cobolforgcc-announce — Low volume list for official announcements about COBOL for GCC

[Cobolforgcc-announce] Status update on cobol compiler cobolforgcc

[tim <te...@me...>]

Hi,

Sorry this update is somewhat belated. I have been working hard on the
compiler. 

Total lines of code is up to 17,000 or so now. The bits that are done
are 
- The lexical analysers for the preprocessor and main compiler,
- The preprocessor parser and implementation of the copy and replace
verbs,
- The main compiler parser, and
- The main compiler abstract syntax tree (AST) handling (validation and
cross referencing). 

Of the 17,000 lines, about 3,000 is 

This is all for a subset of the COBOL nucleus, part from the COPY and
REPLACE verbs which are part of the "source text manipulation module".

The next step is the code generation, which involves hooking into GCC. I
am also looking at using LLVM for code generation.

The web pages have not yet been updated. That's on my TODO list.

So far I am happy with the decision to code the compiler in Lisp. I am
seeing massive productivity benefits compared to C, which I used in the
earlier version. The lines of code ratio is about 3-4:1 for C versus
Lisp, though it varies depending on the problem being addressed. The AST
code was only about 2:1 for example.

The GCC interfacing code will have to be in C, but I will keep that as
small as possible. LLVM requires C++ I believe, same deal there.

The next milestone is the code generation using GCC4.3. This will
probably take 1-2 months.

Tim Josling

[Cobolforgcc-announce] Status of the COBOL For GCC project

[tim <te...@me...>]

Work on the project stopped for a number of years, for good but
uninteresting reasons, but I have recently started work again. 

In fact I left my job recently to work on the project. I took a holiday
first and I'm now working full time on the project. I plan for work on
the project for at least six months and I think that will suffice to get
a substantial part of the work done. 

My plan is to get a minimal subset running and then put it up on
sourceforge, at which point I will be inviting others to help.

Our first attempt produced 80,000 lines of code and I would estimate we
were about 30% of the way there. I have started again with a different
approach using LISP as the main language for the compiler front end
instead of C. The interface to the GCC back end will still be in C but
that will only be a few thousand lines. The runtime will be in a
combination of C and a subset of COBOL. With this approach I think the
whole thing will be less than 70,000 lines of code. Quite a bit of the
code fromt he first attempt is reusable.

Expiriments I have done indicate that LISP programs require only about
10-20% of the lines of code of a comparable C program with almost
proportionate productivity improvements and reductions in bugs, and with
similar performance. LISP is not just for processing lists; it is a very
powerful programming language and good open source implementations exist
(eg SBCL, GCL, CMU Lisp). Some of these run on Windows as well as Unix
variants. There are a lot of good books, documentation and other
training materials about LISP available on the internet.

The opportunities I see for people to help will be

COBOL Skills:
* Test your COBOL programs and submit bug reports.
* Work on the some of the runtime routines.

Lisp skills
* Work on the compiler front end (parser etc).

C Skills
* Work on the runtime routines.

C Skills and GCC Internals skills
* Work on the GCC back end interface.

As this is a Free Software Foundation project it will be necessary for
anyone who makes more than a very small contribution to legally assign
the copyright to the FSF, and to get a waiver from their employer or
school or college as well.

I will issue more updates as I progress, at least on a monthly basis.

Regards,
Tim Josling

[Cobolforgcc-announce] Open Source COBOL Update

[Tim J. <te...@me...>]

Sharing Code
------------

We now have a project 'COBOL-UTILS' for common components of the three Free
COBOL projects. Initially the project will be used for a common version of the
NIST test suite. Later we intend to put major subsystems such as sort and
report writer into this project. 

COBOL For GCC and Open COBOL projects have been talking about combining
efforts. In the meantime competition is keeping us all working hard.

Rumor File
----------

I keep hearing rumors that a major software vendor is considering making its
COBOL compiler Free Software or Open Source. This would be a great development
particularly if the compiler were written in a mainstream language, rather
than assembler. 
One thing that gets in the way of companies releasing software as Free
Software is that often proprietary software is in pretty bad shape internally
and needs considerable renovation before it can match the coding and
documentation standards of other free software.

Open COBOL
----------

Keisuke Nishida (kni...@co...) has released version 0.9.6. At the
moment the focus is on making the software more robust and usable rather than
adding lots of new features. Release 0.9.6 has many bug fixes and improvements
to configuration. It has support for internationalisation, using GNU gettext.
I get the feeling that the pace of work on Open COBOL has slowed down
slightly, which is only to be expected given the frenetic pace of the previous
few months. 
Tiny COBOL

Activity was mainly focussed on bug fixes and improved support for different
environments.
Release 0.58 is now available. It includes better support for comments and
debugging lines, concatenation, inspect converting, some support for
declaratives. The compiler is now a lot easier to use, due to better default
options. TC now supports embedded SQL statements. Work on Tiny COBOL remains
very active.

COBOL For GCC
-------------

I now have the CALL verb working. The support at the moment is for
COBOL-to-COBOL calls. I will shortly add supports for COBOL to C and C to
COBOL calls. Once this is done I will be packaging the software up for my
first release. The compiler now contains a sufficient subset to allow useful
programs to be written, and I plan to write all future runtime routines in
this subset.

In addition, I converted the compiler from version 3.1 of GCC to version 3.2,
which required major changes. I had to totally rewrite the memory management
routines.  A major bug hunt was also needed because all the code optimisation
stopped working. That's all fixed now. 
Open COBOL generates C code, whereas COBOL for GCC integrates into the GCC
code generation back end. Tiny COBOL generates X86 assembler code directly. I
have found that integrating into GCC has created a lot of extra work, although
in the end I think the result will be a better compiler.

I am still working hard on COBOL for GCC. 

My other 'hobby' is managing my investment portfolio. Fortunately I had read
enough investment history to recognise the US markets of a couple of years ago
as a bubble of historic proportions, so I got into cash in time. One good clue
was when people started to say "This time it's different". Another was when
people who had never invested in their lives started asking for stock tips. At
the time of writing, I still think things still have a fair way to fall,
especially the NASDAQ, but the downside risks are reducing. To give some
perspective on how far markets can fall, remember that the Japanese market
fell from about 35,000 to under 10,000. They don't always stop at fair value.
I find the combination of psychology and technical issues involved in
investing to be quite absorbing.

For me the free software world is a refreshing change from the lies and deceit
so often seen in business and in politics. 

Tim Josling

[Cobolforgcc-announce] Status summary

[Tim J. <te...@me...>]

I have more or less finished my DBase to Java conversion project
which sidetracked me for about five months, and I've started
intensive work on the COBOL4GCC project again.

While I was away, Ted Seward wrote a lot of runtime code and test
cases:

- display
- move edited
- conversion between binary and decimal
- stop statement
- 128 bit arithmetic routines

Steven Ellis wrote the 'string' runtime routines.

Daniel Ardison has written about 1/2 of the runtime for the
inspect verb.

Matthew Vanecek also contributed a cobol.el file. This provides
cobol source editing facilities under the emacs text editer.

I also kept the code up to date with the GCC weekly snapshots.
The current cvs code supports the 22 November snapshot.

Current work under way is:

Tim Josling - finish the code generation for the COBOL nucleus
subset to be used to write most of the compiler runtime and of
the phase B compiler (ie in COBOL).

Daniel Ardison - complete the inspect runtime (quite complex).

Steven Ellis - complete test cases for string runtime.

Ted Seward - currently on leave due to work commitments.

Upcoming tasks include unstring and sort, and interfacing to gmp
for infinite precision/size arithmetic.

Could I suggest anyone who is developing for Cobol4GCC to
register at sourceforge. That makes it a lot easier to track
tasks etc. For the same reason I prefer to have any emails (other
than with code attached) sent via the cobolforgcc-devel mailing
list. That way there is a history of specs etc. If anyone objects
to my copying correspondence to the cobolforgcc-devel mailing
lists please let me know.

Tim Josling

[Cobolforgcc-announce] FW: 2-week reminder

[William M. K. <wm...@ix...>]

From Bill,

For those of you with various "issues" (e.g. Public/private, etc) that
want them on the Agenda for the upcoming J4 meeting, please note the
following.

FYI,

A) "input" documents that do not actually propose a change PROBABLY will
not be "time-limited" if they aren't submitted by the two-week deadline.

B) IMHO, there is ALREADY too many papers available for J4 to get thru
them all in 5 days (the length of this meeting) so papers submitted now
MAY OR MAY NOT be discussed at all.

C) The J4 "goal" (expectation?) is have the OUTPUT of the December
meeting be sent to ISO for a FCD (Final Committee Draft) review.
Therefore, (again IMHO) if you want a CHANGE to the draft and you are
just raising an issue, then you MUST get it discussed at the Oct meeting
to meet the draft deadline. (This would allow someone to draft a change
for approval at the December meeting.)



-----Original Message-----
From: <snip> Don Schricker
Sent: Tuesday, September 12, 2000 10:33 AM
_____
J4,

Next Monday (September 18) is the deadline for submission of documents
before the 2-week deadline for the October meeting.

Don

Don Schricker;  Massachusetts, USA;  Don...@me...
J4 Chairman
MERANT manager of language standards

[Cobolforgcc-announce] Fw: IBM "drop service date" for COBOL for MVS & VM

[William M. K. <wm...@ix...>]

FYI,
  (for those of you "interested" in IBM mainframe COBOL variations - and
who haven't already seen this yourself)  I just posted this in
comp.lang.cobol

"William M. Klein" <wm...@no...> wrote in message
news:<8p32cr$ht8$1...@sl...>...
> FYI,
>   Today  IBM announced a Dec 31, 2001 "drop of service" date for IBM
COBOL
> for MVS & VM.  (Usually such announcements are for the US - check your
own
> "support dates" if you are in some other country.)
>
> This product is "replaced" by IBM COBOL for OS/390 & VM.
>
> See:
>  http://www.ibmlink.ibm.com/usalets&parms=H_900-220
>
> (This announcement also includes "drop dates" for various other
> product/releases such as TSO/E, IMS, DB2 etc.  It also tells what
> product/releases replace them)
>
> It *might* seem as if this means that there will be ONLY one
product/release
> of COBOL available for the "MVS" environment for the first time since
the
> early 1990's - HOWEVER, I have every reason to believe that there will
be
> another announcement this month (that hasn't been made yet) of a "new"
> release (version?) of COBOL for OS/390 & VM.
>
> Keep your eyes open (and I will "post" an announcement when I see it).
>
> --
> Bill Klein
>     wmklein <at> ix dot netcom dot com
>
>

[Cobolforgcc-announce] FW: CD 1.9 on the web

[William M. K. <wm...@ix...>]

FYI

-----Original Message-----

CD 1.9 is available on the NCITS web site. See,

	http://www.ncits.org/tc_home/j4.htm

Annex B shows as Annex A.  This is corrected in the printed version and
will be fixed later in the PDF version.

Don Schricker;  Massachusetts, USA;  Don...@me...
J4 Chairman
MERANT manager of language standards

[Cobolforgcc-announce] "Whither" MERANT COBOL (was: MERANT licence fees

[William M. K. <wm...@ix...>]

FYI -
  Two posts that I just did to comp.lang.cobl

From a MERANT press release published today,

> I found the "full" MERANT stuff from today.  It is at:
>
>  http://www.merant.com/resource_center/investors/press2.asp
>
> Of particular interest to this group (I suspect) is the following,
>
> "Cobol revenues decreased at a faster than expected rate, and this
decline
> could continue. This phenomenon was experienced across the industry,
as other
> enterprise software companies have recently reported shortfalls
associated
> with mainframe revenues. We will be reducing our cost base in Cobol to
ensure
> future profitability for this revenue stream."
>
> PERSONALLY, I think that last sentence says a LOT about what you can
expect
> in the future for both NetExpress and MainframeExpress.
>

    ***


"August 22, 2000 04:12 AM

LONDON, Aug 22 (Reuters) - British software tools supplier Merant Plc
MRNT said on Tuesday it would begin a share buy-back programme and
consider a number of strategic alternatives after reporting a first
quarter loss.
 ...
 "The company is considering a number of strategic alternatives to
maximise shareholder value," Merant said in a statement.
 ...
Merant is shifting its focus from traditional Cobol (Common Business
Oriented Language) software products to its new Egility e-business
software -- which allows companies to mix computer applications
programmed in different languages.

But it has been unable to replace services as fast as expected, with
demand for its traditional products vanishing at a rapid pace.

Merant's total first quarter Cobol revenue fell 24 percent, and Cobol
licence fees dropped 43 percent. But its Egility business grew seven
percent and now made up nearly two-thirds of its business, Merant said.
"