[SLL] Arguments against unknown process of open-source: Where is Eric Raymond?

[Xeno Campanoli]

I really don't want anyone to answer this post in terms of the second 
paragraph.  The first paragraph is an important introduction, but if you 
take the time to read and think about my topic, you'll see the matter I 
wish to address is rather different, but vitally connected to said 
second paragraph.

In a situation I am finding myself facing frequently these days (how's 
that for euphemistic subtlety?) I commonly hear the argument that open 
source, and particularly firefox, doesn't get QA attention in a 
formalized process which can be reasonably recognized from a business 
perspective, and that this explains quality problems with these 
applications/systems and that generally speaking they should not be 
trusted if you can possibly avoid it.

I now allege that the problem is we open-source / free-software folks 
don't think about the specific substance and falsehood of this argument. 
  You can call this pooh poohing, but what it really comes down to, is 
we don't know our own virtues in terms of metrics that can be talked 
about in meetings when people who have been able to get metrics from 
others who have been paid to collect them are easily able to sell with 
these latter available numbers.  I guess what I'm saying here is an 
adjunct to the larger statement I made a few months back that optimal 
open source is generated with the same breadth of inputs, from hardcore 
coding, to newsgroup posts like this one, as added resources to the 
projects.  I suspect this is what made Linus so successful, is he saw 
the many details and subleties which allowed him to manage more than 
just simple code projects (please excuse the bootlicking there).  But 
that diversity should also be in terms of personnel, not just the skills 
of the participants, and specifically, I would really like it if there 
was more being done in terms of breadth on hardcore statistics and 
usability information.  Wouldn't it be nice, for instance, if 
Distrowatch could tell you numbers on crashes, or reported install bugs 
as well as uploads of the distros, and wouldn't it be nice if Mozilla 
could give substantive bug handling numbers in terms of tester and 
developer hours with some kind of launchpad tool or something?

Okay, I'll stop there for now.

xc

-- 
There is more safety in diversity; more danger in great power.
There is love in effort to understand; hatred in refusal to.
http://brotherbases.org