Re: [Gnu-arch-users] Conflicts in .arch-ids

[Aaron Bentley]

James Blackwell wrote:
> Catalin Marinas:
>
> > > A good start would be a public Linux mainline BK->arch gateway (one way)
> > > but this probably cannot be achieved with the free BK because of its
> > > license,
>
>
> BK's license prevents developers of other revision control systems from
> writing a BK->arch gateway without purchasing a license (about $2k
> IIRC). As such, Tom Lord, Robert Collins, Aaron Bentley and myself can
> not legally work on such a gateway unless a couple thousand dollars is
> forked over (per person).

It's a bit stronger than that.  Even Alan Cox couldn't use BK, because 
Red Hat works on CVS.

> What's more, the clause is rather ambigious, so its potentially possible
> that anybody that wrote a BK->Arch gateway could be considered as
> "working on arch".

One clean-room approach would be for an Archer to specify the API they 
needed, and for a BK-clean user to write a library conforming to that API.

The API would not, then, be Arch-specific, and in fact, could also be 
the basis for BK -> svn, BK -> darcs, and BK -> perforce gateways.

I'm not necessarily recommending this approach, but one of the cscvs 
maintainers could probably spec up a good API.

> So what to do, what to do? Larry McVoy, speaking as the CEO of BitMovers
> Inc, has said that anybody that wants to perform this work can license a
> copy of bitkeeper.
>
> So I'll ask this... is there anybody that, if a licensed copy of
> bitkeeper were provided to them, has the time, the inclination
> and the credentials to work on a BK->Arch gateway? 

Not I.

Aaron

--
Aaron Bentley
Director of Technology
Panometrics, Inc.