[uf-discuss] RFC: Using <img>'s alt attribute for Microformats

[Tantek Ç elik]

On 12/14/05 3:24 PM, "Ryan King" <ryan at technorati.com> wrote:

> On Dec 14, 2005, at 3:04 PM, Scott Reynen wrote:
>> On Dec 14, 2005, at 4:15 PM, Ryan King wrote:
>>> On Dec 14, 2005, at 12:47 PM, Charles Iliya Krempeaux wrote:
>>>> So, maybe hCard's, hReview's, etc should allow this too?  For
>>>> example:
>>>> 
>>>>     <address class="vcard">
>>>>         <img class="fn" src="my-name-in-styled-lettering.png"
>>>> alt="Charles Iliya Krempeaux" />
>>>>     </address>
>>>> 
>>>> (And, yeah, I know this would give the problem that would can't add
>>>> the sub-structure to the "fn" that hCard's allow.  But you could
>>>> always break it up into multiple images if you wanted.  And at least
>>>> then you'd support this very common practice.)
>>> 
>>> For properties that don't expect images it might be reasonable to
>>> use the alt attribute value for the value of the property.
>>> However, it might be a bit too open ended (and how often do people
>>> do this, anyway?).
>> 
>> Using images of stylized text with alt attributes containing the
>> text is quite common:
>> 
>> http://www.google.com/
>> http://www.apple.com/
>> http://www.yahoo.com/
>> http://www.microsoft.com/
>> and, um...
>> http://www.technorati.com/
>> http://microformats.org/
> 
> Eh, good point. I guess I wasn't thinking about logos and such. But
> in the case of technorati.com, we could reasonably make our logo into
> an hcard with:
> 
> <span class="vcard">
> <a class="url" href="http://technorati.com/">
> <img class="org fn logo" src="http://static.technorati.com/pix/tn-
> logo.gif" alt="Technorati" />
> </a>
> </span>
> 
> With the interpretation being:
> 
> fn=Technorati
> org=Technorati
> url=http://technorati.com/
> logo=http://static.technorati.com/pix/tn-logo.gif

Ryan is correct, and this has been specified for quite some time in the
hCard parsing spec:

<http://microformats.org/wiki/hcard-parsing#parsing_hCard_properties_and_val
ues>


>> If you're asking about TV show titles specifically, I'm sure
>> Charles could answer that better, but here's one:
>> 
>> http://tv.yahoo.com/
>> 
>> My question: which should take precedence between alt and title
>> values?
> 
> Good question.
> 
> From the spec [http://www.w3.org/TR/REC-html40/struct/
> objects.html#adef-alt]:
> 
>> alt = text [CS]
>> For user agents that cannot display images, forms, or applets, this
>> attribute specifies alternate text. The language of the alternate
>> text is specified by the lang attribute.
>> Several non-textual elements (IMG, AREA, APPLET, and INPUT) let
>> authors specify alternate text to serve as content when the element
>> cannot be rendered normally. Specifying alternate text assists
>> users without graphic display terminals, users whose browsers don't
>> support forms, visually impaired users, those who use speech
>> synthesizers, those who have configured their graphical user agents
>> not to display images, etc.
> and [http://www.w3.org/TR/REC-html40/struct/global.html#h-7.4.3]
> 
>> title = text [CS]
>> This attribute offers advisory information about the element for
>> which it is set.
> 
> It seems to me that title would not make sense for content (except
> for in the case of the abbr-datetime pattern). Alt seems more
> meaningful here.

Agreed.

In general we should not be using "title" for this kind of purpose unless
*absolutely* necessary, like in the abbr-datetime pattern, as Ryan pointed
it.

The rule still is to keep the data as visible as possible.  We made an
exception in the abbr-datetime case only because of another rule which is
humans first, machines second.

Thanks,

Tantek