I'd like to gather and discuss information
on the topic of conversion tools, partly to,
but mostly from SGML/TEI.
In our institute for literature in Munich,
we use mostly PC-based and non-SGML-aware
software like WordCruncher, if we work with
computers at all. We are now investigating
the transition to SGML aware tools, but they
have their own restrictions, and are expensive.
To make the transition as smooth as possible,
and also to be more versatile in using TEI-tagged
documents, a flexible conversion tool seems
very necessary to me. `Flexible' here shall
mean both that it can do a lot and that it is
not too hard work to make it do it.
The conversion from older formats or unfor-
matted text to TEI tags is not the main objective,
because to my mind it is mainly a task for
a responsible human editor -- simply because
the information given by TEI tags is very rich
and often interpretation-dependend. (That's
why we have tags in the first place, right?).
The objective _is_ the mechanical translation
from TEI to plain ASCII (with various information
filtered out, perhaps), to TeX, to WordCruncher
markup or data base input. The preparation of
a new target system (like TeX) may always be
a bigger job, but the tuning of singular parame-
ters, like "print variant reading in square
brackets" or "omit variant reading" should be
So far, I've encountered three conversion tools
on the `free' marked:
* sgml2: Simple and straightforward, but far too
simple for TEI dtd, I'm afraid. I don't
consider it further.
* gf [general formatter]: Very flexible, once it's
programmed; it's author assured me that
adapting it to the TEI dtd would be a life
project, though. I'm afraid, this one won't
do either; but Gary, the author pointed me
* tpt [Technical Publications Toolkit]: This is software
from the Open Software Foundation (the
authors of motif, as far as I know), they
seemingly have written it for their docu-
mentation purposes but give it away as free
source. It provides table-driven conversion
from SGML, using sgmls's output as input.
This comes as close to my wishes as possible,
it seems -- maybe just because my wishes are
I'd be very glad both about comments on the general
idea and about experience or further hints concerning
the software from anybody!
For getting "out" of SGML, another tool you should look at is
the SGML Hammer from SoftQuad (be *sure* to tell them I sent
you when you call them). Last I remember it was a few hundred $,
and SoftQuad at least used to have good academic discounts.