Re: Open Source RPG ?

Pat Barber wrote:
> Very well said David. I have often wondered just exactly what
> business in their right mind would actually use "open source"
> for anything other than "just looking".

Actually, that's not my point.

On the iSeries, we are dependent on the system staying the same ... or
at least being 100% compatible with previous versions.

If the compiler suddenly became variable, then there's a chance we
couldn't depend on it.

Don't get me wrong ... I think open source is a wonderful thing ... it's
encourages collaboration and peer review.  I depend on open source
software heavily.  ALL of midrange.com runs on open source software.

There is absolutely no reason a company shouldn't use open source
software ... so long as they are confident in a) Their ability to self
diagnose and support it (which isn't bad, if they have people that
understand the mechanism and have the source), b) their vendors ability
and willingness to support the product, or c) the community that uses,
maintains, and enhances the project.

Unfortunately, there's always the chance that someone will take a
version of an open source project, fork it, and start developing their
own version.  That's fine if that person is going to support that
version on an ongoing basis.

Redhat, Suse, etc, are doing a fine job with that ... but they largely
depends on the community to find & fix bugs.

With RPG, we're all dependent on IBM doing the job of making sure
everything works the way it's supposed to (and a fine job they do, too).

If RPG were open sourced ... then we would be dependent on the community
finding & fixing bugs (I'm sure IBM would be part of that community),
but could we be assured that the version of the O/S compiler that is
released has been fully tested by IBM to ensure it's compatible.

I guess my point comes down to assurance of quality.  If the IBM would
continue to certify the compiler the way they have been, then maybe
making it a O/S project isn't that bad an idea.  But I still have my
doubts that the community would be big enough and active enough to
provide the necessary level of peer review and collaboration.

A few things to note ...

1. IBM hasn't released the source modifications it made to Apache2 http
server for the iSeries, even though Apache http server is an open source
project.

2. Last time I checked, Sun hadn't released the source for the Java
compiler.

> I'm sure this whole open source thing has some great value,
> but to actually create a complete system written with a
> language that "might" or "might not" have support sounds like
> insanity to me.

Again, it depends on the development community that is willing to invest
time and effort into the project.

. If the community is big enough, there is no problem.  Look at Linux,
gcc, apache, etc.

. If the community is NOT big enough, then you've got problems. Case in
point: htdig, the search engine I use on midrange.com's archives.  This
is a fine application and does a good job ... but it's languishing ...
and hasn't been updated in years.  The developer community that is
willing to work on it just isn't big enough.

Sorry if I rambled ...

david