A somewhat improvised version of the ARIGFH indentification infrastructure.

This was used to re-do some identification work before we finally decided
to read all data ourselves (res/liesgrammar, q.rd/import_gfh).  Still
the remarks here apply to what's going on there, and the infrastructure
in here can of course be used for debug purposes.

Also, this is the source of the idfailures.nid files that import_nids
uses.  These should be equivalent to the id?.nid file with the highest
number. so they are basically used for convenience only; but it's
perfectly possible that some of the old nid files are in some way broken,
so think twice before reverting to them.

These are basically some proj-data/s07 sources with minimal changes to make
things work in our more modern environment.

The purpose of all this is to generate the id and nid files in the catalog
directories; the nid format differs from what old ARIGFH had by
including all metadata (kennx4, kennx8) and leaving out the padding.
We need this to be able to come up with useful records in arigfh.unidentified.

You'll need a German native speaker to make sense of this -- all original
comments and docs are in German.

The way arigfh was indended to work, each catalog has a read routine, lies.f.
Together with a main from identframe, the result of this can read and match
objects from the source catalogs against a master catalog in several steps.

The basic identification routine is run-ident, a shell script that arranges
all data such that identification can commence.