Re: LONGDESC vs D-link (was LONGDESC for OBJECT)

> <OBJECT TYPE="image/gif" DATA="somepic.gif" WIDTH="33" HEIGHT="33">
>   <IMG SRC="somepic.gif" WIDTH="33" HEIGHT="33" ALT="something sensible">
>   <A HREF="somepic.htm" TITLE="D-link for somepic.gif"><IMG SRC="dot.gif" 
>     WIDTH="1" HEIGHT="1" ALT="D-link"></A>
> </OBJECT>

why, as an author-by-hand or as a authoring tool developer, would I
wrap my IMG into OBJECT like that, rather than just providing the IMG ?

what is the value added ?

Also, I think it's too complex, most people on GL can hardly follow
what we're saying by now, so imagine what our targeted audience will
think if we release that.

We should keep things much simpler.

OK, our base constructs are not that clean, we have some historical
markup legacy here, but our role is also to straighten that up.

The way it is is: IMG is for still image, OBJECT is for other richer
media. For all I know, with XML ramping up, we might even see a
multi-media tagset for HTML5 that provides VIDEO, AUDIO, etc elements
(i.e. move away for generalization of OBJECT into specialization,
which is what most people seems to like, all but computer
scientists...)

For short desc: ALT for IMG, content for OBJECT.

So far so good.

For long desc, which is very useful, but is still a niche application
on the web: 
   recommend LONGDESC for IMG (it's HTML4 standard, so soon to be
     implemented in IE, Opera, Netscape and Lynx)
   mention dlink as current practice (interim technique)
   recommend <A rel=longdesc> for OBJECT long desc in content

Despite the fact that as a computer scientist, I think using
PARAM for OBJECT longdesc is a cleaner solution, I think it's so
unlikely that it be implemented by AU and UA that I prefer to drop the 
idea right now, in favor of long desc handled in the content of
OBJECT, with a sugar coat convention rel="longdesc" to allow for UA
support.

Received on Friday, 31 July 1998 10:08:40 UTC