One major weaknesses with PBCC and PBDLL is in file handling.
Both products have adopted the MS file handling which works but
is inflexible when it come to random access files. It only
supports fixed length fields. Which are major problem when
you have dynamic strings.
What would be nice is if you could add random access file access
with support for variable length fields. The old CB86 had these
functions and made file handling extremely flexible.
Basically the function would be similar to reading and writing
sequential files expect there is one more variable, the record
number. The record would be padded or truncated to fit the
record length.
This method of file handling means it possible to have different
record structure within a random access files such as in B+tree
index system. You can have 10 fields in one record, 2 fields in
others, 1 fields in another which is a dynamic string which is
full record size.
------------------
Both products have adopted the MS file handling which works but
is inflexible when it come to random access files. It only
supports fixed length fields. Which are major problem when
you have dynamic strings.
What would be nice is if you could add random access file access
with support for variable length fields. The old CB86 had these
functions and made file handling extremely flexible.
Basically the function would be similar to reading and writing
sequential files expect there is one more variable, the record
number. The record would be padded or truncated to fit the
record length.
This method of file handling means it possible to have different
record structure within a random access files such as in B+tree
index system. You can have 10 fields in one record, 2 fields in
others, 1 fields in another which is a dynamic string which is
full record size.
------------------
Comment