3
$\begingroup$

The thread Policy on R questions: Stackoverflow vs Stats Analysis suggests that we have some difference of opinions as to how to approach this issue. To get to some sort of resolution I propose that we use this thread to offer several different proposals and let the votes decide what we should do. I would suggest that we follow the following ground rules:

  1. The suggested policy measures are mutually exclusive. So, please vote up the one policy that you think we should adopt.

  2. Do not downvote any proposal. I suggest that we do not count downvotes at the end of the polling period to decide what to do.

  3. If you wish to add another policy proposal feel free to do so with the caveat that your proposal must be distinct from the ones already existing.

Let us run the poll for one week i.e., till Nov 5th, 2010. The proposal with the highest number of upvotes will be the proposal we will adopt as far as R questions are concerned.

$\endgroup$
5
  • $\begingroup$ I believe it is a topic for meta.SO; we don't have a power to decide about SO. $\endgroup$
    – user88
    Commented Oct 29, 2010 at 16:18
  • 1
    $\begingroup$ @mbq We can decide what is acceptable to us. Then we can go over to meta.SO and ask their opinion. $\endgroup$
    – svadali
    Commented Oct 29, 2010 at 16:22
  • $\begingroup$ @Srikant ok than. $\endgroup$
    – user88
    Commented Oct 29, 2010 at 16:24
  • 3
    $\begingroup$ If there are relatively few upvotes, or some people decide to go ahead and downvote, that would indicate lack of consensus. In that eventuality, we should not rush ahead to encourage changes in policy that are manifestly controversial. Thus, I would encourage readers to contemplate the options carefully and not hesitate to vote up or down as you see fit. $\endgroup$
    – whuber Mod
    Commented Oct 30, 2010 at 18:23
  • 1
    $\begingroup$ whuber - thanks for making this point. This was my concern on the last thread: the StackExchange system just isn't made for developing consensus, because consensus requires knowing where everybody stands. The voting so far has been rather anemic, but does that mean people are undecided or just don't care? $\endgroup$
    – Matt Parker
    Commented Oct 30, 2010 at 18:36

2 Answers 2

Reset to default
10
$\begingroup$

I vote that we not have any R-specific policy at all. We are better off making policy related to the site's purpose rather than attempting to single out any particular software platform (which is just cutting heads off the Hydra).

Furthermore, I would like to suggest not migrating existing questions on any basis. Let this site grow as it will, but leave the history of its growth unchanged.

Further reasons are outlined in my response to the antecedent thread.

$\endgroup$
7
  • $\begingroup$ @whuber Perhaps, I am not being clear. The post is not about existing R/Matlab/.. questions. But, about what is a desirable policy for the future. 'Migration' refers to migrating questions from one site to another site as soon as they are posted. As far as I know this migration can be done by the mods. Hence, the need for some consensus on a policy. I agree with you that the focus need not be on R although R was the trigger for this discussion. $\endgroup$
    – svadali
    Commented Oct 30, 2010 at 18:23
  • $\begingroup$ @Srikant But both your proposals are prominently and exclusively about R-related questions! $\endgroup$
    – whuber Mod
    Commented Oct 30, 2010 at 18:26
  • 1
    $\begingroup$ I think any policy we make on R questions can be easily adapted to other languages. R is currently more pertinent simply due to the amount of questions relating to it (it is the most frequent tag). Some languages won't have a SE site to migrate to, but I think answering this question in regards to the R software helps to clarify the scope of the site. $\endgroup$
    – Andy W
    Commented Oct 30, 2010 at 20:00
  • 4
    $\begingroup$ If the policy were on deciding which questions are acceptable here, that would be fine. But it centers on what ought to be migrated from SO. I think this is the exclusive concern of the SO community, and not within our purview. Especially since one solution advocates what type of programming questions ought to be migrated -- interfering with the explicit purpose of the flagship SO site: programming. We had previously decided to allow some latitude in accepting programming questions, because it's hard to set up a policy. I don't see how it's become easier by targeting one language. $\endgroup$
    – ars
    Commented Oct 30, 2010 at 20:55
  • $\begingroup$ @whuber @ars @Andy - @Andy's point is spot on. I really feel that the focus on 'R' is not meant to single out that specific language. The focus is an artifact of R's popularity on SO relative to other languages. Would it help if I edit the question to discuss migration of all data analysis questions to our site? Reg the issue of what ought to be migrated from SO- we need to first decide what we will accept. Then my plan was to open a similar question on meta.SO on the same issue to seek their opinion on what they think about the issue given our consensus. $\endgroup$
    – svadali
    Commented Oct 30, 2010 at 23:27
  • 2
    $\begingroup$ @Srikant: But we've already decided what's acceptable here. See: [ meta.stats.stackexchange.com/questions/252/… ], [ meta.stats.stackexchange.com/questions/35/… ], [ meta.stats.stackexchange.com/questions/1/… ], [ meta.stats.stackexchange.com/questions/276/… ]. Confused about how this question adds anything new to that discussion. This is one reason we need a community FAQ so we don't keep rehashing old ground. $\endgroup$
    – ars
    Commented Oct 31, 2010 at 2:30
  • $\begingroup$ +1 I agree completely; thought that this was already decided. $\endgroup$
    – Shane Mod
    Commented Nov 1, 2010 at 13:29
3
$\begingroup$

All R questions that are related to data analysis should be migrated from SO to our site.

Prototype Examples For Migration:

  1. How should I do [stats test/plot/estimate a model/import data/..] in R?

  2. What R packages can I use to perform [stats test/plot/estimate a model/import data/..]?

Prototype Examples Not For Migration:

  1. How can I speed up my R code?

  2. How can I avoid loops in R?

  3. How can I vectorize in R?

  4. How can I use R to interface with Excel/Python/...?

$\endgroup$
1
  • 3
    $\begingroup$ I think it would be more pertinent to consider what questions, if any, should be migrated from here to SO. Personally I think migration is a bit draconian and we should just encourage people one way or the other by means of helpful comments. That said, stuff in your second list does seem like it would get better answers on SO, while I'd like to think more analysis-related stuff would better served at SA/DA/CV/whatever. $\endgroup$
    – walkytalky
    Commented Oct 30, 2010 at 23:22

You must log in to answer this question.