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

Re: a question about alt

From: Leif Halvard Silli <lhs@malform.no>
Date: Thu, 06 Sep 2007 12:17:44 +0200
Message-ID: <f36f5e1c2d84593a5a64b3c5f34dd101@10013.local>
To: "Steve Faulkner" <sfaulkner@paciellogroup.com>
Cc: "James Graham" <jg307@cam.ac.uk>, HTMLWG <public-html@w3.org>

2007-09-06 11:49:47 +0200 "Steve Faulkner" <sfaulkner@paciellogroup.com>:

> James wrote:
>> Distinguishing the cases alt="" and alt=" " would make it very easy to
>> typo a meaningfully-different value and very hard to spot the mistake.
>> If such an explicit indicator is desirable, using alt="" and noalt seems
>> like a better solution.

So, James, you are thinking about the semantics for the authors, then - the handcoder/eye investigator.

> I understand what you are saying. the reason I have suggested this is that a
> new attribute would not be backwards compatible with assistive technology.
> The alt=" " suggestion is treated by the assistive tech i have tested it
> with, the same way as alt="" (the image is ignored with default settings).

I think the keyword in what James said is «spot». When you *look* at a single element, it might be difficult to see whether it is alt="" or alt=" ". But,

1. One will, most often, not «spot», one will do a search. And, unless we want to create a spec with which we can test the RegEx skills – or tools – of the authors, then it is much more simple to search for alt=" " than it is to search for for IMG-elements which are lacking an @alt. One doesn't even need a RegEx tool to find an alt=" ".
2. Often, authors will use author tools that hide the code. And then the author doesn't need to think about this.
3. As I have said, it is difficult to have two kinds of negative alt. Authors will not understand the difference between no@ALT and @ALT. (And that is _my_ word about the semantics for the authors.)

I also actually think that it can be difficult to «spot» that there is not @alt - but that might be just me. I would be quick to not trust what I see.
-- 
leif halvard silli
Received on Thursday, 6 September 2007 10:18:10 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:49 UTC