Re: creating an equivalent of r-help on r.stackexchange.com ? (was: Re: Should there be an R-beginners list?)
On Sat, Feb 8, 2014 at 4:41 AM, Patrick Connolly
<[hidden email]> wrote:
> |> Dear Don and Bert,
> |> Allow me to address some of your concerns below.
> Which you do very clearly by positioning your responses underneath
> what you're commenting on. That doesn't seem to be possible on SE.
In addition to Yihui's remarks (including "mailing lists are good for
discussions, and SO/SE is good for Q&A's"), I would only add that on
SE commenting inline is a non-existent problem.
On the Q&A site all communication is restricted to three types,
clearly separate forms of interaction: Question, Answer, or Comment.
The user may ask only one clearly defined question, well, per
Question. And each proposed Answer is supposed to answer that very
specific clearly described question. Everything else, going from rants
to requests for clarifications go (mostly) in comments (and are mostly
ignored). If the question is vague, the OP doesn't need to sift
through ML-like threads and comment inline, but simply edits the
original Question and adds the required information to make it clear.
Same mechanism works nicely for Answers.
This means that when dealing with a complex situation what you do is
break down the problem in clearly identifiable parts; then in the
Question you explain the background and ask a simple question; then in
a 2nd Question you re-explain the background (or link to the 1st
Question), and ask a second simple question; and so on. This requires
a self-discipline that helps the help-providers in understanding where
the issue lies, and how it could be addressed.
So while on a ML a discussion can quickly digress from a clearly
defined question to something extremely more diffuse, threads or no
threads (as Yihui mentioned, What was my original question?; and What
are we discussing right now?), on a Q&A web interface moderators (and
the community) systematically force the users to stay on topic. And
personally I find that useful: no more "I stop monitoring a thread
because I can't follow it anymore" (anyone?).