Re: gnubol: Anyone actually working on Cobol

[Tim Josling <tej@melbpc.org.au>]

I was very much in the "let's start coding" school, until I wrote
the preprocessor. I now think the choice of parser tool is quite
imnportant. 

For me the choice is down to pccts and btyacc. I will be deciding
what I myself will use over the next few days. After my weekend's
research I am leaning back to btyacc and flex.

My strategy will be as follows:

* Do nucleus level one first and gradually expand.

* Try and get the key things right so it can then be
incrementally grown. By whoever wants to contribute.

* Generate code using GCC's back end from day one ie cobol to
machine code, not cobol to c. The GCC back end supports every
platform known to man so this is portable enough.

* Function first, efficiency later. Serious effort to be standard
conformant.

* Target IBM mainframe compatibility but with native data types
(ie I am not going to emulate big endian binary arithmetic on
Intel).

A project like this will always get its fair share of blowhards
with lots of ideas but who don't contribute any code. As linux
Torvalds said 'show us the code'!.

Much of the compiler has to be in C for various reasons and a lot
of those who would contribute only know COBOL, so it is important
to get a basic cobol thing going so those people can contribute
by writing cobol eg for a lot of the verb support routines.

Tim Josling

Glen Colbert wrote:
> 
> I've been following this thread off and on for over two years now.  Over
> this time, I still can't get a clear picture of what it is this group is
> trying to do.

--
This message was sent through the gnu-cobol mailing list.  To remove yourself
from this mailing list, send a message to majordomo@lusars.net with the
words "unsubscribe gnu-cobol" in the message body.  For more information on
the GNU COBOL project, send mail to gnu-cobol-owner@lusars.net.