One thing I am in the process of doing is to move the schema object
output to its own file. The part of this project I am working on now
is to add indexing and the index will have it's own file, the tuples
will have their own file and the schema object will have it's own file
so that I can move loading the schema or creating the schema to be the
"responsibility" of the schema_types and creating the table to a
table_types package.

This would move the insertrec, updaterec, deleterec, and selectrec to
table_types as well. From those methods I could remove any operations
where I have to read past the schema information; that could just be
read from the schema spec file for each table. There would be three
data files. <tablename>_SCMA, <tablename>_PIDX and <tablename> (for
the tuples). Then I can eliminate the byte_start and byte_end
attributes of the attribute class. Also, I could remove record_size
from the schema class because it is never used.

Doing all this might help facilitate unit testing better, of which I
need to do to track down the problem. Thanks for your help!!