On Mon, 5 Feb 1996, Michael Johnson wrote:
> >Would ignoring the style specification be preferable?
> Yes, I think it would be. The browser still ought to produce some kind of
> error message however.
I agree that (if the STYLE attribute goes through), the style specification
should be ignored if no stylesheet notation is given.
On a more important note, however, i strongly recommend that all future
HTML specifications explicitly suggest or even require that compliant
user agents must display *some* sort of indication (however inobtrusive)
when they have received invalid markup.
After all, when a browser runs across something it does not expect or
something it was not programmed to understand, it must warn the user
against information loss. The entire communication has not made its way
from provider to consumer.
Also regarding information loss, i feel that no information should be
thrown away by the user agent, as is so common in current practice. Often
the only way to get certain fields from HTTP headers is to use telnet,
because the browser won't give you the information. Complete header
information should be available somewhere.
Likewise, all the information in the HEAD, and ID, CLASS, and (for 2.0)
NAME identifiers should be accessible without requiring the user to read
the source. It's very annoying to have to parse out information that the
browser has already processed, but refuses to give you.
Ping (Ka-Ping Yee): 3A Computer Engineering, University of Waterloo, Canada
[log in to unmask], St. Paul's College, Waterloo N2L 3G5, 519 725-8008
CWSF 89, 90, 92; LIYSF 90, 91; Shad Valley 92; DOE 93; IMO 91, 93; ACMICPC 94
! Ayukawa Madoka ! Hiyama Hikaru ! Tendou Akane ! Hayakawa Moemi ! Amano Ai !