[Zope-dev] PostgreSQL and Zope (was Re: [Zope] Summary of DB Discussion)

[Ross J. Reedstrom]

On Tue, Nov 09, 1999 at 06:31:19PM -0600, Karl Fast wrote:
<snip>
> 
>  - There are DA's for MySQL and PostgreSQL, the two open source databases
>    (yeah, yeah, MySQL isn't truly open source--although one of the older
>    version is). They work, although both are still in development (is this
>    active?) and some people have had troubles getting them to compile. 
>    This seems especially true of the MySQL DA judging from pleas for help
>    sent to the zope list, although this may be more of a reflection of
>    MySQL's popularity over PostgreSQL.

Since there have been a number of posts recently in the "works for me"
catagory, I like to think it reflects fewer problems. ;-)

> 
>  - DC is looking for someone in the zope community to take ownership of
>    the MySQL DA (what's the status of the PostgreSQL DA?)
<snip>
> 

Other have addressed some of these points. Since the PostgreSQL DA is
on my Zope member page, I thought I should comment on the current state
of development.  I recently took some vacation: just before I left,
the original author of the ZPyGreSQLDA resurfaced in email. Since I've
been back, I haven't caught up at paying work, so progress on the DA
has suffered. The biggest problem has been that my project at work has
gone completely ColdFusion (Zope wasn't there early enough), although we
do use PostgreSQL. So I haven't been able to justify much daytime work
on it.  The biggest problem is probably that I've written many more
lines of Tcl than Python, so I'm climbing more than one learing curve
at a time. (What I could really use is some good open source time and
project management software. Anyone have any recommendations? And what
happened to the ZopeCalendar <-> PalmPilot conduit project? )

My experience with the PostgreSQL developers and mailing lists are very
similar to the Zope lists: lots of clueful, helpful people (some of them
are the same people ;-) Bugs exist (what software is perfect?), but get
diagnosed and fixed very rapidly. Features are added (and limitations
removed) on a weekly basis. I get the feeling that the pgsql project has
turned a corner in the last year or two: the code base has been cleaned
up, making it easier to understand, and contribute code. New developers
are being attracted, and quickly becoming contributers. It feels like
it's reaching critical mass.  The pgsql-hackers list reminds me of the
linux-kernel list from about '92 or so, in that respect.

If you haven't looked at PostgreSQL since it's Postgres95 days, it's
definitely time to look again.  To get this back on topic for the Zope
lists, in particular, I think DC should take a serious look at it:
the licenses are compatible (pgsql is Berkeley software, after all) and
the corporate attitudes are compatible.  (The core developers recently
formed PostgreSQL, Inc. and are learning how to be a business, as well as
a project: could probably use some pointers from you folks ;-) As another
poster mentioned, pgsql is an EORDBMS: Extensible Object Relational DBMS,
and not only because you get the source. Building a ZStore for the ZODB on
top of PostgreSQLs object features should be easier that relying totally
on relational calculus, and might reveal some interesting synergies.

If I could unleash the crystal ball (rose tinted), I can see future
where Zope and PostgreSQL are the defacto next generation tools for
web applications, taking over from perl and MySQL as more critical,
core business functions move to the web.

Boy, now that I've expressed all that in words, I better get cracking on
that DA, shouldn't I? Now that 2.0 seems to have settled down a little,
(and 2.1 is in beta) any one want to take a crack at helping? Patches
always welcome!

Ross
-- 
Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> 
NSBRI Research Scientist/Programmer
Computer and Information Technology Institute
Rice University, 6100 S. Main St.,  Houston, TX 77005