On Fri, 24 Nov 2006 15:19:05 -0600, Greg Kuperberg <[EMAIL PROTECTED]>
wrote:
On Fri, Nov 24, 2006 at 01:42:41AM -0800, Andi Vajda wrote:
All exceptions are listed in PyLucene's README [3]. Again, if something
in the PyLucene documentation is wrong or misleading, contributions,
corrections and patches are welcome.
The real problem is not that the README file has any mistakes in it,
but rather that it is haphazard documentation. It just didn't connect
the dots for me. You can only expect so much from documentation recycled
from a conference presentation and stashed in a file called README.
again, patches would likely be welcome.
If PyLucene had a wiki, that would give it a better chance to
evolve into good documentation.
probably, wiki's are useful, but you also have to combat wiki spam, and it
takes time to organize them usefully, their not a magic pancea.
But since you mention it, the README file has not been adapted
to Lucene 2.0.
The "Lucene in Action" book [4] is also a great source of documentation
and
samples for using Lucene.
It's a pretty good book, when the authors are not too busy singing
praises to Doug Cutting and Lucene. I don't even disagree with
these praises, but they get in the way.
Almost all samples from this book were ported to
python and PyLucene [5].
Since you mention it, LuceneBenchmark.py is not adapted to Lucene 2.0.
Maybe someone among the 90% of people not yet using PyLucene is going
to step
forward and make some contribution improving on documentation,
installation
or otherwise ?
Of course, you can only expect people to start using it first, and
then contribute.
i've been contributing since i started using it (initial fsdir impl and
first unit tests way, way back when) and its been great to watch as
pylucene has come lightyears forward since then. i think the current state
of installation is unavoidable till patches to gcc aren't needed, which is
just now becoming true. i think the reality is that its cutting edge in
terms of architecture, and it needs good toolchain support, and as long as
that toolchain needs patches its going to be work to install from source.
there are binary builds for several platforms. but to me, the bottom line
is that andy has done a tremendous job of maintaining and supporting
pylucene imo as a one man solo effort part time, and if folks want to make
it easier to install or better documented, they should step up and make
that happen, not whine about it. if you just want something easy to
install there is also xapian.
cheers,
kapil
_______________________________________________
pylucene-dev mailing list
[email protected]
http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
