[Asterisk-Dev] benevolent dictatorship, or inclusive developper community?

[C. Maj]

On Thu, 8 Jan 2004, Chris Albertson waxed:

> OK, We can argue CVS vs. BitKeeper.  But it will become a
> short argument when you read the fine print in BitKeepers
> license terms.  Basically Digium would have to pay not
> a few bucks as all their work is not Open Source.

There's the rub.  And although this thread seems dead now
with the Good News of some milestones for releases from
Mark, I still remember this one time back in 3rd grade when
CVS beat me up on my way home from school -- and I'll never
forgive it :)

> No doubt BitKeeper has a smarter merge algorithum than does CVS.
> CVS is pretty dumb.  But Asterisk is a relitively small software
> ssystem with few developers.

CVS IS RETARDED.

> While BitKeeper is now used to hold the Linux Kernel sources, this
> is a recent development.  For years people would simply e-mail
> Linus a patch and he'd apply it (or not) by hand and no version
> control system of any kind was used.  In fact Linus resisted using
> one for a long, long time and prefered to do the work manually.
> So to argue that something like BitKeeper is _required_ is not
> quite true as _most_ of the kernel was developed using only "diff"
> and "patch".  

I thought BK proved itself to even the doubtful in November,
when a kernel exploit was squashed in less than 24 hours.

Of course it's not required, but that's like asking someone
to light their crack pipe with a spark from smashing two
rocks together instead of using flame from a butane torch.

> But keep in mind the size of the Asterisk code base.  It is not
> all that large.  You can count lines of code (LOC) many different
> ways.I like to simply count semicolons with something like this:
>   grep ";" * | wc
> Every other metric I've seem is roughly proportional to the above
> so for relative comparisons I think it is "good enough"
>
> Using the above, we can compare Asterisk to both the PostgreSQL
> DBMS and to the Linux kernel.
> 
>   Asterisk      41,000
>   Postgresql   130,000
>   Kernel     1,300,000
> 
> The kernel is about 30X larger in terms of LOC and maybe 200X
> larger in terms of the number of developers working on it.
> The kernel is about two orders of magnitude larger problem.

That's not really fair, defining _one_ metric and using it
to make your case.  But ignoring the straw man, let's take a
look at PostgreSQL from a Developer's point-of-view:

    http://developer.postgresql.org/

PG has more docs than code !  There's a TODO list, unapplied
patches list, developer's FAQ, etc.  Yes, PG uses CVS, but
it is a mature project with a lot more flesh to make crappy
CVS not seem so crappy.

--Chris


-- 

Chris Maj <cmaj_hat_freedomcorpse_hot_info>
Pronunciation Guide:  Maj == May
Fingerprint: 43D6 799C F6CF F920 6623  DC85 C8A3 CFFE F0DE C146