>Sourceforge is a wonderful resource and community.  Today, NONE of the
tools
>used to make/build/test software in that (sourceforge) world can help you
>when you have to update a CICS COBOL program to fix a bug, or get a patch
>from IBM for the operating system and apply it across 20 LPAR's with no
>disruption of business.
1) Edit the COBOL source in a PDS member.
2) Submit a compile.
3) View the compile output
4) Test the transaction
5) Submit a job to install to production.
All of these features are available today in WebSphere Studio Enterprise
Developer (WSED). WSED uses Eclipse and functions such as submitting jobs
and viewing output are IBM provided plugins. But those same plugins
*could* be created by an open source MVS community.

If that CICS source is currently under the control of a Source Manager,
such as SCLM, CVS, PCVS or Serena Version Manager you probably already
have access to it through your favorite Windows or Linux based IDE.

ANT is a great tool for building and installing software. I currently use
ANT to build java code and install to z/os with an ANT FTP task. There are
currently few ANT tasks available for z/os but I'm sure more are on the
way.

The lines are blurring. Many of the strong tools created in the *nix/Win
world are moving towards operability with a z/os server. IBM is meeting
them halfway with support for various remote protocols.


>CAN something like sourceforge be used to perform controlled maintenance
of
>CICS COBOL and the other MVS projects we must support?  Yes, they *can*,
but
>unless IBM writes it, promotes it, and supports it, no business manager is
>going to allow you use it.  Too much business and career risk.

IBM is writing it and promoting it. What is internet service delivery?

We're throwing around the term "Sourceforge". Let me make a distinction
here. I think the sourceforge model could be a stronger method for
collaborative development of public domain software like that found on the
CBT tape. I'm not suggesting a public "sourceforge" be used to maintain in-
house programs. But there may be benefits in adopting a private
sourceforge -- strong version control, communication, continuos
integration, more flexibility in programmer and admin tools.

>Just NOT for their installed base of batch programs and transaction
>processing.  And it *is* hard to use CVS for anything when you can't
compile
>CICS COBOL and link to a standard PDS LOADLIB for testing and QA and
>Production.  There aren't any "promotion" tools available for standard
>change control procedures

There are products today that drive remote z/os builds from a workstation
or remote server. Eclipse supports SCLM (pretty standard promotion tool)
for source maint and builds with ANT. How about Serena ChangeMan? I'm sure
there are others.

>When z/OS Unix Services is indistinguishable from a standard GNU linux
(and
>not just posix) environment AND supports appropriate interfaces to current
>MVS technology (i.e., transparent access to MVS/EBCDIC files and data from
>all utilities), then maybe there will be a need/desire for a sourceforge
>equivalent of CBT.

There's an assumption that all these new tools must execute on the z/os
box. I don't think so. z/os only needs to provide interfaces to data and
tasks. For example, we run our CVS servers on Windows but z/os jobs can
access the data through a CVS client running in z/os Unix Services.

>Until then CBT does the job.  It isn't broken, so
>there's no need to fix it, much less replace it.

I agree that the CBT has served us well and will continue to do so.
There's nothing stopping someone from starting a new MVS sourceforge
project, not as a replacement but as an adjunct to CBT. The sourceforge
project could automatically distribute to the CBT, providing both choices
for distribution. If the new approach is valuable, it will prove itself
out.


-Rob

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to