Development process (was: scipy.stats)
Hi, Sorry - I moved this thread over because it seemed to be of general importance - both for scipy.stats but also for our working process. I really think it's a discussion we need to have in public - either here on the list - or in a public meeting in scipy. On Mon, May 31, 2010 at 7:16 AM, <josef.pktd@gmail.com> wrote:
Since Travis seems to want to take back control of scipy.stats, I am considering my role as inofficial maintainer as ended.
I would have appreciated his help almost 3 years ago, when I started to learn numpy, scipy, and started to submit patches for scipy.stats.distributions.
But by now, I have pretty strong opinions about statistics in python, after almost three years, I'm a bit tired of cleaning up the mess of others (and want to clean up my own mess), and there are obviously big philosophical differences for the development process between me and Travis (no discussion, no review, no tests). http://projects.scipy.org/scipy/log/trunk/scipy/stats/tests
Returning to Josef's frustration. It seems to me there is a general thirst for a more formal understanding of code ownership, and the process for discussion and review of changes. I think Josef's frustration is very important; my impression is that Josef's work on scipy.stats has been a very large step forward for a module that previously was seriously broken and not properly tested. It's obvious I think that Josef is the de-facto maintainer of scipy.stats, and it's obvious that, as maintainer, he should first - be asked about changes, and second - should have the right of review before acceptance. Any maintainer would expect this, and any maintainer is going to get annoyed if those accepted rules of behavior are not observed. Travis - hoping you are reading this thread - does that also seem reasonable to you? What do you think would be the best way of proceeding? Best, Matthew
On Mon, May 31, 2010 at 2:15 PM, Matthew Brett <matthew.brett@gmail.com>wrote:
Hi,
Sorry - I moved this thread over because it seemed to be of general importance - both for scipy.stats but also for our working process. I really think it's a discussion we need to have in public - either here on the list - or in a public meeting in scipy.
On Mon, May 31, 2010 at 7:16 AM, <josef.pktd@gmail.com> wrote:
Since Travis seems to want to take back control of scipy.stats, I am considering my role as inofficial maintainer as ended.
I would have appreciated his help almost 3 years ago, when I started to learn numpy, scipy, and started to submit patches for scipy.stats.distributions.
But by now, I have pretty strong opinions about statistics in python, after almost three years, I'm a bit tired of cleaning up the mess of others (and want to clean up my own mess), and there are obviously big philosophical differences for the development process between me and Travis (no discussion, no review, no tests). http://projects.scipy.org/scipy/log/trunk/scipy/stats/tests
Returning to Josef's frustration. It seems to me there is a general thirst for a more formal understanding of code ownership, and the process for discussion and review of changes.
I think Josef's frustration is very important; my impression is that Josef's work on scipy.stats has been a very large step forward for a module that previously was seriously broken and not properly tested.
It's obvious I think that Josef is the de-facto maintainer of scipy.stats, and it's obvious that, as maintainer, he should first - be asked about changes, and second - should have the right of review before acceptance.
Any maintainer would expect this, and any maintainer is going to get annoyed if those accepted rules of behavior are not observed.
Travis - hoping you are reading this thread - does that also seem reasonable to you? What do you think would be the best way of proceeding?
Best,
Matthew _______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
[Quoted from the original thread] On Mon, May 31, 2010 at 2:59 PM, David Goldsmith <d.l.goldsmith@gmail.com>wrote:
I don't think I'm being unjustly personal: I think if certain parties (is that unpersonal enough?) flaunt community standards of practice, then they should be called out on it - what other teeth does a policy in this sort of community have? But if I'm mistaken about this Policy not yet having been promulgated/accepted, i.e., if there is no infraction on account of a technicality, then please, say so: I'll happily retract my accusation and apologize.
But that doesn't give you free rein to indulge in psychoanalysis as to motives and such.
I'm sure you mean (for it's the nature of this list, is it not, that we all have free rein to be as diplomatic, or not, as we wish) something along the lines of: if you tone it down a bit and make it less personal, i.e., be a bit more 'diplomatic,' then people are more likely to take you seriously. In fact, I feel I am being diplomatic by refraining from stating what I think should really be done; you yourself in your first reply to the original post intimated that person(s) who shall (now) remain nameless (but whom it was fine for you to name when the thread began) have *repeatedly* flaunted the standards in this manner, and yet we're asked to remain patient and continue to try to "housebreak" these infractor(s) (your metaphor - who's being disrespectful there, comparing the infractor(s) to puppies). I've said my piece, I'll shut up now. DG
Hi David,
I'm sure you mean (for it's the nature of this list, is it not, that we all have free rein to be as diplomatic, or not, as we wish) something along the lines of: if you tone it down a bit and make it less personal, i.e., be a bit more 'diplomatic,' then people are more likely to take you seriously.
Sorry - I should have replied to the earlier thread after starting this one. I think that is indeed Charles' point, that the best thing to do, is to identify the general problem, where the problem does not start with 'if only X would not ...' but is more on the lines of 'there must be a problem in our process because the following things happen fairly often ... ' That's what I am trying to do with this thread. I think we have structural problem in organization, where it is not clear what the process for code maintenance is. I think many people believe that we need such a process, but, given we do not have one, it is inevitable that things like this (significant portions of untested code suddenly appearing in trunk) are going to happen. What we need is a ) agreement that there is problem and b) an idea of how to go forward. I think it's also obvious that that conversation has to happen in public and on record so we can all have our say and agree. I'm sure it's possible to do that. And - Travis (sorry - I am sure you are doing more enjoyable things for Memorial day) - of course it's essential that you join in with and / or lead that conversation. See y'all, Matthew
On May 31, 2010, at 9:09 PM, Matthew Brett wrote:
Hi David,
I'm sure you mean (for it's the nature of this list, is it not, that we all have free rein to be as diplomatic, or not, as we wish) something along the lines of: if you tone it down a bit and make it less personal, i.e., be a bit more 'diplomatic,' then people are more likely to take you seriously.
Sorry - I should have replied to the earlier thread after starting this one.
I think that is indeed Charles' point, that the best thing to do, is to identify the general problem, where the problem does not start with 'if only X would not ...' but is more on the lines of 'there must be a problem in our process because the following things happen fairly often ... '
That's what I am trying to do with this thread. I think we have structural problem in organization, where it is not clear what the process for code maintenance is. I think many people believe that we need such a process, but, given we do not have one, it is inevitable that things like this (significant portions of untested code suddenly appearing in trunk) are going to happen.
What we need is a ) agreement that there is problem and b) an idea of how to go forward.
I think it's also obvious that that conversation has to happen in public and on record so we can all have our say and agree. I'm sure it's possible to do that.
And - Travis (sorry - I am sure you are doing more enjoyable things for Memorial day) - of course it's essential that you join in with and / or lead that conversation.
How many people interested in this discussion will be at SciPy this year? It may be a good idea to have a discussion about this at the conference. We could phone conference others in as well so that every voice can be heard. I do think we need to address this issue. I did not realize I was offending people with my enthusiasm for having a chance to work on SciPy. I have always resisted too much "procedure" and "policy" so that it becomes difficult for people to contribute. I really think technology changes and DVCS can help with this process. -Travis
On Mon, May 31, 2010 at 11:02 PM, Travis Oliphant <oliphant@enthought.com> wrote:
How many people interested in this discussion will be at SciPy this year? It may be a good idea to have a discussion about this at the conference. We could phone conference others in as well so that every voice can be heard.
I think it should be done here on the list. This makes it easier for all to review and refer back to. Also makes it more open, NOT that you are trying to do it behind closed doors.
I do think we need to address this issue. I did not realize I was offending people with my enthusiasm for having a chance to work on SciPy. I have always resisted too much "procedure" and "policy" so that it becomes difficult for people to contribute. I really think technology changes and DVCS can help with this process.
I am all for DVCS. (I posted this on another thread but it is more appropriate here or maybe I just want to repeat myself :) ) The diversity of perspective as to who has done what and more importantly if it was right or ok to do seems to imply that there is lack of clear roles/policies. If there was it seems that we could hope there would not be this diversity in perspectives.
-Travis
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
Hi, On Tue, Jun 1, 2010 at 6:32 AM, Vincent Davis <vincent@vincentdavis.net> wrote:
On Mon, May 31, 2010 at 11:02 PM, Travis Oliphant <oliphant@enthought.com> wrote:
How many people interested in this discussion will be at SciPy this year? It may be a good idea to have a discussion about this at the conference. We could phone conference others in as well so that every voice can be heard.
I think it should be done here on the list. This makes it easier for all to review and refer back to. Also makes it more open, NOT that you are trying to do it behind closed doors.
I agree very much that discussion on the list is better. I think it helps solidify the idea of numpy and scipy being a community project, where all discussion is public and open. I know that can be a little tough sometimes, but that too has its benefits in clearing the air and making people feel that the discussion is open. See you, Matthew
participants (4)
-
David Goldsmith -
Matthew Brett -
Travis Oliphant -
Vincent Davis