[SciPy-dev] Last week's Skypecon Minutes and request

David Goldsmith d_l_goldsmith at yahoo.com
Mon Jun 29 16:28:29 EDT 2009


Hi, all.  Below are the Minutes of last week's Summer Marathon Weekly Skypecon.  Before that, you'll note in them that the participants decided to fix the Skypecon time to Wednesday's, 19:00 UTC.  Well, I already need to request a "variance" for Wed. July 15 - anyone object strongly to Tuesday July 14, 18:00 UTC that week?  Thanks!

DG

PS: Note that that's three Weds. from now - the next two, July 1 and July 8, the Skypecon will be at the designated time, 19:00 UTC; please email me any items you'd like to see on the agenda.  And now, without further ado, I give you... the Minutes!


Numpy Doc Summer '09 Marathon
June 24, 2009 Skypecon Minutes

Start: 19:00 UTC
End: 20:00 UTC

Present:
David Goldsmith, Secretary
Stefan van der Walt
Ralf Gommers
Pauli Virtanen

Discussions:

0) Clarifying the Doc Standard vis-à-vis classes and their constructors, attributes, instances (created by numpy), and constants

    Agreed that, for now at least, c-tors will be documented in the class docstring

    Agreed that, because we might want to group attributes into meaningful categories within their class, we’ll stick with manual generation and annotation of such lists of attributes

    Instances: Pauli explained how things work “behind the scenes,” and it was decided that the way it's currently done is okay. For what that way is, see patch attached to ticket #1146, http://projects.scipy.org/numpy/ticket/1146

    Constants: At first, said it’s moot because we can’t easily modify constants’ docstrings, but then Pauli pointed out that the development version of NumPy has the subpackage numpy.doc.constants, wherein constants can be documented at module scope.  Did not settle on exactly how it should look but were in agreement that it should be a) kept short, and b) use the same sections as for functions where applicable.

1) Two "pre-proof" reviews: technical (“Is it correct?”) and presentation (“Is it clear and as simple as possible, modulo being correct?”)

    Note: the presentation issue is seen as distinct from whether or not the docstring is grammatically correct and free of typos, these latter to be assured by the final proof

    Discussed the “pros” (our labor pool, in general, consists of persons who are stronger in one area or the other, but not both; dividing the review process in this way makes it more manageable and thus less intimidating for prospective reviewers) and “cons” (compounding and complicating the overall process in this way will just push back completion that much further); decided that the pros outweigh the cons, and thus this division should be added to the review process, but, recognizing that these reviews are beyond the stated goal of the Summer Marathon, decided that required changes to the Wiki can be designated as “medium” priority.

    It came up after the meeting that there is still confusion about the meaning of the proposal, and that if the “non-technical” review is as described above, then at least one meeting participants opted to change his “vote,” resulting in a “tie” regarding “endorsement” of the proposal;  we await clarification from the proposer regarding his intended meaning.

2) Adding a FAQ page to the Wiki

    Discussed how recent questions/issues arising on the mailing list and in the Comments sections of the Wiki demonstrate the need for the Wiki to have something like a FAQ page.

    Pauli agreed to create a “mock-up” for dissemination and discussion, which he has already completed and published at http://docs.scipy.org/numpy/Questions+Answers/

3) Meeting time

    Decided that 19:00 UTC Wednesdays is a good time, and that, all else equal, it's probably better to fix the time than have it "float."  Consequently, next Summer Marathon Skypecon is scheduled for:

    Wednesday, July 1, 19:00 UTC

Notes:

    Ralf noted an inconsistency on the Milestones page between the Goal statement and the color scheme, which the Secretary has corrected

    Again, had periodic dropout problems all around.




      



More information about the SciPy-Dev mailing list