[uf-discuss] making img machine-readable

[Benjamin Hawkes-Lewis]

Many CSS image replacement techniques are highly problematic for users 
with visual impairments. Let's say you hide some text off-screen and 
replace it with a background image. Let's say one of your users has 
colorblindness or otherwise impaired vision and needs to force 
particular background and foreground colors. They'll need to disable 
background-images as a matter of course. At that point your text is 
basically invisible unless they disable all your CSS. By contrast, if 
they disable HTML images, they'll see your alt text.

I haven't yet seen a flavour of image replacement without any 
accessibility issues (feel free to point me to one), and I don't think 
microformat parsing should assume the use of such fragile techniques.

--
Benjamin Hawkes-Lewis

Andy Mabbett wrote:
> In message <4699D714.1000100 at xtra.co.nz>, Paul Wilkins 
> <paul_wilkins at xtra.co.nz> writes
> 
>> Andy Mabbett wrote:
>>
>>> Furthermore, you would speak:
>>>
>>>    <span class="fn"><img src="GothicLetterA.png" alt="A">ndy
>>>    Mabbett</span>
>>>
>>> as "Andy Mabbett", which is why microformats should also interpret it 
>>> that
>>> way.
>>>
>>>
>> That should be marked up in a more modern manner using CSS image 
>> replacement, which entirely resolves any issues.
>>
>> <span class="fn"><span class="gothicletter a">A</span>ndy Mabbett</span>
> 
> That's a matter of opinion, and an option not open to users of some 
> systems.
> 
> Besides, using microformats shouldn't require a publisher to use one 
> valid method over another; the example I give above *is* valid.
>