On 25 Jun 2009, at 12:14, Upayavira wrote:
On Wed, 2009-06-24 at 16:10 +0100, Ian Boston wrote:
On 24 Jun 2009, at 10:04, Chris Chabot wrote:
If anyone (from mentors to contributers and especially mailing list
lurkers)
feels differently, then please let us know so we know to fix it ...
but from
where I'm standing I would say we look about ready!
I agree with everything that Chris has said (I would not have dared
ask the question otherwise :) ).
With a Lurker hat on:
Sometime I feel that we dont discuss motivation or design of a change
as much as we might, which makes it hard for new commers to
understand
what is going on, or about to happen. In some respects this is
unnecessary since we are implementing the Gadges and OS specs.
However
there are plenty of TLP's that are hard to work in as a result of
being a) complex and b) fast moving. IMHO its someting we have to
work
at. That would be my only critisism.
With a committer hat on:
The review of submissions from commiters and contributers alike, and
rapid resolution of differences of opinion without hurting personal
possitions or emotions typifies Shindig. (IMHO)
I presume you are not a Google employee. I'd appreciate it if you
would
speak about Shindig from a non-google perspective.
This is because, understandably, Google does dominate Shindig. For
graduation, the major concern is that sufficient space has been
created
in the project for non-Google contributors to participate on an equal
footing.
What I do notice is that activity here is fast. I read very little of
the traffic on this list. As a mentor, I simply cannot keep up. Does
this form a barrier for someone who would like to work on/with
Shindig,
but cannot do so (a) full time (b) on the West Coast USA (c) based
from
a Google office (d) etc?
Upayavira
With a non google employee hat on, a personal view:
I Started getting involved in Shindig about 16 months ago, and found
the community immediately open and friendly. In spite of being Google
dominated, more so in those days, there was an certainly an openness
to talk about things on list. Its hard for me to judge if the onlist
activity is the tip on icebrerg representing all the communication,
but I cant remember of any instances over the past year where I have
felt that I was watching a replay of a previously agreed outcome.
Shortly after joining I noticed a number of potential improvements in
the areas of the Social Model, the Service Provider Interfaces and the
JSON serialization. All of which I felt able to contribute to,
refactoring the service layer into an SPI with seperate sample
implementation, tweaking the model and providing an alternative
implementation JSON serialization (although this was not that great a
piece of work, subsiquently fixed by Paul L). At that time I certainly
was not working full time on Shindig and I live in the UK, but in
general then and now I feel I am able to keep up with the pace of
change. If I compare shindig-dev to Jackrabbit dev which I also follow
I think they are quite simular in rate of change and complexity. So
does the pace of change form a barrier to contributions to a), b), c)
and d) ? I think the answer is no, with the caveat that the attention
to detail that is reqiured is high, which requires carefull thought on
the part of anyone contributing.
I suspect that this is driven by the unusual circumstnaces that mean
the code at trunk/head is appearing in significant production, as I
understand it, shortly after being committed. Shortly when compared to
many other Apache projects. This has real benefits and significant
drawbacks. The real benefits are that the code base is driven by
production evidence, the drawbacks is that changes represent risk, and
the likelyhood is that those not running in heavy production will make
contributions without a detailed knowledge of the impact.
I my mind this is a teathing problem for the community that could be
addressed by decoupling the production deployments from trunk/head and
greater use of a sandbox space for experimentation. Having said all
that shindig-dev is a list that I try and read daily, even though I
dont have enough time at the moment to make significant contributions,
or review the patches of others.
I hope that helps put a GMT, non Google perspective on it.
On the whole I think that there a good ballance of Google/non Google
committers, but slightly less production risk for contributions would
reduce barriers.
All IMVHO, others mileage may vary, but I would like to hear from
others.
Ian
