> > Regarding Matthew's earlier comments, which I can't find, it would be
> > _very_ _bad_ to use XHTML and serve it as text/html - without
> > enforcing the well-formedness constraint on XHTML from the very start,
> > you are throwing away the world's best chance to have a HTML-esque
> > format which is zero-cruft, and can be understood by lightweight
> > parsers which don't need to deal with broken code.
> >...
>
> I don't believe that for a moment. And the reason I don't believe it is
> that XHTML 1.0 Transitional (which is what I propose we use) exists at
> all. XHTML Transitional wouldn't exist, if XHTML hadn't been intended to
> be served to clients which only understand text/html.
What _benefits_ are there in using XHTML over HTML 4.01 Strict? I'm not
going to post Henri's opinion here a third time, but I do agree with it -
using XHTML without enforcing well-formedness at the browser is just
asking for trouble.
I remember reading, in a bug I think, that they had a big argument about
this on the W3C list. I am pretty sure that the conclusion was XHTML as
text/xml only, and HTML as text/html, but I'm not certain. dbaron (among
others) would know.
> > > 6) We need to settle on filesystem layout and URI layouts. Some have
> > > been suggested, I'm not aware of anything we should be working to.
> > > It's make our minds up time.
> >
> > Again, this problem goes away if we use a content management system.
>
> No, it doesn't. We need to minimize linkrot whether or not we use a
> content management system. A content management system might make it
> easy to keep links on our own site up to date, but it will do nothing to
> update links to mozilla.org which reside on other people's Web sites and
> bookmark lists.
If we use a content management system, there's no such thing as a
filesystem layout. Therefore there's no argument as to what it should be.
The only thing is the navigation layout (see below.)
> > More important is the navigational layout, which is a separate thing
> > which keeps getting conflated with the filesystem layout and even the
> > URI layout.
>
> To mazimize usability, the URI structure should match the navigational
> structure as much as possible.
But the navigational structure won't be a tree. At least, I would have
thought it wouldn't be - the problems we've had deciding where things live
are symptoms of the fact that, navigationally, some things live in two or
more places.
> > If it works on Nav 4.x on any platform, it should work on the others -
> > right?
> >...
>
> No. 4.x versions for different platforms have different crasher bugs.
But the rendering is basically the same? It's not like IE 5 for
Mac/Windows. Right?
Gerv
