[Asterisk-Dev] benevolent dictatorship, or inclusive
developper community?
John Todd
jtodd at loligo.com
Wed Jan 7 11:32:41 MST 2004
[thread mismatch intentional so as not to sink multi-threaded reply
to bottom of list]
[posted only to -dev to reduce hot air chatter]
At 1:12 AM -0500 1/7/04, Brian Capouch wrote:
>Jeremy McNamara wrote:
>
>>You do know there is a Asterisk Documentation Project also.
>>
>>
>>Jeremy McNamara
>>
>
>Who is writing the book, and for whom?
>
>B.
There is a book being written. No, I'm not writing it.
In followup to the main thread of this Subject:
Digium has done a good job to this point of keeping the Asterisk
project moving forward. Until about 6 months ago, there was a fairly
small user community, and they all knew Mark personally. With
Asterisk's rapid rise to stardom, and the implementation of the bug
tracker, there are now many more people that have ideas, bug fixes,
and feature updates than was ever experienced in the past.
Additionally I will assume Digium is now busier than ever, taking
Mark's time away from pure Asterisk coding and making him more of a
business person and less of a coder as a full-time profession. At
least, less of an open-source coder.
So, what to do? I agree with some of the comments about bugs and
patches languishing in the bugtracker, but Mark is also 100% correct
when he says that not enough people are testing the submissions so
that they are pre-tested upon submission. The first thing to do is
get people testing and working on the bugs that exist in the system,
and commenting with "Yes, I tested this in X manner, and achieved Y
results, and it is [ready/not ready] for CVS integration." Hold your
flames: I realize there are many bugs in the tracking system that
have been tested, or do not need to be tested (very simple changes,
or documentation changes, or what-have-you.)
Having been one of the first "bugtracker marshalls", I will say that
even in cases where there have been extremely minor changes it has
been an effort to get Mark's attention for more than an hour or two
during a week to work on bugs or to comb patches for integration,
because he's been trying to earn money and keep the company growing,
or he's trying to have <gasp!> a life outside of the IRC channel.
Most of the items in the bug tracker are not "minor" in nature; they
require quite a bit of thought and consideration for repercussions
elsewhere in the * package. Some bugs need to be hunted down and
killed, which takes a significant amount of time in and of itself
just to identify the issues. Mark doesn't have the time to do this,
and the user community (with some obvious exceptions) has been slack
in helping out with giving valid and tested data in the bugtracker,
so Mark doesn't have the resources to pursue many of these valid
reports. I cannot see any blame here, given the extraordinary amount
of work that has and does happen on the project from his efforts.
So, what's the solution? Mark is in a tough spot: he's trying to
make his paying customers happy - the ones that you don't know about
who pay Digium to do extensive non-GPL modifications to Asterisk,
plus the customers that buy hardware. He's trying to run a business,
with many employees. This is a _full-time_ job. Now, at the same
time, he's trying to manage a full-time coding effort with Asterisk,
which gets more complex and tangled every day. It sounds like one of
four solutions is in order:
1) Hire someone full-time to handle Asterisk coding issues
that Mark can trust to maintain the "bottleneck" of Digium-only CVS
updates and code implementations
2) Hire someone full-time to handle Digium so that Mark can
get back to doing more time with open-source Asterisk efforts
3) Allow some small pool of "trusted" people to make CVS changes
4) Branch Asterisk to a new CVS server somewhere else that is
then more exposed to various programmers
Number 4 is the least likely and "worst" option in the tree, but I
include it as an option to be complete.
Number 3 is the best option in my opinion, but is perhaps not
feasable in Mark's mind since he feels personally responsible for the
code and has no method of allowing others to determine if a
particular patch/bugfix's author has disclaimed the code (legal
risk.) This also would force the issue of CVS being
"development-only" and coming out with a more formal release and
bugfix structure, which has made some progress in the past few months
but hasn't quite gelled yet.
Number 2 is the second-best option, since it gives the community
Mark's excellent coding skills and his full attention, but I suspect
it will not happen due to the requirement that Mark demote himself at
Digium.
Number 1 is third best but perhaps most pragmatic from my perceptions
. It would require someone coming 'up to speed' with Asterisk, and
would require funding from somewhere. However, it may be possible
for the community to pool some funds together to help in this cause.
I'd be willing to put $200 as a donation towards hiring someone, as I
believe in putting my money where my mouth is. There are many
changes that I'd like to see in Asterisk, but I can't personally
afford the full $150 (or 100$?) an hour consulting costs that Digium
would charge me. I'd be very happy to donate to a "pool" of funds,
though, as I'm sure would many others on this list. If perhaps some
skilled programmer that already exists could be hired by Digium for
the express purpose of fixing bugs with Asterisk, then maybe that
would reduce the cost as well and eliminate the curve for familiarity
with code. However, I would say that for me to donate funds would
require that they went towards fixing open-source problems that were
NOT requested by Digium customers who were paying for solutions. In
other words, I'm not going to fund someone whose time is being billed
out to someone else.
Martin to some degree has been working on public bugs, as his 95
CVS changes since Sept 1 attest. Most of his changes, though, seem
to not originate from the bugtracker, though I am pleased to see
un-noted bugs and features getting added via other channels. I
think, however, someone who is dedicated to public Asterisk
bugtracker issues would be a welcome resource for the community, and
one which may actually get some funding from those of us who will
actually cough up some money for the effort.
I am sure that I've missed possible solutions; discussions welcome.
This is a critical stage in the development of a platform which will
change telephony; doing nothing is not an option if we want to see
this "movement" continue to grow at it's current pace.
JT
More information about the asterisk-dev
mailing list