[pymvpa] Upcoming PyMVPA 0.2.0: Request for help

[Michael Hanke]

Hi,

as already mentioned, we will have a new release next week. I spend some
time over the last two days to improve/extend/polish the documentation.
Although still insufficient in several places the usability should now be a
lot better than before.

However, I'd like to request your help. Some parts of the docs say
'to be written' and others are just stubs. I think some of them can be
significantly improved with minimal effort, when done by the people who
wrote the corresponding code.

For example, we have a list of classifiers. I already started to add
references to papers (and even wikipedia pages) to explain what the
respective classifer does. It would be extremely helpful if the author
(or somebody else who is familiar with the algorithm) could add a few
sentences (or links) to explain why it was worth to implement the
classifier -- although we already had kNN ;-)

Additionally, the 'Feature selection' section is basically empty and
_any_ contribution will improve it.

Since a few days ago, the manual is checked with doctests
('make testmanual'). So, if you want more examples, snippets, you can
add them now, knowing about the decreased likelihood that they will get
out of sync.

I also added a new file ('scenarios.txt') that is supposed to contain
outlines of analysis scenarios, i.e. schemes/sketches/drafts of analysis
workflows. Currently, it only contains a searchlight demo. If you want
to share your 'best practices' -- this is the place to add them.

More examples where help is helpful:

  * We have an (empty) glossary. If you think some of the PyMVPA
    terminology is confusing: Tell me -- and we will have a non-empty
    glossary ;-)

  * Is there some literature that _you_ would recommend to people working
    with PyMVPA?


I'd like to mention that your contributions will not be hidden in some
PDF on the web. We recently switched to python-sphinx to generate _all_
docs. The PyMVPA website will be completely auto-generated from the docs
and will contain all your contributions. It provides an online and
offline full text search facility (in addition to the access by a keyword
index), that will make the docs a really useful reference. This docs will
also be shipped with the binary packages.

The new website will go online in-sync with the new release.

If you want to help, just start editing the .txt files under doc/. If
you do not want to fiddle with Git, download a snapshot ...

http://git.debian.org/?p=pkg-exppsy/pymvpa.git;a=snapshot;h=refs/heads/master;sf=tgz

... and send diffs.



Thanks in advance!


Michael


BTW: If you want to take a look at the new docs, just do a

	make website

in the root of the source tree and point your browser to

	build/website/index.html

(you need latex and python-sphinx for this)


-- 
GPG key:  1024D/3144BE0F Michael Hanke
http://apsy.gse.uni-magdeburg.de/hanke
ICQ: 48230050