Stephen, I would be very grateful if you could somehow make you "value added" code available, one way or the other. It sounds like this could be useful stuff to other people. Besides, and that touches upon you other point, I found that "example code" is really a great way to learn using a package, especially if one's programming experience is not very thorough. Just seeing how other people have solved problems is very useful (take NumPy, for example). And while PyTables has very good documentation (thanks to the authors!), it is true that there is not yet a very large code base to be found in the wild that one can study, although the software seems to be used by quite some prestigious institutions. Unfortunately at the moment I would belong to the group of people who would profit from other people's work, rather than the other way around. However as soon as I think the situation has improved I would certainly contribute.
I don't have a strong opinion on the h5py / PyTables point. But I am very satisfied with, and grateful for, PyTables. It has served me very well over the years, and the authors have show great commitment, which is worth a great deal. And so far in my work I haven't come across any advantages in using h5py over PyTables (especially now that PyTables also supports fancy indexing for arrays), but they may exist for other situations. In any case, I am glad that h5py exists; I think diversity is a good thing in the long run; and the project don't seem to have the same goals, so they may complement each other. Cheers, Gabriel On Tue, 2009-05-12 at 21:36 +0200, Stephen Simmons wrote: > - Are there opportunities to create "value added" packages giving > higher level functionality? For example I've written disk-based sort, > merge and split routines for tables too big to fit in memory, > iterators > that return chunks of tables as numpy recarrays, iterators with hooks > for progress bars, etc. I don't recalls seeing any community > discussions > about contributing back to the PyTables ecosystem. > > - Can the audience for PyTables be increased by a cookbook with > recipes > combining PyTables with numpy? Speaking for myself, it took many > months > for me to realise my "for row in tbl" code could be parallelised with > the idiom "arr = tbl.read[100:1000]". In other words, to see PyTables > as > a super-flexible persistence layer for numpy. Especially as a newbie, > I > would have a benefitted from more examples to get me thinking. ------------------------------------------------------------------------------ The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your production scanning environment may not be a perfect world - but thanks to Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700 Series Scanner you'll get full speed at 300 dpi even with all image processing features enabled. http://p.sf.net/sfu/kodak-com _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users