[Zope-DB] Re: R: R: ZODBCDA for Python 2.3.3
Matthew T. Kromer
matt at bane.mi.org
Fri Apr 2 13:21:28 EST 2004
Please remember that just because software is released to you for free,
> that does not mean that it has no cost to you. When using OSS, you must
> be prepared to invest the time needed to integrate it. That may mean
> getting your hands dirty and writing some code, or paying someone else
> to do so.
and I'd like to point out that (being the last person who compiled
ZODBCDA) all it takes is visual c++. I am pretty confident you could
even get it working with cygwin and gcc, but you'd have to be more of a
windows person than I.
I say this because, whereas people enjoy using no cost software, they
are often reluctant to vest any resources in maintenance of same.
Now as a practical matter, ZODBCDA should die -- its entirely SWIG
driven from what I think is ODBC 1.0 headers. As such it is not what I
consider to be optimal software. This is a reason I can endorse
mxODBC -- not because I have a lot of experience using it (I dont) but
because it's actively maintained by and supported through nominal usage
fees. I know for a fact that the code it contains receives a lot more
care and attention than ZODBCDA does, which had ZERO attention in the
past four years, with the possible exception of me compiling a Python
2.1 version of the binary.
The dilutive nature of receiving the benefits of open source software
often means that the value flows are often unidirectional (from the
creator to the public) and not bidirectional. Each individual consumer
feels that their incremental value is so small as to be negligible, and
never contributes even that negligible value back, either to the
software's origin, or elsewhere in a community value chain. "Value"
here often doesn't need to be money -- people can derive value from
other sources -- but even positive feedback tends to be rare, and well
We can attempt to monetize value at times, since money is a reasonably
good exchange medium for value. To the best of my knowledge Zope
Corporation (or its prior incarnation, Digital Creations) never
received any supplemental funding from any source for ZODBCDA. This
would suggest that because the individual user benefit was perceived to
be zero, the aggregate value was also zero, and as such, development
was stopped on a product that was valueless.
As someone who wrote a different database adapter (DCOracle2) and other
than a few "thanks" from folks in the industry I never saw anyone come
back and say "We'd like to buy a support contract to help fund
development." You might imagine that this removes anything but
self-serving reasons for developing such products. Certainly one can
understand that it is difficult to envision a sustainable business
model that revolves around working for free, when one's expenses are
For those most in need of ZODBCDA and Python 2.3.3 on Windows, I
strongly suggest investing money in Visual C++, or time and energy in
cygwin and gcc to get a functioning build. Because the individual cost
of either of these appears to be greater than that of using mxODBC, it
would seem to me that migrating to mxODBC is the more rational thing to
do from any individual user perspective.
More information about the Zope-DB