Bitkeeper in Linuxdevelopment :(

[Alessandro Rubini]

> It somehow worries me, that non-free software now is at the core
> of the prominent Free Software kernel linux.

This is a problem, but it's true that CVS is somewhat unsuitable for
large projects. I recall some talking about Linus possibly using
bitkeeper some three years ago (I think), so this isn't really
asonishing news.

Fortuantely, other smart people is working in revision management
systems, and the stuff looks promising.  While Linus often tends to
value practical benefits more than being a saint, I don't expect him
to become dependent on bitkeeper.  Even if he chooses to continue with
bitkeeper after testing it, as soon as other systems offer the same
level of functionalities and are Free Software I expect him to switch.

Meanwhile, looking at those systems may be interesting.

Tom Lord recently announced arch (see http://www.regexps.com , which
points to tarballs in ftp://ftp.regexps.com -- don't have the direct
pointer now).  He posted an interesting description of the system
to fsb at crynwr.com -- look for <200201160939.BAA27937 at morrowfield.home>
dated Jan 16.

There are now some arch-specific mailing lists (subscription addresses
in brackets):

        arch-users[-request]@regexps.com
        bug-arch[-request]@regexps.com
        arch-announce[-request]@regexps.com

Today, Jonathan Shapiro talked about the tool of his group, currently
called CPCMS. His comments in <003a01c1af27$0c1172e0$848e10ac at vmware> ,
posted on Feb 6.  His draft paper is at
http://srl.cs.jhu.edu/~shap/cpcms2001.ps (or .pdf).

Jonathan also comments (very briefly) on bitkeeper and subversions.

Archives of fsb are at http://www.crynwr.com/cgi-bin/ezmlm-cgi/0/

/alessandro, who hasn't yet looked at either one