I was wondering what the possibilities were for have the option of
byte-compiling the gap libraries. It is somewhat annoying to have to wait
each time a new routine is called upon to have gap interpret a text file.
In addition, of course, byte-compiling would also provide opportunities to
optimize the code somewhat leading to potentially faster execution times.
Of course, the optimal solution for time would be to have a dynamically
linked library in the machine code of the computer it runs on. Of course this
is a much more complicated task and makes the portability of GAP more
problematical. However, it is an attainable goal, I think.

I have looked into the parsing routines in the source code but have not had
enough time to see exactly how the parsing process proceeds to determine how
easy or hard it would be to include byte-compiling. It should just be a
matter of saving the internal format of the code and data structures
associated with it to disk and then recalling them when the library is needed.
This is exactly the manner in which emacs operates. And in emacs you even
have a choice as to whether or not you desire to have the libraries byte-
compiled or not and having both around does not pose any particular problem.

Anyway, that's just my two cents worth for improving the efficiency of what
I think is a fabulous system.

--John Neil

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
John Neil, Graduate Teaching Assistant           e-mail:  neil@math.mth.pdx.edu
Mathematics Department                         NeXTMail:  neil@dehn.mth.pdx.edu
Portland State University
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=