Re: LYNX-DEV ISMAP/USEMAP alt text

[Klaus Weide]

On Mon, 7 Apr 1997, Foteos Macrides wrote:

> "Hiram Lester, Jr." <address@hidden> wrote:
> >On Sat, 5 Apr 1997, Al Gilman wrote:
> >
> >> Can anybody think of a reason to keep the [ISMAP] legend on the
> >> screen if a USEMAP has been furnished in the web page?  I would
> >> agree with Hiram that it's best it just vanish if there is a 
> >> USEMAP that Lynx can use.

[ To which Fote replied with some more historical perspective,
then: ]  
>       When I implemented MAP/USEMAP handling in Lynx, we had set
> up the server-side imagemap "workaround" of sending 0,0 coordinates
> so that Lynx-friendly sites could use that point, presumeable as
> part of the background "default", for sending a text-suitable
> alternative page to Lynx.  The MAP/USEMAP markup does not include
> a "default" attribute -- one of its many, frequently criticized,
> weaknesses -- and few providers have figured out how to work around
> that handicap.  The consequence is that it can be difficult to work
> a text-suitable AREA into the MAP without conflicts for GUIs.  I
> therefore ignored the provision that an ISMAP attribute, if present,
> should be ignored if a USEMAP attribute is present, 

Apparently Scott, who started this, would prefer a behavior *without*
ignoring the "provision to ignore".  He does want the ISMAP attribute
to be ignored - although on his page at http://sol.slcc.edu/index2.html
he also provides a useful response for the ...?0,0 A link. 

> and set up the
> code to use the ALT string, if present, for the client-side MAP
> content, but prepend an "[ISMAP]" link followed by a hyphen to
> indicate that it is associated with (an alternative to) the
> client-side MAP content, and could still be used to send a 0,0
> coordinate pair for the text-suitable page at the Lynx-friendly
> sites.  

Continuing to speculate about Scott's wishes [:-) but he is here
to correct me] -- maybe he would be happy with the ISMAP link
remaining, if he could just find a way to give it a less cryptic
link text than "[ISMAP]".  

I note that ther *is* a way to have different text - just put it
between the A and IMG tags, as in 
  <A HREF=...>Additional Text<IMG ISMAP USEMAP=...></A>.
That Additional Text will be selectable in Lynx, and will appear
in addition to "[ISMAP]-" or, with my patch (and without '*' ON),
instead of it.   
But I don't know how graphical browsers would react to it,
probably not in a useful way.

[...]
>       If Klaus' patch is used, it needs code to eliminate the
> associated Anchor for the server-side imagemap, and not yield a
> dangling link when links_are_numbered is ON.  

(I assume with "dangling" you mean just that it is not selectable
because it has zero-length text.  If there is something left dangling
in some other sense, pointing to nothing or similar, please clarify.)

Yes, it is ugly to have that numbered link without being able to select
it (other than by its number) and without indication what it is.
But having just "[ISMAP]" and no way to change it is also ugly..
I think default strings like "[IMAGE]" or "[LINK]" make some sense
to a user who doesn't know any HTML (as long as he knows the English
words), but "ISMAP" doesn't.

If "code to eliminate the ... Anchor" means diving into GridText's
(etc.)  form handling code and adding yet more special treatments
there, then I cop out at this point.  Better to retract my patch then,
unless somebody else wants to do that.

> That wasn't needed
> for the vanilla code, which creates links for both.

Would it be reasonable to (mis?)use some other attribute on the A tag
(or even the IMG tag) for the purpose of supplying a different anchor
text?  Maybe TITLE?  [CLASS?]  Just speculating, but this seems to me
the cleanest solution, if it can be done without contradicting any of
the defined purposes for such an attribute.  It would also eliminate
the "dangling link" problem.

   Klaus

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;