Re: [Plplot-devel] State of the Union

[Rafael L. <lab...@mp...>]

Geoffrey,

Thanks a lot for this impressive proposal.  It is gigantic and I confess
that I just skimmed through it (BTW, are you just a single person, or a
collection of individuals responding to the name of "Geoff Furnish"? :-)

I share your opinion about the lack of documents summarizing our discussions
and ideas and find the PIP idea very appealing.  I took a look at the Python
PEP system (http://python.sourceforge.net/peps) and vote for adopting it for
the PIPs.  The PEP format is quite simple and well documented, and the
scripts for htmlizing the source files are available. Let us avoid
reinventing the wheel.  Let us suck it from elsewhere.

As regards location for PIPs in the CVS tree, a better place would be under
$CVSROOT/pip/ and not $CVSROOT/plplot/pip, as you suggested.  This would
have the benefit of avoiding the branch & tag problems in the source tree.

I find this PIP idea particularly interesting because I was envisaging to do
something similar for the AM/LT project.  The work on that branch has been
stalled for a while and I simply forgot its present state (BTW, the same
must be said about the gnome driver, the Debian packages and the Perl
bindings...)

So here is the immediate roadmap: convert your previous post into the PEP
format and give it the title "Index of PLplot Improvement Proposals (PIPs)";
call it PIP #0 and cvs commit to $CVSROOT/pip/pip-0000.txt; rip out the
scraps of text for each proposed project and put them in separate files
pip-xxxx.txt; set up a rudimentary home page under
http://plplot.sf.net/devel to host the PIPs; setup the automatic htmlization
(seems to be as simple as "$ python pep2html.py -i NUM").  [An aside note:
in the PLplot-Docbook project, we use scripts for updating the web site,
although they are not generated automatically (by cron jobs, say).]

These hacks should last a couple of hours.  Having this proof of concept
working would be great.

-- 
Rafael