Hannes Magnusson skrev:
Long long long long post, which I haven't read fully yet.
Just wanted to clarify few things you seem to have misunderstood.
In order not to burden this list with irrelevant stuff, I've discussed a
few things with Hannes in private emails. There was some confusion,
mainly coming from different usage of terms.
*Strict*
When I talk about being "strict", I refer to using a strict DTD in the
Doctype, i.e. HTML 4.01 strict or XHTML 1 strict.
We agree that there is no speed penalty in browsers for being strict.
*True*
When I talk about XHTML sent with an XML MIME header, usually
application/xhtml+xml, I call that "true XHTML".
We agree that Firefox before version 3.0 did not render true XHTML
incrementally, and that MSIE <= 8.0 and most probably 9.0 won't handle
it at all.
We agree that true XHTML is unusable PhD, now and for all foreseeable
future.
So what flavor should PhD use? (My personal opinion follows.)
Currently it uses XHTML 1.0 transitional according to its DTD, sent out
as text/html, in accordance with the infamous appendix C of the XHTML
spec, that allows it - as a transitional measure.
I recommend HTML 5 - using that subset of elements and attributes that
is supported by todays browsers. To start with only one single change is
needed. Shortening the Doctype to:
<!DOCTYPE html>
Google is already doing this on some of their pages. I am doing it for
all my new projects.
It will trigger standards mode in all current and future browsers. It
will also put MSIE 8+ in "super standards mode", and not the awful MSIE
7 compatibility mode.
To check whether the markup is acceptable, one does not look for *valid*
code according to the DTD, but for *conformant* code, according to a
conformance checker.
Being HTML 5 conformant should be a long term goal. It is not an
immediate need. IMHO.
Lars Gunther
