[Asterisk-Dev] Asterisk Maintenance and Development

Nick Bachmann asterisk at not-real.org
Mon Jan 3 17:54:55 MST 2005


Howard White wrote:

>VCCH, Inc. _is_ contributing (and reselling :))  Yes, I am looking for a
>role in which to contribute. 
>
There are lots of ways to help develop Asterisk without coding, but you 
really have to the the initiative for most of them.
You can:
-Help out in Mantis, testing patches and offering feedback
-Sponsor bounties (this is how you can steer the roadmap... new 
features/changes must be developed somehow)
-Look for and report bugs
-Write documentation
-Sponsor conferences
-Answer questions on irc and the mailing lists
-Pay for prompts from Allison to go with new features
-Get involved with Russel's maintenance group
-Ask developers directly how they need help

If you can code C (even marginally) there are even more ways you can 
help.  You can also:
-Write bug fixes
-Write new features (you can control feature development even more if 
you do it yourself!)
-Document and reformat code
-Optimize code
-Ask developers directly how they need help

Many developers (including myself) have pet projects that we're doing on 
the side.  These pet projects usually ofter cool functionality but are 
put on the back-burner in favor or paying work.  I think most of us 
would enthusiastically accept help to get them done.  This is also a 
great way to get a handle on the source; most coders are willing to do 
some tutoring in exchange for your help.  Even if you can't code, you 
can still help!  The best advice I can give for getting involved is ASK 
how you can be helpful.

>>>Yes, I know that Asterisk is an open source project that
>>>depends on the generous support of uncompensated programmers
>>>in their spare time.
>>>      
>>>
That's not always true.  Many of us are paid (or at least have some 
motivation beyond altruism) to do at least some of our Asterisk coding.  
At least personally, though, most of the programming stuff I do at work 
related to * is boring and not generally useful outside of the company.  
If I find a bug, I'll report it (and hopefully fix it), but neither my 
boss nor any of my "other clients" wants to pay me to develop a feature 
(or completely develop it) that's not immediately useful to them.  Which 
brings me back to those side projects get put on hold...

>Mark Spencer is in direct control of what I describe and shall have to
>delegate in order for anyone to contribute.
>  
>
Just one example of delegation that already exists: bug marshals (and we 
are just about all non-Digium and I don't think any of us are 
masquerading as Mark) have some of this delegated power.

>>Frankly, Mark and Digium contribute enough already. What we really need
>>is more people actually contributing, and less folks actually just
>>talking about what is wrong.
>>    
>>
Yes, Jim!  I think the Byrds would tell you that it's a time for action.

>Jim, these last few thoughts of yours are where we disagree.  Mark is
>"the single point of failure" that is in charge of Asterisk.  
>
I disagree... and I think that's a point you have the burden of proof 
on.  Look through CVS... there's more than just Mark coding.  Instead, 
he's more of the architect and "keeper of consistency".

>What I
>propose by introducing established Product Life Cycle tools is a means
>by which Mark may chart the future course of Asterisk development and
>then delegate the management thereof to other Digium staff or Asterisk
>volunteers.
>
Can you (or somebody else) explain "PLM" without buzzwords?  
Buzzword-ridden, "Xtreme Programming"-esque marketing is all I can find 
on the www.  I suspect, though, based on its name and what you have 
said, is that what you are suggesting is probably much more easily 
replaced by a text file called "todo".

My perception of the Asterisk developer community is that it mostly 
consists of the no-nonsense developers that come from certain 
backgrounds--telcos, Unix programming, etc.--that encourage more 
classical no-nonsense style over trying to use every new tool and 
buzzword that manages to escape into the software market.  This is the 
type of developer that uses vim/emacs over a GUI because he can get more 
done, faster, in an intuitive console program than that shiny IDE with 
more buttons and menus than I could ever know what to do with.  These 
are the same kinds of people who know better than to use any tools that 
replace simple files and effective communication.  'Simple' is a 
paradigm that means a lot to these people, and they're not interested in 
shifting that paradigm.  That's my perception of the pervasive culture 
here. It doesn't include every developer, but I think most would find 
most of that description fits.  I think if you understand those feelings 
a little better, you might understand not only why I disagree with PLM, 
but also why other stuff happens here.

>  Asterisk is now larger than what Mark may by himself manage
>  
>
Again, do you have some example of this?  I think Mark does his part 
with Asterisk exceedingly well.

>(along with Digium) and I am trying to spark the discussion by which the
>Asterisk community may positively encourage Mark to recognize the scope
>of what is going on and respond accordingly.
>  
>
That's why we have Bug Marshals and that's why drumkilla (and his newly 
forming team) manage -stable.  There's a lot of history in all these 
decisions...

>I also recognize the the open source community resents and rejects the
>regimentaion of planning and documentation.  Such resentment shows up,
>bluntly, as the shortcomings of Asterisk.  
>
I wouldn't call it resentment.  I'd say that Asterisk gets released when 
it's ready for release, and not before.  I would definitely agree that 
it would be helpful for everybody if we had a firmer idea of what is 
being shot for between releases, rather than the soft of hazy ones.

>Mark has to delegate control
>and the developers must accept the means of delegation.
>  
>
I think you grossly underestimate and trivialize the developer community 
and their role in Asterisk development.  Please, stop into #asterisk-dev 
and #asterisk-bugs someday and see what we do there.

Nick
(hermie on irc)



More information about the asterisk-dev mailing list