And though each project's final phase involves a great deal of manual
work resulting in a polished text that is the basis for both the versions
released to PG, it's interesting to note that there is no facility provided
for actually preserving this foundation text version. Oversimplifying a bit,
the text version removes a bunch of information, and the html version
adds a bunch of stuff. Arguably the most valuable text (for content and
metadata) is, at best, on someone's PC somewhere. Or more likely
discarded.
Well, be careful you are not laboring under misunderstandings.
The output of DP (that is to say, what comes out of round F2)
is _not_ a finished text.
Yes, there is some propriatory markup used at DP. However calling
it a "format" is going too far. I would call it "suggestive markup".
That is, its purpose is to record some information about layout
and format for the post-processor to use when they produce
html and/or text files.
The output of the rounds at DP usually contains plenty of
ambiguities (such as the propriatoy <tb>, which you may
want to ignore or render in different ways depending on the text),
proofer's notes (such as "is this a typo?"), and other
inconsistancies. It is the job of the post-processor to
take all this, ask for help if needed with any of the
issues with the particular text, and produce the final texts
which are submitted for posting.
--Andrew