Re: Old HTMLversions in guidelines

> > I made a list of the  HTML4.0 features that 
> > do not degrade well to HTML3.2. Here it is the whole list.
> 
> Are you suggesting we should mention this list somewhere ?

No. It was just to show that backward compatibility issues with HTML3.2 
are rather minor, and therfore shouldn't be mentioned that often.

In my view only the bidirectional algorithm, character entities and, 
TFOOT and BUTTON should be noted.

With HTML2.0, there are more issues but the major ones are noted already.
One is that MAP and AREA are not recognized, but we have priority 2 to 
add redundent text links to imagemaps.

Using OBJECT and %block; in MAP is also compatible with HTML2.0, but 
not with MSIE4, and pardoxically it is theoretically problematic 
with HTML3.2.

Another HTML2.0 issue is TABLEs, but here the guidelines correctly 
encourage using them for their purpose, rather than using the 
information-poor PRE as some older non-W3C "accessibility guides" 
proposed. 

Back to HTML3.2,
INPUT type=button is widely supported by browsers that support 
scripting, so it is rather a theoretical issue. 

Concerning deprecated colors and the S element, we discourage 
that anyway.

About having headings in LI, it makes some little uglyness problems 
in Opera3.2 (but not in 3.5) and Lynx (depending on version) but is not
critical, as no content is lost. Heading in DD seems to be handeled
without problems.

Concerning %Length; vs. %Pixels; both are defined as CDATA in the DTD, 
but the wording allows precentage values in %Length; i.e. width ="80%"
I think in practice more problems occur with cellpadding and cellspacing 
rather than with IMGs width and height, but this is not that important 
for WAI guidelines as authors never use precentage in cellpadding and
cellspacing or in IMG's border.

By the way, some browsers (Netscape 3) don't display INPUT elements 
outside of FORM. This structure is allowed in HTML3.2 and HTML4.0, 
but not in HTML2.0... So they implemented frames before 
implementing HTML3.2 correctly.

Regards, Nir.

> 
> In any case, it's an interesting exercise, and I'm asking our HTML
> staff to check it for completeness.

I ommitted a few minor things: there are some attributes that are required 
in HTML3.2 but are optional in HTML4.0. (like name in TEXTAREA) Again, this 
is a theoretical issue, as it wouldn't in practice break any browser, I 
think.

Received on Monday, 9 November 1998 07:27:29 UTC