W3C home > Mailing lists > Public > public-html@w3.org > January 2010

Re: The harm that can come if the W3C supports publication of competing specs

From: Shelley Powers <shelley.just@gmail.com>
Date: Sat, 16 Jan 2010 22:00:40 -0600
Message-ID: <643cc0271001162000y29acce8dx4fff6d16f69aea8d@mail.gmail.com>
To: Aryeh Gregor <Simetrical+w3c@gmail.com>
Cc: Boris Zbarsky <bzbarsky@mit.edu>, HTMLWG WG <public-html@w3.org>
On Sat, Jan 16, 2010 at 9:21 PM, Aryeh Gregor <Simetrical+w3c@gmail.com> wrote:
> On Sat, Jan 16, 2010 at 11:06 AM, Shelley Powers <shelley.just@gmail.com> wrote:
>> I do not work for an implementor, which seems to imbue one with super
>> human markup skills, so you'll have to excuse me if what I perceive
>> support for competing standards by the same standards organization
>> seems to me to be, well, forgive my frankness, the dumbest thing the
>> W3C has done since blink.
>
> I wouldn't go that far.  But as I've said in the past, I agree that
> working on two standards that do almost exactly the same thing is a
> bad idea.  Since no one can agree on which to actually support,
> though, there seems to be no choice but to let them fight it out.
>
>> I suppose one could say that HTML and XHTML were competing, but both
>> were really a different serialization of the same model. I suppose one
>> could say the same about Microdata and RDFa, but they are not the
>> same, they both don't support the same models. More importantly, there
>> is a major difference in impetus for both groups: one wants to support
>> metadata, the other seemingly wants to kill the first.
>
> Both want to support metadata.  I think MediaWiki should support
> metadata, for instance, and believe it should be in microdata format.
> I could just as easily have argued that we shouldn't support any kind
> of metadata, but I didn't, since I don't believe that.  Likewise, if
> Ian really disagreed with the idea of metadata in HTML, he could have
> refused to accept any metadata format into the spec, instead of
> accepting one and trying to promote it over the competitor.  The same
> goes for all the implementers who've said they want to implement
> microdata but not RDFa.
>
>> No, after reading the Wikipedia tech emails, I do object to a FPWD of
>> Microdata.
>
> For what it's worth, I (the MediaWiki developer who proposed we use
> microdata) don't really care what the W3C says.  I use the WHATWG spec
> myself, recommend others use the WHATWG spec, and will continue to do
> so if it diverges significantly from the W3C spec.  I don't think W3C
> endorsement or lack thereof will figure significantly into what
> MediaWiki does -- we use plenty of standards endorsed by non-W3C
> bodies, as do all sites.

This is a W3C group. We can't control what other groups do. Won't even
try. If the WhatWG wants to go its own course, have fun, send
postcards.


>
> On Sat, Jan 16, 2010 at 2:59 PM, Shelley Powers <shelley.just@gmail.com> wrote:
>> I respect your disagreement David. In this case, though, I believe two
>> specs will result in harm. Getting people to incorporate metadata has
>> been a challenge in the first place. Now, if you have members of this
>> WG having metadata wars in the email groups for other applications,
>> people are going to decide to not support either. And probably
>> politely, or not so politely, tell us all go buzz off.
>
> Yep.  That's what I said.  That's what I'd argue for MediaWiki, too,
> if not for the fact that we have a clear-cut use for metadata-in-HTML.
>  The solution is to kill one spec or the other, but that gets us
> nowhere if we can't decide which -- and we obviously can't.
>
>> What's also bothersome in this email exchange is that there's no
>> indication in this discussion that the web is not run by Google,
>> Mozilla, Apple, and Opera alone. Seemingly forgetting that when comes
>> to metadata--heck, when it comes to anything on the web--there's a
>> host of other tools and organizations, and people, that also have a
>> stake in the result.
>
> Microsoft, Mozilla, Google, Apple, and Opera are the only browser
> implementers that matter.  So for anything that requires support in
> browsers, yes, it's their decision, not anyone else.  A spec that no
> one uses is pointless.  Microdata and RDFa don't require support in
> browsers, so those five aren't as important here, and I did indeed
> only mention them in passing.  I mostly emphasized the technical
> benefits of microdata, and why I thought it would win in the end.
>

I believe that Mozilla's, Opera's, Apple's, Microsoft's, and Google's
customers also have some say. Unless any of the companies want to go
on record for saying that they don't give a darn what Web developer
and authors want. Could make things interesting in the browser
rankings.

But you are right, metadata doesn't impact on the browsers.


> On Sat, Jan 16, 2010 at 3:03 PM, Boris Zbarsky <bzbarsky@mit.edu> wrote:
>> The main issue with XHTML2 was that it was actively incompatible with XHTML1
>> and HTML (in the sense that you couldn't sanely implement both in the same
>> UA).  Is Microdata in such a situation wrt RDFa?  I have to admit I've long
>> since tuned out most of the discussion on both, since it seemed to be going
>> nowhere.
>
> It would be entirely possible to mix microdata and RDFa in the same
> document, if you liked.  It would also be easy for a consumer to
> support both -- HTML5 specifies algorithms to convert between them.
>

Bit messy, not to mention cluttered. The real key, though, is that
there are any number of libraries in any number of languages that
support RDFa. I'm not sure there's one that supports Microdata.

Of course, the RDFa support is for a released specification, RDFa in XHTML.


>> I do think we should make it abundantly clear for ALL our FPWDs that they
>> are NOT recommendations, NOT necessarily calls for implementation and NOT
>> "standards".  This has been a major issue of confusion in the past with all
>> sorts of half-baked W3C stuff, whether it competes with other W3C stuff or
>> not.
>
> Both RDFa+HTML and microdata are ready for implementations and author
> adoption, though, right?  So isn't it incorrect to suggest that they
> aren't?
>

No, they are not. Well, if anyone wants to hack around with a changing
spec. Only one is FPWD, though, and it's not mature enough for
production.

It is irresponsible of us to encourage the use of unreleased and
unstable specifications. Even ones we like.

Shelley
Received on Sunday, 17 January 2010 04:01:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:57 GMT