Hi Jochen, Thanks for you mail. I'm copying this reply to [email protected] so that other developers get the benefit of the discussion.
There are a number of statistical procedures still missing. You might like to consider the posthoc tests for the ONEWAY command, or oblique rotations for FACTOR. We've never actually had any complaints about PSPP being numerically unstable. That said, we could always use advice about what could be done better. For example, it is well know that variance is much more stable when calculated with 2 passes though the data than with one. However in PSPP we prefer to make as few passes as possible. So any advice from an expert on when and where the two pass method should be prefered is welcome. So far as coding standards are concerned, these are available online at http://gnu.org/prep/standards/ Please feel free to post a message to [email protected] or drop by at #pspp on irc.freenode.net Regards, John On Wed, May 18, 2011 at 11:38:37AM +0100, Jochen M??ller wrote: > Hello, > > I am a young mathematician currently engaged in PhD studies in numerical > analysis and would like to contribute to the PSPP project to gather > experience. I am totally new to open source projects so I don't know > where to start... > > Have you got suggestions what I could do, maybe implementation of a > special statistical procedure that is still missing? On the old project > homepage you asked for help by a numerical analyst to check possibly > unstable/unsafe algorithms, is this still a current problem? Are there > coding guidelines so that my code will fit in without much changes? > > Kind regards, > > Jochen -- PGP Public key ID: 1024D/2DE827B3 fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3 See http://pgp.mit.edu or any PGP keyserver for public key.
signature.asc
Description: Digital signature
_______________________________________________ pspp-dev mailing list [email protected] https://lists.gnu.org/mailman/listinfo/pspp-dev
