As much as I hate to say it, the current draft <draft-ietf-html-spec-01.txt>
is not ready for RFC status. I just made one more review of this document
(I've been busy with other things, as you all know) and I am shocked at
the number of errors that are in the text. There are 16 errors in
the major section 1 (Overview) alone, and this whole section does not even
belong in the main body of the spec; it should be an appendix as a
quick-reference guide.
There is no logical flow to the section numbering, thus making it
near impossible to find where something is discussed [the same comment
can be applied to the DTD, but we can't do much about that].
This would be obvious if the TOC was generated down to three levels.
Various parts of section 2 (not in order) are the true Introduction;
it should also include a section on terminology (what is currently
section 7). The references are incomplete and in an unusual format.
The Author's addresses are wrong. The acknowledgements can be trimmed.
The main body of the spec is in better shape, but still averages one
error/inconsistancy per page. The DTD itself is in excellent shape
and all of the features of HTML 2.0 have (in my opinion) been decided.
The problem is that the WG (including me) has neglected the content
of the textual portions of the specification.
Now, before people start complaining about how I'm just bitching and
not providing the actual corrections, let me explain. First, it would
take me most of two days just to provide an annotated list of corrections
to the text version of the spec, and I won't have the time to do that
until Saturday (due to the need to get the HTTP/1.1 pre-draft done
before the I-D deadline). Second, it would be a waste of time to provide
diffs to the text version given the nature of the changes.
Instead, I would like to offer to make my changes directly to the
FrameMaker version -- all that I would need is the MIF file from draft 01.
I would then restructure and reformat the specification along the same
lines as the HTTP/1.0 specification (i.e., using the same formats) and
produce a base copy prior to making any content changes. Then, I would
make my own content changes (with change bars on) and make both versions
available at my web/ftp site. After that, I think Dan Connolly should
take up the Author's chair again and produce an HTML version of the spec
(Henrik can provide help with that because it would use the same WebMaker
configuration as the existing HTTP spec).
My part in the above can be accomplished this weekend if I receive a
copy of the MIF file by Friday night. So, the question is this:
Is the above plan acceptable to the WG? And, if so, can Eric and folks
at Spyglass provide me with the MIF?
If not, then I will write up my comments on the spec in normal, e-mail
form and send it in sometime Sunday or Monday. And please, don't bother
to tell me it's too late to do this -- the existing draft will not pass
muster with the IESG (it won't even come close).
Sorry,
....Roy T. Fielding Department of ICS, University of California, Irvine USA
<fielding@ics.uci.edu>
<URL:http://www.ics.uci.edu/dir/grad/Software/fielding>