W3C home > Mailing lists > Public > public-html@w3.org > September 2007

Re: More about <alt>

From: Sander Tekelenburg <st@isoc.nl>
Date: Thu, 6 Sep 2007 18:14:37 +0200
Message-Id: <p0624069cc305d4d12051@[]>
To: public-html@w3.org

At 16:36 +0200 UTC, on 2007-09-06, Olivier GENDRIN wrote:

> On 9/6/07, Sander Tekelenburg <st@isoc.nl> wrote:

[... defined keywords to indicate type of equivalent]

>> even we we define some generic predefined values like
>> "short", "long", "audio", "captioned video", etc. we would still have to
>> allow authors to use some non-predefined value: "table", "slide", "pdf",
>> ".doc", "x", "y", "z".
> But can't we define in the spec a short track for thoses issues ? A
> way to update the spec (not the draft, the final spec) in for example
> two month only for adding items in predefined names lists (role, rel,
> shape, ...), without changing the HTML version number ?

That's actually the case right now with @rel. See

I cannot really imagine this can work well though. The later a value is added
to a list of keywords, the bigger the risk that those keywords are already
being in use with a different meaning... So it would burden us with the task
to first research to what extent a newly added keyword would introduce

I suppose that in theory you could require authors to only use pre-defined
keywords, and when they need something else, to first propose that to the
HTML WG. But even if you'd provide an open and easy to use channel for that,
I doubt that many authors would actually do that.

And no doubt UAs will want to make the best of a web page that uses an
undefined keyword, so it seems safe to assume that at least as far as UA
behaviour is concerned, any 'keyword' will be accepted -- the UA would just
not be able to map it to user-friendly text.

Sander Tekelenburg
The Web Repair Initiative: <http://webrepair.org/>
Received on Thursday, 6 September 2007 16:15:58 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:26 UTC