Re: Thoughts on professional portrait (headshot) images?

On 19/08/2022 23:50, John Foliot wrote:
> (i.e. alt="Photo: John Foliot", and *not *alt="Photo of John Foliot who 
> has a white handlebar mustache

I think anything beyond the first of these examples reveals more about 
the the writer of the alt text than the subject of the picture, as there 
is so much information that can be inferred from a portrait, that even a 
longish description will be incomplete, and the selection is likely to 
represent what the coder thinks, or is told, is important. Even if we 
are following the "looks like me" trend, it might be some completely 
different attribute that any particular sighted viewer identifies with.

Also, what can be gleaned from a photograph may not represent the 
classification into which subject would put themselves, or to which they 
are generally considered to belong.  For instance, given such a portrait 
of Megan Merkle, if I didn't know her history, I would probably describe 
her as a white woman.

Some of the other attributes suggested in the long versions are just not 
things that you could reliably deduce from a head and shoulders 
portrait, or even a full length one.

Radio broadcasting is possibly a good example of a situation where 
speakers are introduced without the ability to see them.  In the UK, we 
have the BBC as a public service broadcaster, who therefore puts out a 
lot of current affairs material involving experts in subjects.

They do not give physical descriptions of those experts, and only give 
relevant professional background.  Taking the black/white division, I 
hear experts used who have also appeared on TV, and would be understood 
as black from their video image, but not from their name or accent. That 
fact is never mentioned; they are just subject experts.

Even when race relations, or say, particular musical or food traditions, 
is relevant, the introduction might say something about where they or 
their parents were born, but it would be up to the listener to make the 
black/white inference.

Received on Saturday, 20 August 2022 21:51:08 UTC