W3C home > Mailing lists > Public > public-html@w3.org > May 2012

Re: Issue 31c: Meta generator

From: Daniel Glazman <daniel.glazman@disruptive-innovations.com>
Date: Sat, 19 May 2012 09:43:12 +0200
Message-ID: <4FB74F10.6080705@disruptive-innovations.com>
To: public-html@w3.org
Le 19/05/12 00:19, John Foliot a écrit :

> To your knowledge, is it used that way (to give useful information), in any form, today?  Or is it, in its current incarnation, still used for "marketing purposes" only?

Marketing purposes only, as far as I know. Please note some editing
tools add another meta tag to indicate a document was modified/edited
by their tool, preserving the original generator if it was another
tool. Again, marketing only.

FWIW, it can be useful for project management, just to tell authors
what editing tool was used. But making assumption on the document's
contents, structure, bugs or whatever based on generator or modified-by
meta tags seems to me a no-go. Totally unpredictable.

> So when HTML5 becomes a Recommendation, this new rule will be something that no authoring tool has had to deal with previously? I would suspect that this would also be the case for conformance checkers as well, correct?

Correct.

> If the meta-generator string and subsequent rule proved to introduce an accessibility issue, would you as a tool vendor remove this marketing string from your tool/templates?
>
> Do you have any opinion or knowledge of what other tool vendors might do under this same circumstance?

In short, you're asking tool vendors to remove the only hint in the
markup that will let the viewer of the source of a given page know that
the page was created by their tool?-) Honestly, no, I don't want to
remove that and I suspect other editing tool vendors will have the
same opinion, at first glance. The HTML WG will have to be extremely
persuasive to change that opinion, IMHO...
And I don't see why it should have to be persuasive. HTML5 is here to
make html grow on the basis on common practice. Common practice is that
the generator meta tag is used w/o constraints at this time. Adding
constraints to it seems to me contrary to the design of html5 and the
needs of the industry.

>> In that sense, basing any kind of rule on the presence of such a meta
>> tag seems to me a pretty serious conceptual error.
>
> I believe this is the heart of the problem, and reason for the request to re-open the issue. The Chairs appear to be seeking "concrete evidence" of, or "identification of explicit harm" based upon future events. As we can only hypothesize on future events, your opinion as a tool vendor will at least hopefully guide the discussion towards what we might expect from tool vendors in the future.

I agree. Asking for proofs, metrics, evidences or making assumption on
the impact of a totally new behaviour touching marketing is impossible
at this time. There's only one way to know: ping all the major editing
tool vendors (Coda, iWeb, Microsoft, Adobe, ...) to get their feeling
on this. Good luck on that !
 From my own perspective as a tool vendor and W3C Member, I already said
it, attaching a behaviour to the presence or absence of a meta
generator tag seems to me a conceptual error and something web authors
will skip.

I renew my deep concerns about the @alt handling rules. They're not
verifiable by a validator.

Side note: I'll be away from the keyboard most of saturday and sunday.
Don't make any assumption if you don't get a fast answer to extra
messages to me.

</Daniel>
Received on Saturday, 19 May 2012 07:43:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 19 May 2012 07:43:38 GMT