[Info-Ingres] Ingres Janitors

[Mark R. Winston]

Roy Hann wrote:
[snip]
> I want to get the UK Ingres user association to do more to promote community 
> development projects.  In my new role as acting chairman I've asked Marty 
> Bowes to be our "Community Projects Czar".   

Seems appropriate, especially since he tackled the compilation of INGRES
and stuck with it despite little substantive help from Ingres Corp.

I'm really taken with your
> suggestion above because most of us have day jobs or other limitations that 
> would prevent us taking on a major project but as you point out there is 
> probably lots that we could do for our own amusement or convenience, and in 
> principle it could relieve other developers.

Yes, and I think it would serve to spawn a deeper level of expertise of
existing INGRES users as they are exposed to the information garnered
and published as a result of the effort of the Janitors and the Janitor
Mentors.

> There are a number of things that stand in the way at the moment.  I am also 
> curious to know how Ingres Corp "passively killed" your first attempt.  Can 
> you explain in more detail what you tried and why you think it didn't work?

Ingres Corp. uses some nifty tool, for which I've forgotten the name, to
parse the INGRES source code to look for issues, e.g. orphaned
structures, syntax errors, compilation warnings, etc.  The idea was to
take subsets of the report produced by said nifty tool and attempt to
tackle the "simple" issues.  Myself and another non-Ingres Corp. member
were going to take first pass at this and work out the details for
receiving bugs, coding a fix, testing said fix, and then submitting the
fix to Ingres Corp. for integration into the main code base.

There were concerns by Ingres Corp. that exposing these bugs publicly
would perhaps encourage their exploitation by crackers, damaging the
reputation of INGRES technology, and consequently, Ingres Corp.
Opinions on how to minimize this risk were voiced but an official
decision of how Ingres Corp. would be involved in facilitating this core
INGRES code bug list was never made, either for or against the approach.

Additionally, even enhancing some of the Open Source connectivity tools,
e.g. the Perl INGRES DBI, is difficult because traversing the INGRES
code base without being able to reference its specific layout or
consulting a knowledgeable party is frightening.  Karl has done a nice
job in a presentation of providing a "Gentle Introduction" or "101"
discussion of how to approach the INGRES code base but as far as I'm
aware, this is the *only* thing of its kind that's publicly available.

Tackling Perl DBI enhancements is one type of challenge where we need an
idea of where we should be looking to determine how INGRES performs X
function; fixing trivial or non-critical bugs in the INGRES code base is
another thing entirely, both from a standpoint of general complexity,
and perhaps proximally, how to identify such bugs to begin to attempt to
fix them.

I think it's a problem with a resolution, and although it's great to see
Karl volunteering to assist, I don't think it should be on Karl's, or
your shoulders, to introduce the INGRES code base to the Open Source
Community, or to foster the Community's involvement.  There will be some
bridge that will require the knowledge of INGRES Engineering to cross
and it would nice if Ingres Corp. would acknowledge that A) this is
necessary and B) they aren't the only ones with the skills to bug-fix or
even enhance INGRES technology.

If we had the likes of Karl, you, and the others (who know who they are)
providing a list of bug fixes and/or enhancements to someone like Marty,
who in-turn was distributing said tasks to persons such as myself, we
might all be surprised at how much we're able to accomplish, and how
quickly we gain the inertia required to refine the process and add new
members to the project, promote existing Janitors to Senior Janitors,
Junior Mentors, and so on, and so forth.

But perhaps I'm being too idealistic :)