Martin wrote (in arguing against my proposal that the template design
rule should be X/HTML strict rather than transitional):
Basically it means we can still use tables, that's
the main difference as far as I'm concerned. Personally, I'm not
comfortable with completely dropping them, since table-less
design is still so new and IE is still so painful at handling
it.
Also, XHTML Strict implies that ALL content in a page will be strict, and given that teachers and students are editing the content this can not always be true.
Lastly, there can definitely be absolute separation of styles
from the templates, this is true even of HTML 4. For example even
the tables can be styled using CSS.
I'm sorry for harping on about this, but I do think it's worth
investing a little effort in teasing out now, before template
design is substantially underway, rather than coming back to it later.
As David has pointed out, the distinction between using tables
for layout versus CSS for layout is essentially separate from
using strict or transitional X/HTML. True, "philosophically",
using tables for layout is rather against the "spirit" of strict
HTML (which basically is trying to push all styling, which would
ultimately include layout, out of HTML and into CSS). But there
is no technical conflict: one can certainly use tables for
layout in completely valid X/HTML strict. So, since Martin goes
on to acknowledge that all table "styling"
can be in CSS, there is no argument here for transitional over
strict.
Whether or not to actually use tables for layout is another
question; my point here is simply to emphasise that it
is another question, and shouldn't muddy the debate
about strict versus transitional.
Secondly, Martin makes the very good point that even if the
templates produced X/HTML conforming to a strict DTD, the
included user content may well violate this again. This is very
true. But it also again largely transcends the strict versus
transitional question. User content may well be "bad" in lots of
different ways, and if we don't do anything on the server side to
control for that, then we won't be able to deliver valid HTML of
any kind. OK, transitional is more "relaxed", so there
is a better "chance" of not generating bad code if we label
conservatively as transitional rather than strict. But I basically
agree with Petr that this is a separate problem, that needs
to be addressed separately. In any case, even if we
wind up labelling the code sent to the client side as
transitional, just to allow for poor mark-up quality of user
content, that is hardly an argument for endorsing the use of
transitional markup in the templates as a matter of design
policy?
Finally, Martin writes:
The very long and hard work will be converting to
templates, so I want to do that first. Converting later from
XHTML Transitional to XHTML Strict will be much, much easier and
will be something that web designers with no knowledge of PHP can
fiddle about with to their heart's content, since all they have
to deal with will be XHTML files and CSS.
So, again I want to (respectfully!) register my disagreement -
and I'm acutely mindful of the fact that it is other people, not
me, who are generously stepping up to the plate to actually
do this work. And yes, converting to a templated
architecture is going to be long and hard work. But it just is not
apparent to me that that is going to be made any less so by
allowing transitional html in the templates. (Again, the issue
of tables vs. CSS for layout is another, but disjoint, question.)
I may be completely missing the point here (and please, somebody
clobber me, if I am!) but I don't yet understand any specific
feature of transitional HTML that it is actually desirable to
retain in the templates...