Re: Clarification of case where 'there simply is no text that can do justice to an image'

 From an actual site.

<div>
<h2><img src="orangeBeachText.gif" alt="Featured Location: Orange  
Beach"></h2>
<img src="orangeBeach.jpg" alt="Orange tinted photo of Orange Beach" />
<p>
Lorem ipsum dolor sit amet, consectetuer adipiscing elit.
Duis egestas solicitudin lacas. <img alt="" src="common/trans.gif"  
id="spacer"/>
Ut dui dui, pretium ut, fermentum in, commodo in, massa.
</p>
<a id="visit" href="orange.php">Visit This Place</a>
</div>

In this case the designer wanted to have an outlined div of a fixed  
height with a button to click absolutely positioned in the bottom  
right of the div with the paragraph right above it but the lower  
portion of the paragraph needed to flow around the button.

Simply including the image at the end of the paragraph would not give  
the desired appearance.
The CMS would not allow <br> tags so the data entry people couldn't  
just put a break where needed to prevent text from going behind the  
button.

The best solution seemed to be to just put a transparent gif 110px by  
30px after the first available whitespace following the 90th character  
of the paragraph and float it right.

With 200 randomised possibilities for the paragraph text this achieved  
the desired visual result 96% of the time.

I've since changed the gif image to just be another div but lets  
pretend it was still an image.

What I'm wondering is if I should put an alt value on the image and if  
so, what would be the best thing to say?

Would people using AT even want to waste their time reading about this  
particular pointless image?

On Aug 23, at 5:10 AM, Ian Hickson wrote:

>
> On Sat, 16 Aug 2008, Steven Faulkner wrote:
>>
>> The current HTML5 spec contiains this expalnation and example of a
>> conforming alt text [1]:
>>
>> "Sometimes there simply is no text that can do justice to an image.  
>> For
>> example, there is little that can be said to usefully describe a
>> Rorschach inkblot test.
>>
>> <figure>
>> <img src="/commons/a/a7/Rorschach1.jpg" alt="{inkblot test}">
>> <legend>A black outline of the first of the ten cards
>> in the Rorschach inkblot test.</legend>
>> </figure>"
>>
>>
>> In light of the example in the spec, is the following conforming to
>> HTML5 or not?
>>
>> <figure>
>> <img src="/commons/a/a7/Rorschach1.jpg" alt="a black vertically
>> symmetrical shape, which contains 4 small white areas and 9 smaller
>> black blotches seperated from the main body of the shape.">
>> <legend>A black outline of the first of the ten cards
>> in the Rorschach inkblot test.</legend>
>> </figure>
>
> While I think that describing a Rorschach inkblot, even for the  
> purposes
> of the alt="" attribute, is an act with such irony that I would  
> personally
> not think it appropriate, I could see an argument that would  
> convincingly
> suggest that the above is valid, so I've changed the spec to in fact
> require a description of some sort for such images.
>
> In fact, on the advice of Ben in an off-list e-mail, I have split  
> the "key
> content" section into three subsections to make it clearer that  
> there are
> three different levels here; images that must be given full  
> replacement
> text (such as photos on a blog), images that somewhat defy  
> descripton but
> must nonetheless be described (inkblot tests, fractals), and images  
> that
> should be described but for which the HTML generator might simply  
> not have
> any means to describe the image (blind photographers, CGI scripts  
> dealing
> with unannotated images, Webcams, etc).
>
> -- 
> Ian Hickson               U+1047E                ) 
> \._.,--....,'``.    fL
> http://ln.hixie.ch/       U+263A                /,   _.. \   _ 
> \  ;`._ ,.
> Things that are impossible just take longer.   `._.-(,_..'-- 
> (,_..'`-.;.'
>