[asterisk-dev] Release schedule ?

Howard White howwhi at vcch.com
Fri Nov 10 15:05:44 MST 2006


Greg Boehnlein wrote:
> On Fri, 10 Nov 2006, Luigi Rizzo wrote:
> 
>> On Fri, Nov 10, 2006 at 01:20:36PM +0100, Johansson Olle E wrote:
>>>>> I also suggest postponing any discussion about release procedures and
>>>>> stick to what we have and focus on the bug tracker.
>>>> and that was the first part (and the subject) of my mail.
>>>>
>>>> 	Is there a release schedule at all ?
>>> We are many months behind the schedule, so I think the current
>>> schedule is "when we have fixed all known and fixable bugs".
>>> Let's focus on the bug tracker, all of us. Now.
>> ok that's not a schedule.
>>
>> you are welcome to consider this discussion a waste of time,
>> but it is a commonly accepted concept that in order to make
>> things happen, there need to be reasonable and achievable deadlines,
>> so that there is a chance that people may convince themselves
>> (as that's the only way in a volunteer project) to address it.
>>
>> "now" is not reasonable - too early.
>> "when we have fixed all known and fixable bugs" is not achievable;
>> the list grows with time.
> 
> I agree. There is never a time when a piece of software is perfect. I also 
> agree there deadlines are good things that help prioritize and focus work. 
> It is to be expected that any .0 release should have a basic, expected 
> level of functionality. From the community perspective, this does not seem 
> to be well documented. I.E. what are the specific, measurable tests that a 
> release must achieve to be qualified for release? At Astricon, my initial 
> requests for regression testing in the Developer's conference met with 
> interest from only a few people, and consequently it got dropped off the 
> agenda in favor of other topics. Mark made the comment "Testing.. isn't 
> that what a user base is for?". Obviously, he was joking, but I'm sure 
> that many developers feel that way. They are in it for the thrill of the 
> hunt.. writing new code.. they don't want to spend a lot of time debugging 
> things and going through the heavy lifting and insanity of support. In one 
> respect, I think that is great.. you have to have brilliant people to turn 
> ideas into code. On the other hand, the great unsung heroes of Asterisk 
> are the legions of people that contribute to fixing bugs, improving 
> compatibility and  stability. In many cases, we forget about all the work 
> that these people do.
> 
> In that respect, I was extremely excited about the fact that Digium, as 
> shepards of Asterisk, have recognized and committed resources to 
> maintaining the bug tracker. This is a departure from the past, and is 
> an important change in mentality that needs to be supported and recognized 
> by the community. We, as a community, need to help ourselves by helping 
> Digium. Imagine what we could do if everyone on this list chose one bug on 
> the bug-tracker to review and attempt to fix!
> 
> I can honestly say, that in the past few months, I have seen a large 
> increase in the number of commits to 1.4. For my needs, it is very close 
> to meeting the base level of stability that I need to put a few 
> "production test cases" into play. VLDTMF has passed Interop Testing with 
> all of my carriers, and once some of the T.38 pass through issues are 
> addressed, I'll be putting it into limited production for a small subset 
> of my customers.
> 
> In once sense, I'd rather have software that takes longer, but is 
> thoroughly tested in my environment, over software that is pushed out the 
> door because of an artifical deadline. As the old saying goes "We will 
> sell no wine before its time". I believe that Asterisk 1.4 will never have 
> all bugs fixed, but I believe that we can try to fix as many as we can by 
> focusing our energies on thorough testing and by developing a set of 
> criteria that we (as a community) believe is important for a release.
> 
> I'd like Blitzrage to chime in at some point about the Asterisk regression 
> testing that was thrown around in the Code Zone at Astricon. I think this 
> will go a LONG way towards quickly finding, fixing and being SURE that 
> Asterisk meets that certain basic level of functionality between releases!
> 
This wouldn't hurt so bad if I hadn't suggested more product life cycle 
management two plus years ago.

<Flamesuit = on><Earplugs = in>

The reality here is folks that to have schedules, plans, delivery dates 
and releases; people have to get paid for doing the work.  If everyone 
asking for the listed items wishes to pay license fees equivalent to the 
price of Cisco Call Manager or the like...

</rant>

Now go back to fixing bugs like we are.

Howard White
president - VCCH, Inc.


More information about the asterisk-dev mailing list