RE: alt and authoring practices

>>> What data would you like me to collect?
>>
>> Well, the data from a web crawl that seem germane would be along the
>> lines of percentages of images for the oft-mentioned three cases:
>>
>> . have no alt attribute
>> . have an alt=""
>> . have an alt="(a descriptive string)"
>>
>> Obviously that still gives you no sense how often the alt text is
>> useful, but it's a start.
>
> A start towards what?
>
> As I said to Karl, before I spend money on this, I'd really like to know
> what hypothesis we are testing.

Sorry, I should have included more context in my original reply.
This hypothesis, from Anne van Kesteren, replying to Steven Faulkner:
( http://lists.w3.org/Archives/Public/public-html/2008Apr/0486.html )

> Given that authors make mistakes there are nine possibilities of authoring

> images:
> 
>   1. <img alt="..."> - available -> Correct usage
>   2. <img alt=""> - available -> Incorrect usage
>   3. <img> - available -> Incorrect usage
>   4. <img alt="..."> - missing -> Incorrect usage
>   5. <img alt=""> - missing -> Incorrect usage
>   6. <img> - missing -> Correct usage
>   7. <img alt="..."> - empty -> Incorrect usage
>   8. <img alt=""> - empty -> Correct usage
>   9. <img> - empty -> Incorrect usage
> 
> It seems your assumption is that on average 9 is more common than 3 and 6

> combined and that therefore <img> should be equivalent to <img alt=""> as

> far as user agents go and we should have an alternative solution to cater

> for 6.
> 
> It seems the assumption from the editor is that on average all incorrect  
> usage is about as likely and that therefore 3 and 6 should win from 9 and

> that therefore <img> might as well be used for this case.
> 
> With nobody having data of usage on the Web the position of the editor  
> seems more reasonable to me.

regards,
Matt
--
Matt Bonner
Hewlett-Packard Company


-----Original Message-----
From: Ian Hickson [mailto:ian@hixie.ch] 
Sent: Wednesday, April 16, 2008 3:49 PM
To: Bonner, Matt (IPG)
Cc: public-html@w3.org
Subject: RE: alt and authoring practices


Trimming cc list.

On Wed, 16 Apr 2008, Bonner, Matt (IPG) wrote:
>>
>> What data would you like me to collect?
>
> Well, the data from a web crawl that seem germane would be along the
> lines of percentages of images for the oft-mentioned three cases:
>
> . have no alt attribute
> . have an alt=""
> . have an alt="(a descriptive string)"
>
> Obviously that still gives you no sense how often the alt text is
> useful, but it's a start.

A start towards what?

As I said to Karl, before I spend money on this, I'd really like to know
what hypothesis we are testing.


> One could imagine trying to run algorithms against the alt text to get a
> sense of the frequency of useful alt text, but that sounds more like
> research than data collection. Not sure the alt-trust-level value can be
> determined algorithmically. ;-) [...]

Indeed, anything that requires human intervention isn't something I can
do without significant help.

--
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'