[tei-council] P5 Howto - how?

[Syd Bauman]

CW> I retrieved Roma V.1.11 from SF and successfully generated a
CW> schema with that version.

I have been having similar problems with 1.13 myself. Even though
Sourceforge will do a very nice job showing the differences[1], it
can be quite difficult to figure out. I believe (but I'm not sure)
that one of the major changes between 1.11 and 1.13 is that the
former generated the 'compiled.rng' file by applying
expandincludes.xsl to the .rng file that was the output of step 1
"make Relax NG from ODD"; the latter tries to generate the
'compiled.rng' file by applying odd2odd.xsl to the ODD file specified
on the Roma commandline (and fails for lack of odd2odd.xsl).

So it looks like backing out to 1.11 will *run* as long as the value
of --xsl= is set to a local directory. (If you want to use the web
for --xsl, try commenting out line 132.) I.e., it looks like all the
component parts that 1.11 calls exist. However, I don't know that
there haven't been significant changes to the .xsl files that are
called that depend on them being called in the way they are by 1.13
instead of by 1.11. It is certainly worth a try, though (and I plan
to do so right after breakfast ... :-)


CW> [Error from Roma web interface] makes sense, since when I look at
CW> the Modules page, there seems to be no module selected at all, at
CW> least no one shows up in the list. I suspect that the input file
CW> has not been parsed correctly.

I got this error once out of the two times I generated a schema via
the web yesterday. I fixed it very quickly by manually adding the
required modules: core, tei, header, & textstructure.

Notes
-----
[1] Try, e.g.,
http://cvs.sourceforge.net/viewcvs.py/tei/P5/Roma?sortby=date&r1=text&tr1=1.11&r2=text&tr2=1.13&diff_format=h