On 6/26/06, Ross Gardler <[EMAIL PROTECTED]> wrote:
David Crossley wrote:
...
> It is now time to assess whether we want to continue.
> I have been wanting to raise this issue for a while
> The discussion at Incubator [2] spurred me.
>
> Is it successful for us at Apache Forrest?
Success can be measured in two ways:
1) development progress
2) community development
In my opinion, we have achieved no actual development that could not be
achieved via the mailing lists. In fact, quite the opposite. Our first
session saw an attempt to create an XHTML2 structure that saw a fair bit
of work by some people who were able to allocate allot of time to the
session.
Others who were not able to allocate as much time were unable to comment
on the approach.
The end result was a spat on how it should be achieved and various paces
of work were thrown away and there has been no progress on the XHTML2 in
core work since. [note this is not a complaint about work being thrown
away, just an acknowledgment that it would not have happened this way if
discussion had been had onlist first]
Community development also suffered as a result of this.
Ultimately the XHTML2 issues were (IMHO) one of the the root causes of a
considerable amount of tension within our community, and was the thin
"end of a wedge" in the development community that is only now starting
to loose its significance.
I think it's unfair to suggest that the communication medium had any
effect on the XHTML2 debacle. I think the disagreement would have
occurred regardless of it being on irc, email, or even skype. We
disagreed largely because of a fundamental misunderstanding of the aim
of the task. Given such a fundamental misunderstanding, a great deal
of work would likely have been done before the disagreement would have
surfaced even if it was onlist first. That happens sometimes. I
actually think there is still some solid, usable work in svn that came
out of that session.
Having said that, I think this was unique in its ability to be
effective with IRC because it was a big problem area with many aspects
that single devs could tackle - even thought we weren't effective at
it. Most issues aren't that large and so don't lend themselves to a
multi-dev collaborative IRC solution - or at least I don't work like
that. I think things like prepping a release, working through JIRA
issues, etc. would still be effective with IRC.
It is also true that there has been some valuable community development
work in these sessions. It is nice to "chat" and this cannot happen as
easily in email.
I agree.
It is also true that some valuable discussion regarding approaches to
technical problems have also been had, but these could, in most cases,
have occurred in email.
Yeah, it's true, but I'll say that my threshold for quality thoughts
is much higher with email. In other words, with IRC, I'm more willing
to just brainstorm without worrying so much about wording, goodness,
etc. Something about the persistence of email archives that makes me
think much harder before I communicate. (I know, irc logs live in svn
too - I didn't say it was completely rational).
So my view is:
No, it is not sucessfull from a development perspective and it is only
partially successful from a community perspective.
I think it is quite successful from a community development
perspective. I think it could be much more effective at forrest
development if we were to select topics that particularly lend
themselves to the communication medium. Some topics are by nature
solo-dev issues and shouldn't be addressed on a FF.
> Does it enable the whole community to be involved?
We can only speak about this on a personal level. I find it difficult to
be present for much of the day. Fridays are a busy time for me. When we
had it on a Tuesday it was better for me, but was worse for the majority.
So given my personal experience, I would say no it does not involve the
whole community.
It doesn't on any single FF, but I feel like as a whole I've been able
to spend quality time with the whole community, albeit not at the same
time.
> Can we address the concern about the summaries?
It is difficult to write a summary unless you were present during the
discussions. The logs can be hard to follow. This means they need to be
written as we "go along" by the people involved.
However, if someone is dropping in for 10 minutes between work
commitments they are unlikely to use that time to write a summary.
When I did op, I wasn't able to attend the whole FF so my way of
keeping up with what was happening was to read where I left off and
add to the summary doc. I didn't commit as I went but since I was
adding to the summary as I went, I could have modified the script that
was commiting the log to also commit the summary. This is not that
hard, if FF doesn't continue because of it's own lack of merit I can
accept that, but this would be an unfortunate reason to not continue.
> Should we let it continue?
I think the experiment should continue. I think if we periodically
reevaluate what works and what doesn't, we can make these sessions
much more fruitful in the long term.
What works:
o) Problems that lend themselves to collaborative solutions.
o) JIRA weeding
o) Social chat - community strength
What doesn't:
o) Technical issues that only take a single dev to fix - may as well
tackle outside of FF.
o) Problems that aren't really understood.
o) User assistance
--tim
