Re: [open-cobol-list] OpenCOBOL release 1.0

[Tim Josling]

On Sat, 2007-12-29 at 23:56 -0500, David Essex wrote:
> Tim Josling wrote:
> 
>  > As for COBOL for GCC, I left my job recently and have been
>  > working full-time on COBOL for GCC since 3 December 2007.
>  > I plan to work on it for six months full-time.
>  >
>  > I have started from scratch again, and this time I am writing
>  > it mostly in Lisp. The GCC back end interface will be in C
>  > and so will the run-time routines that cannot be written in
>  > COBOL.
>  > ...
> 
> Given the advanced state of OC and your knowledge of the GCC
> back-end, is there any particular reason why you choose not
> to use OC ?
> 

The main issue is that if the compiler is going to be incorporated into
GCC, then a copyright assignment to the FSF is required. This also
requires a disclaimer from the school/college/employer. This applied to
all contributors beyond the trivial level.

I don't think this is possible to arrange for all the OC contributors.
If people have changed jobs or left college, we would need to go back
and ask for disclaimers which are hard to get at the best of times.

Also, my experience with my first attempt at cobol for gcc was that
linking in to the gcc back end is very difficult, mainly due to the huge
amount of churn in the code base and APIs. So I want to have as little
code as possible in the same process as the gcc back end. 

Without looking at OC in detail, I would confidently predict a major
upheaval to get it to comply with the gcc way.

At the same time I don't want to go via C code, because of the loss of
debugging and optimisation information. So I am going to create a
special binary interface that will be fast and pre-parsed in effect.

There is no reason why other cobols can't use the same interface so all
is not lost.

There is still a lot of scope to collaborate eg with utilities such as
sort/merge.

Regards,
Tim Josling