Marc Salomon <> wrote:
>I'm concerned more with VALUE/ALT being supported by them when image loading
>is disabled than with VALUE being ignored by GUI UA's that are loading images.
>I'd prefer to lace my WWW application interfaces with ALT's (so that the user
>can simply disable image loading to achieve an image-less version of the
>application) as opposed to dealing with the overhead of generating image-less
>pages for text-only support over slow connections.

	Think about what is intended to be done if the INPUT has a NAME
and is to generate content in the submission.  There is no why to do
what's specified unless the image has been loaded, and you can click
somewhere in the image.  If doesn't have a NAME, such that it will not
generate content, adding an ALT simply complicates the DTD for INPUT,
without adding anything that can't already be done via the VALUE
attribute, which already is being used, by all clients, as "alternate
text" to be displayed for INPUTs with TYPE="submit" (otherwise, there
most certainly would have been requests to have an ALT added, long ago,
and persistently 8-).  We're talking about an INPUT field for a FORM,
not just an inlined IMG.

>I'm not sure which is worse:
>1.  Adding Yet Another Attribute to INPUT for TYPE="image" to follow the ALT
>convention for unavailable images.  or
>2.  Recycling VALUE and breaking the ALT convention making image unavailablity
>with INPUT into a special case.
>Its all well and good that lynx works well with VALUE, but I'd prefer to see an
>single-implementation-independent standard.

	Again, don't just think about what might be displayed with image
loading turned off, but also about how two name=value pairs indicating
the X and Y coordinates upon which the user clicked could be generated
if there is no image upon which to click...


 Foteos Macrides            Worcester Foundation for Biomedical Research
 MACRIDES@SCI.WFBR.EDU         222 Maple Avenue, Shrewsbury, MA 01545

Received on Wednesday, 30 October 1996 19:38:31 UTC