Re: [Nmh-workers] gnu arch revision control system

[J C Lawrence]

On Thu, 23 Sep 2004 14:17:54 -0500 
Chad Walstrom <address@hidden> wrote:
> J C Lawrence wrote:

>> Similar can be said in favour of BitKeeper with the addition of MUCH
>> better merge tools than either Arch or Subversion along with a large
>> host of other developer-centric niceties.

> IIRC, BitKeeper is not a viable option for people who hack on CVS, GNU
> Arch, OpenCM, or similar Open Source or Free Software projects.

Yup.  Such individuals either need to purchase a license or require a
second person to act as a patch<->BK gateway.  Given that BK makes such
a gateway operation pretty trivial (even fully automated), this is a
minor inconvenience.

> That's a very convincing argument to stay away from it, IMHO.  

Rather than argue I'll simply point at the many threads on the Linux
kernel list and the effectiveness of the product for that project.

> BitKeeper isn't exactly decentralized, by the way.  

Violently disagreed.

> If you don't purchase the BitKeeper server...

There is no BK server per se.  Every "client" is a first class node
which makes no distinction between server and client modes.  

> ... for your own internal use, your statistics are reported back to
> the central BK server.

What is reported back is what you can see at the openlogging.org site:
changelog comments (ie collected checkin comments).  BK operation
doesn't depend on that reporting, just the license does.  In the case of
Open Source software this is hardly a concern as that data is already
public by nature of the license.

-- 
J C Lawrence
---------(*)                Satan, oscillate my metallic sonatas.
address@hidden               He lived as a devil, eh?
http://www.kanga.nu/~claw/  Evil is a name of a foeman, as I live.