[Pydotorg-redesign] Re: More on the white paper idea

[Skip Montanaro]

    Kevin> There is definitely overlap with the SIG pages and much of the
    Kevin> content of various SIG pages would be good for topic pages
    Kevin> (eg. edu-sig -> Python in Education). I'm not sure I like the
    Kevin> idea of replacing the better pages like the Python and Scientific
    Kevin> Computing topic with a white paper and a set of links at the
    Kevin> bottom of that, but Skip may have an example white paper in mind
    Kevin> to use as a model for the organization.

    Kevin>   http://www.python.org/topics/scicomp/

    Kevin> This is definitely an area where it would be good to get feedback
    Kevin> from the "scientific computing" customers about whether the pages
    Kevin> are useful to them.

More in line with what I'm interested in is, what information would a person
considering using Python for scientific computing need to know?  The scicomp
page is essentially a lot of links with very little motivation about why
Python would be a good choice for scientific computing and no case studies.
A white paper in this area would ideally speak about things like

    * How is Python integrated with existing Fortran libraries?

    * Is there a Python interface to LAPACK?

    * How does performance compare to applications written completely in
      Fortran? 

Much of that can be inferred by doing a fair amount of clicking and reading,
but it's not all there for one easy gulp.  In particular, for the technical
guy wanting to plop a three- to four-page synopsis on his boss's desk which
supports his claim that Python is a good choice for this area, there's
nothing.

I'm including Eric Jones on this thread so he can provide some specific
feedback on what he feels a technical/project manager considering Python in
a scientific computing context would look for.  Unless he begs off, please
make sure your replies include him, as I don't believe he's on either list.
Let's focus on just scientific computing for the moment.  I suspect we will
find many of the questions people would ask will be similar across
disciplines.

Eric, the general context in which this thread occurs is marketing Python
more effectively.  Kevin Altis (I think) stated a desire to increase the
usage of Python - however crudely you might measure that - by about 10x in
about two years (the clock started yesterday ;-).  Since there is no
"marketing arm" for Python, that means the website must subsume many of the
tasks of a traditional marketing group.  In concrete terms, a change in
focus of this sort implies a significant change in the organization of the
website.  I suggested one way we could organize part of the site is to have
a series of topic-specific white papers.  In Kevin's note, he focused on one
of my suggested areas, scientific computing, and pointed to the current
/topics/scicomp page on python.org.  If we can solve the problem of
developing a good white paper for Python+scientific computing, I think we
can generalize much of what we've learned to other disciplines.  That's
where you come in. ;-)

Skip