Re: draft-ietf-html-table...

Gerald Oskoboiny (
Thu, 3 Aug 95 23:29:56 EDT

Lou Montulli writes:

> >> We have been very good about not introducing conflicts ...
> >Then how do you explain <CENTER>,...And why do you still advocate its use?
> <CENTER> does not conflict with <? ALIGN=CENTER> in any way and contained
> more functionality than the latter. Now that <DIV ALIGN=CENTER> exists
> it will be implemented and used preferentially

I'm glad to hear it! However, in the meantime, the number of documents
that use <CENTER> continues to grow. Why not put something in your
extensions document [1] to the effect of "we strongly encourage HTML
authors to use <H? ALIGN=CENTER> and <P ALIGN=CENTER> whenever possible,"
and promote this usage by doing similarly in your own pages?

Also, there was public discussion of <DIV> as early as October, 1994 [2];
here we are over 9 months later and it still hasn't been implemented
although many other features have been added to your browser in that time.
If you had said "hey, guys, we want to center multiple things at once;
any ideas?" ahead of time, chances are you could've implemented <DIV> from
the start, saving many HTML authors and browser implementors from future
hassle. (If this was in fact proposed ahead of time, I apologize, but as
far as I know it wasn't).

> >Also, <TABLE BORDER=0> will give a border on browsers which implemented
> >the proposed table specification, which is different than Netscape's
> >behavior.
> BORDER=0 is a "border" case. Since we allowed for an integer value
> we had to have some behavior when it was set to zero.

Yes, but this usage conflicts with the proposed specification, which is
the point I was trying to make.

> People should not use BORDER=0, they should instead be leaving BORDER
> out entirely since that will render correctly in all browsers.

I agree completely! But this is much different than the advice given
on [3], which reads:

By allowing the BORDER attribute to take a value, the document
author gains two things. First [the ability to emphasize].
-> Second, by explicitly setting border to zero they regain that
-> space originally reserved for borders between cells, allowing
-> particularly compact tables.

Many HTML authors are currently using <TABLE BORDER=0> to mean "no border",
and are surprised when told this conflicts with proposed HTML 3.0.

We'll have to agree to disagree on <BLINK>... :-)

I suspect this discussion no longer belongs on this list, but I'd be more
than happy to continue this offline if anyone at Netscape is so inclined.
Thanks for your reply,


Gerald Oskoboiny  <>