On 7/11/06, Geir Magnusson Jr <[EMAIL PROTECTED]> wrote:
This thread may be dead/resolved, in which case just ignore me.

It was only "mostly-dead"...but you've raised some good points that I
agree with.

Cliff Schmidt wrote:
> On 6/23/06, Noel J. Bergman <[EMAIL PROTECTED]> wrote:
>> The use of e-mail as the primary means for communication is part of ASF
>> policy and philosophy, and we can certainly learn lessons from
>> projects that
>> have gone against it.  IRC tends to breed a more closed, albeit arguably
>> more integrated, community.
>>
>> That said, if IRC can be used as a learning tool to rapidly bring new
>> people
>> up to speed, and if the information gathered from those sessions is
>> preserved for others to follow up via web-site and e-mail, how do people
>> perceive that?
>
> I've never done that on a project, but I think it could be a
> reasonable thing for a project to try.  I believe the Synapse folks
> have been doing regular IRC meetings from early on.  I'd be interested
> in their perspective on the pros and cons, particularly as an
> incubating project.

Someone did point out that dev traffic is falling off while commit
traffic is same or increasing.

Yep -- and since asking about the Synapse perspective, I haven't seen
a persuasive argument that IRC has been a particularly positive thing
for them.  The key issue could be whether IRC is used as "a learning
tool to rapidly bring new people up to speed" (as Noel asked, and I
echoed, curiosity about) , or whether it is more for development
discussions (which I think is a dangerous move, particularly for a new
project).

> As a XAP mentor, I know that the committers already understand that no
> decisions will be made over IRC, that logs of each IRC will be
> immediately made available to the entire community, and that they need
> to be sensitive to any concerns from people wishing but unable to
> participate.  But, are there other thoughts from the Synapse folks or
> anyone else who has used regular IRC meetings?

I think that people can have that understanding, but I think that it
doesn't matter - it's been my experience that while people are able to
quote the letter of the law as well as the explain the reason behind it,
people unintentionally make "informal decisions" on IRC and execute on
them, all with the best of intentions.  I know i've seen it with
Geronimo, and it can be very disruptive, even though it may be accidental.

I think lots of decisions made on dev lists are the same - informal -
without the trappings of a vote or such, because many decisions are made
by "lazy consensus" - people discuss things or search for help, and then
continue down whatever modified path the group explicitly or implicitly
agreed to.

+1

In the case of XAP, I'm guessing that many of the committers are
employees or contractors/consultants of Nexaweb.  Were I a mentor, I'd
want to be sure that pre-existing development process is being
sufficiently broken up to make it an Apache community development
project, and would worry that regular IRC meetings might be confused
with periodic development meetings...

I'm not as concerned about this point.  Having a semi-monthly IRC
session to help bring new people up to speed is unlikely to be the
thing that holds back a closed development process from becoming an
open and collaborative one.

The short, sound-bite version of the advice I give companies that are
trying to transition their development process to one like Apache's
is, "commits should make sense with the context of the public dev-list
archive alone, and the dev-list should make sense with the context of
the code base alone." (there are exceptions such as bug/issue history,
etc, but that doesn't fit in the sound-bite ;-)   The idea being to
prevent potential hallway conversations or other communication from
being part of the context of the work.

The kind of IRC session that Noel was asking about is less likely to
be the problem.  However, I agree with your concerns people
unintentionally making informal decisions on development-oriented IRC
meetings.

Cliff

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to