
I’m not sure that I quite agree with your comparison of current W3C endeavours with the early days of CSS, XPath, etc. of yore. I came to web development from an academic background, specifically in analytical philosophy and cognitive science, and I was familiar with markup languages before I’d really even given HTML a go. When I did I was rather surprised to find that there was something which called itself a markup language which had bold and italic elements, color attributes and whose table element was routinely used for layout. CSS came as no surprise to me, and I think there are many “standardistas” who feel the same way. It was exactly what we were crying out for. XPath, XPointer and XLink were extremely welcome additions to the rather weak equivalents in HTML. XML was a brilliant move, one I did not foresee but instantly recognised the value of when it arrived. SGML was just too big, too cumbersome. XSLT was then just the icing on the cake. That is why I feel such a sense of dismay at what the W3C is doing now. How can the same organisation that saw the need for something simple and easy to use like XML be churning out some of the totally arcane material that they are now? Similarly SVG was an answer to a felt need. I can’t prove it, but many of my friends will attest to my enthusiasm for the idea from its very beginnings. (Usually it took the form of my boasting that one day I would answer the question “What graphics application do you use?” with “Emacs”. Inkscape has proved far too good to actually make that boast ever likely to be true!)
Now let’s look at some recent W3C activities:
“Mobile Stakeholders Sketch Future of Device Description Repository”
Ughhh. Back to browser sniffing? Ask any standardista what they want and the answer will be “one set of standards for everything which gracefully degrade across all devices and platforms”. Instead we are getting an official database (which will always be incomplete or inaccurate—remember, people lie, people are lazy, people are stupid) listing all mobile devices and their capabilities so we poor web developers can try and send different content to different devices. No bloody thanks.
XForms 1.1 working draft
Okay I really like the ideas of XForms. I was really excited about it. And then I started reading the specs. I started to lose the will to live. This is supposed to be easy to deploy how again?
Emotional States inubator group
This sounds like fun, and might be interesting. But is this a priority? When XHTML 2.0, CSS 3.0 (hell, 2.1 is still a working draft) are dragging their heels?
DIAL to Improve User Experience by Adapting Web Content
The what now?
Would that be a bit like XML, CSS and Javascript? Don’t we have this already? Or is this a language to describe languages? What would I want that for?
In general, everytime I read something from the W3C about the “Mobile Web” I want to weep. There’s just the web. That’s it. It’s meant to be platform independent, device independent, ability independent. The existing standards already in use are supposed to achieve that, and if they don’t then they need to be improved. We don’t need a whole bunch of extra mobile web crap.
And the semantic web is just bizarre. I come from a philosophy and cognitive science background. It’s just a rehash of 1970s logic based AI. No one does that stuff anymore. It fdoesn’t work. All the stuff being done in academic AI and CogSci departments is statistical. So we have a rehash of old ideas in a totally undeployable form. Give me microformats.
July 27th, 2006 at 3:51 pm
Oh, as a P.S. to my previous comment, I thought I’d provide a little anecdotal evidence about what is and isn’t being done in AI and Cognitive Science these days. Edinburgh University’s School of Informatics—which is an amalgam of the old Department of Computer Science, Department of Artificial Intelligence, the Centre for Speech Technology Research, and the Centre for Cognitive Science (my old department)—no longer offer any of the courses they used to on Logic or Formal Linguistics (or so my old PhD supervisor told me). Undergraduates wishing to take a course on Logic are free to take the one offered by The School of Philosophy, Psychology and Language Sciences.
July 28th, 2006 at 12:25 pm
OK Darren,
Most of your points are good, I’ll give you that
Except maybe for semantic web. I’m OK with the notion that a fully-semantic HTML-based content isn’t possible (if only we could just define what fully-semantic is though). Yet, one can’t deny that a list is a list and should be marked up as such using appropriate tags (
ul,ol,li,dl,dt,dd), a structured (i.e. built up withh*headers and sections in mind) page is more understandable either by a machine or by a human (for instance I’m using very often the Document Map Firefox extension on long pages to get a clear view of the document’s outline), acitemeans a reference of some kind, ablockquotea citation, etc. So maybe it’s the terminology that is a bit disappointing and should be renamed “semantics-oriented web” or something.Yet even that notion lacks a lot of what we’ve come to think of as natural in the book world. Endnotes, anyone? Footnotes? Legends on images (the equivalent of
caption)?All of a sudden I wonder if I’m still on-topic, though
Back to your last note:
Do you mean RDF or FOAF are useless?
July 29th, 2006 at 10:36 am
Thanks for the reply, Stephane. We actually agree on many things. To try and clarify that, let me distinguish between the semantic web (lowercase) and The Semantic Web (capitalised). The lowercase-semantic-web is something I am an advocate of; it involves using markup as it’s supposed to be used—for capturing logical, structural information. For example, see my Image Captions in XHTML (some of the style example won’t show up because a server crash ate my Phlogiston theme). Your examples of proper markup use (
citevsem,dfnvsstrong, etc.) also fall under the umbrella of the semantic web, as do microformats. Note that all the reviews on my blog use microformats, which also contain a link to an XSL tranformation to convert the microformat into RDF. I certainly don’t think RDF is useless, I think it’s fantastic, as it allows us to specify relationships between resources in ways which break out of the hierarchical structure of an XML document. The same goes for FOAF and other such formats. As a standardista, that kind of “semantic web” is precisely the kind of thing I’m after. But it is important to note that the best we’ll ever achieve is to have beautiful strutural markup in our templates and for specific kinds of blocks of data like addresses, events, reviews, etc. Once you have long chunks of narrative data put into a content management system by civilians you won’t be able to achieve this. Do you really think that some junior in the marketing deparment of a company is going to carefully markup acronyms and abbreviations, to distinguish betweenciteandem? No, they just want a WYSIWYG interface with an italic button.The Capitalised-Semantic-Web is much more than this. This is from the W3C’s page on the topic:
The idea being that for example, we have some piece of RDF somewhere saying
author-of(Darren Brierton, Phlogiston)and another bit of RDF somewhere completely different sayingresident-of(Darren Brierton, Edinburgh), and from that a computer can deduce the new piece of information that the author of Phlogiston lives in Edinburgh. But for this to actually work all of the terms of those relations have to be trusted authoritative URIs. A URI for Phlogiston is easy. For Edinburgh? Maybe doable. For Darren Brierton? Well Libraries have name authority databases, but they are hard to maintain, hard to cross-link with other libraries name authorities, and only deal with the much smaller number of historically important people. And what about the URIs for “author-of” and “resident-of”? Where do these come from? Well presumably these come from some kind of OWL ontology. So in order for this to work, we have to have some kind of authoritative, trusted repository analysing every single relation that objects can be in! Good luck with that!I know I linked to this in my original post, but it is worth linking to again: Cory Doctorow’s Metacrap is a fabulous account of why this stuff won’t work. There are practical difficulties (his first three points in particular: people lie, people are lazy, people are stupid) and there are theoretical diffiiculties.
So when I moan about The Semantic Web, please note that it is the capitalised W3C vision of that that I am moaning about. As far as the lowercase semantic web goes, I am a huge advocate and supporter.