[whatwg] ALT and equivalent representation

[Bill Mason]

Smylers wrote:
> Bill Mason writes:
> 
>> Simon Pieters wrote:
>>
>>> For instance it would be reasonable to use two images -- a filled
>>> star  and an unfilled star -- to represent a rating of something:
>>>
>>>    <p> Rating: <img src=1> <img src=1> <img src=1> <img src=0>
>>>    <img src=0> </p> 
>>>
>>> You'd want the text version to be:
>>>
>>>    Rating: 3/5
>> There would probably be the argument for more literal interpretations
>> such as
>>
>> Rating: (3 black stars in unicode) (2 white stars in unicode)
> 
> You could do that, and you could also use <meter> to denote 3/5.  But
> that's irrelevant -- Simon's example of alt=3/5 clearly is a genuine
> alternative way of representing the same information as the images do,
> and authors should be free to decide that's the text representation they
> wish to use.

My point was only that what an author would "want the text version to 
be" is open to interpretation, not that the suggested text and how the 
alt attributes would be defined to produce that result was not a valid 
alternative.

> To think about it slightly differently, obviously it would be possible
> to represent a rating with a single image containing 5 stars of the
> appropriate colours; the author would just need to pre-compose the
> relevant images for each possible rating, in which case 3/5 could be
> implemented as:
> 
>   <img src=3_5 alt=3/5>
> 
> And in that case the alt text of 3/5 is acceptable, and clearly in the
> right place.
> 
> Deciding instead to achieve this as Simon suggested above (have just 2
> images and to display each rating by combining 5 instances of those
> images) is an implementation detail.  There are benefits to each.  HTML
> 5 shouldn't bless either of them as being 'preferred'.

I don't recall suggesting that anyhow.

> And since with images enabled both implementations render identically it
> follows that alt text appropriate in one implementation is just as
> appropriate in the other.  So what Simon suggested does make sense: when
> several images are combined to convey something as a whole, it should be
> a valid alternative to put text conveying the whole on any one of them,
> marking the rest with empty alt text.

And I actually lean toward agreeing with that suggestion.

-- 
Bill Mason
Accessible Internet
whatwg at accessibleinter.net
http://accessibleinter.net/