Re: proposal for "null" alt-text (was "Re: GL's interpretation of null alt-text")

I agree entirely with Nir's position: it is better to advise authors to
use more appropriate techniques, particularly CSS, than to waste time
debating the merits of various text equivalents for so-called "spacer
images". Since this issue originally arose as a query from the ER working
group, our answer should suggest that all uses of null and blank ALT
attribute values (alt="" alt=" " and alt=" " etc.) should be flagged
as errors, with ALT="" being a more serious error than either of the
others.

The HTML specification does not determine how user agents will or should
render ALT=" ". It may be interpreted as a single-character attribute
value, comprising a space, or either leading or trailing white-space, to
be ignored during HTML processing, preceded or followed by the empty
string. As Nir has pointed out, the rendering of so-called "spacer
images" is subject to variation across distinct visual media and among
different user agents. The rendering of the text equivalent in these cases
is equally undetermined, as the present controversy has shown. In the end,
authors who wish to misuse HTML in an attempt to achieve formatting
effects run the risk of having their documents rendered in unexpected ways
by different user agents and/or in different media. Such practices should
be avoided.

Received on Wednesday, 17 November 1999 18:20:55 UTC