[SciPy-user] SciPy Journal

[Joe Harrington]

Thoughts:

1. Let's at least consolidate discussion on scipy-users, so we don't
   have 3 threads going.  It's about scipy and scikits as well as
   numpy, so it belongs on one of the scipy lists, and it's more than
   just for developers.  The users are the customers of the articles,
   and potential editors.  In the future, it would be good for a
   cross-poster to identify a single list to contain followup
   discussion.

2. A journal is a significant time and people effort.  While it's a
   laudable goal, we have a *serious* deficiency in the areas of
   release packaging (i.e., installs for various OS releases that just
   work) and user documentation.  Do we really want to divert our
   efforts to this journal while the majority of people interested in
   using the software are still sitting on the sidelines waiting for
   us to get our house in order on these more basic things?  As you
   pointed out, you can submit these articles elsewhere for now.

3. If one-stop-shopping is a concern, we can easily put up a web page
   listing packages and containing pointers to articles that describe
   the algorithms, wherever they are presented.  We can use arxiv.org
   to post articles for free, reviewed or not.

4. I agree that we want our articles freely available to all without
   any subscription.  There are some journals that fit the bill, and
   we could put links to those on the index page of item 3, for those
   desiring the legitimacy of peer review.

5. I agree about the problem of raising the bar.  On the other hand, I
   think raising the bar on some things is desirable.  The index page
   of item 3 could be a place for the community to call for a reviewed
   article on a major package.

6. Creating our own journal has the benefit of making a place for
   community to work, but we already have that with these lists.  On
   the other hand, it sequesters our good work from the wider
   community, which would otherwise see it and perhaps get interested
   in it if it were published in existing journals.

Bottom line: I think we should start with an index page today, and
suggest that people send their articles to places that will allow them
to be re-posted freely on the net (or negotiate that permission in
advance, which I have done for all my research articles).  We can then
consolidate them on the index page and even republish them in the
future, when we have the critical mass for both up-to-date
documentation and packaging and a journal.

--jh--