Re: Change proposals for ISSUE-31 and ISSUE-80

On Tue, Jul 20, 2010 at 7:54 PM, Matt May <mattmay@adobe.com> wrote:
> Here, you've offered your opinion as a starting point, but demand "data" to counter it.

I didn't demand anything.  Indeed, I'm in no position to demand
anything, since I have no ability to change the spec directly or
indirectly.  I asked a question, to which I did not know the answer
(and still do not, after reading your response).

> I have to suggest that you're suffering from selection bias from working around people who place a higher value on validation than adhering to the spirit of the spec. My feel for things, which is likewise colored by working with my share of developers and designers, not to mention blind users themselves, is precisely the opposite. I'd suggest that more people produce meaningful alt text than fake it for the sake of a validator, but I don't see how anybody is going to be swayed either way.

It seems pretty simple: we could pick pages at random and look.

> Instead, let's look at outcomes. In the case where @alt is mandatory, even when there are bad actors inserting bogus @alt values, it is exceedingly rare that such behavior does real harm to a user who can't see that image, relative to the possibility of not having it. In every other case, the presence of alt text, even when poorly done, either improves the situation for users who need it, or is easily ignored in favor of other repair techniques.

I realize that this is the understanding of those who favor mandatory
alt text.  This is why I specifically asked if anyone had *data* about
the subject, rather than just asking what people thought about it.  I
know what people think about it, but it is not a priori obvious that
the benefits of requiring alt text outweigh the harm, so hard evidence
would be useful to inform the discussion.  If none exists, of course,
we have to deal with what we have, but I thought it was worthwhile to
ask.

Received on Wednesday, 21 July 2010 18:51:53 UTC