[Python-Dev] Python Core Mentorship program

[Laura Creighton]

In a message of Fri, 25 Mar 2011 18:14:02 -0400, Jesse Noller writes:
>Ben,
>
>In principle I agree with you - I would like open archives for the
>specific reasons you cite, but I value the ability for people who may
>not be comfortable with coming out and openly discussing things on a
>list if they know it's open to the magical powers of google and public
>archives. Heck, having open archives makes it *easier to find out*
>about the list itself, serving the purpose even more.
>
>But - weighed in favor of the target audience (those that may not yet
>be comfortable with "full disclosure", or discussing personality
>clashes on the tracker, or those worried about future employers
>digging up stuff) - I want to error on the side of the closed list
>archives for now. In several months, we all might realize it was a
>monumental mistake. At that time, we can fix the problem.

I would have thought that the set of people who were more comfortable
with the closed list was prettry close to zero.  Because the problem
with saying something stupid in public is really not one of perfect
strangers using google to find out that I said something stupid once,
but rather that current members of the target group, in this case
the subscribers to python-dev or python-dev-mentors will find out
that I think stupid thoughts _now_, and think less of me for it, and
maybe say some nasty things about me.

Python-dev historically has been rather special.  The forbidding message
"Do not post general Python questions to this list. For help with 
 Python please see the Python help page." in a red boarded box is
fairly effective at getting the message "do not waste the valuable
time of these people" across.  For a while, I remember, we lied and
said that subscriptions to python-dev needed to be approved, even
when they didn't.  That seemed to deter some people from even trying
to join, which was probably either a good thing, or a bad thing on
the whole (but, of course, we have no way to measure).  And the other
thing that makes python-dev unusual is that it is casually read by
a large number of people who never say a word.  All open mailing lists 
have lurkers, especially those who read them without a subscription,
through some other means, but python-dev is unsual in the number of
people who try to read it 'just in case something important happens',
and 'just to feel like they know what is going on in the python community'.
All of these factors add to the 'don't waste people's time' factor.

Thus there is a lot to be said about having a separate group, where
python-dev contributors answer questions that they have made time for,
even if others might consider them a waste of time.  Because those
others are not forced to subscribe and read them.  But I don't see
such a compelling reason for a closed group.  It's not as if we expect
that mentoring to be a source of deeply personal stories and
anecdotes.  Or that people want the safety to discuss heretical
approaches to changing CPython not expected to go down well with
python-dev.

It all seems to boil down to 'some people would be more comfortable
this way'.  I'd like to get some metrics on how many of those people
there are.  And I'd like to measure them against a different group,
people like me who won't contribute to a closed group, in part because
the whole closed-ness of it makes me undomfortable. My experience
with closed-groups vs open groups has been almost entirely negative,
which would be reason enough for me to hesitate to join one, but
especially when it comes to a _mentoring_ list.  The single most
important reason why I would post something I think might be really
stupid is because 'if I don't understand this, then there are probably
others out there like me in the same boat'.  So I ask such things with
the hope that the exchange will be googled _a whole lot_ in the
future.  And again, when I answer a question fully and completely, I
do so thinking 'I'll bet a lot of people, and not this one soul, will
be interested in this'.  If the answer will only be seen by the
comparatively few people in a closed mailing list, I am comparatively
unmotivated to write anything, or write anything substantial.

I've seen a whole lot of very bad behaviour on the part of self-styled
leaders of closed mailing lists.  They determine the party-line of the
group and then, because it is private, blast those souls who do not
conform with impunity.  Having been on the receiving end of a number
of such exchanges, my conclusion has been that having the whole thing
open is often the only defence one has.  Firstly, most people are
more restrained when what they say can be seen by the world at large,
so some of these incidents would not happen.  But secondly, the ability
to share the mail with others greatly empowers the people on the
receiving end.  But if you cannot get an outside opinion because doing
so would violate the group's closed-ness, then you are more vulnerable.

The point of bringing this up is not because I think that
python-dev-mentors is likely to run into these sorts of problems, but
to let you know that there is a substantial number of people out
there who are not emotionally comfortable with closed groups as
opposed to open ones.  And I don't see why the emotional comfort of
those who like closed groups should have precedence over the emotional
comfort of those of us who find closed groups threatening.

The argument 'we'll try it closed, and then we will see about opening
it up, later' doesn't work for me, because every time I have been
involved in such an effort, except once, the end result has been that
later _never_ happens. The people who like the group closed _always_
successfully resist such change.  It is too much easier for people
like me who only got involved in the first place because of the
expectation of the group's eventual opening to drop out and quit the
group than to spend the amount of energy necessary to open the group
up.  And for me, at any rate, my reluctance to spend the energy is in
large part because of the closed-ness of the group.  I'd care a lot
more about the group, if the group was open.

I don't think that people like me are that rare -- hmm, I should rephrase
that -- I don't think that people with this attitude towards closed groups
are that rare.  And it is hard to find out about us, because we're the
people who don't join such things.  So no matter how well a given closed
group works, according to those who are members of such a thing, it is
impossible to say whether the group would have been _so_ _much_ _better_
if it were open -- or whether being closed was essential to its success.

So 'In several months we might all realize that this was a momentual
mistake and then fix the problem' never works.  No matter what you do,
there will always be some people who like it, so you will never get the
'all'.  But what might be possible to measure is the number of people
who won't be joining a closed list, but who would join an open one,
versus those who will join the closed list but wouldn't join the
open one, by having some sort of vote.  Of course, a vote could only
measure people's intentions, which are often different from what they
really do, but it would at least allow some sort of measurement of
what isn't happening, which is otherwise close to impossible to see.

I for one would feel a lot better about what I was missing out on in
python-dev-mentors if I knew that there were some actual human beings
benefitting from it who would not have joined an open group.  Without
some sort of metric, I will always worry that it was done 'in case
there might be people like that out there' or 'to test the theory 
that the reason we don't have enough python-dev contributers is
because python-dev is an open group' without any actual use-cases
at all.

>
>"The perfect is the enemy of the good." :)
>

"If you can not measure it, you can not improve it."
Lord Kelvin: (Sir William Thomson)

Laura

>jesse