[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: splitting experimental by arch?

>>"Guy" == Guy Maor <maor@ece.utexas.edu> writes:

Guy> Manoj Srivastava <srivasta@datasync.com> writes:
>> I think that is wide enough exposure for packages that were never
>> meant to have a wide audience anyway. I do not know what you meant
>> when you said "ftp.debian.org only distributes the top two levels"

Guy> You seemed to be promoting a third level of stability, on the
Guy> same tier as stable and unstable.  I can't imagine we need such a
Guy> thing.

	Well, some of us _can_ too imagine such a thing. ;-)

Guy> Experimental is for the very small set of packages which
Guy> are truly dangerous or very very different from an existing
Guy> version in unstable.

	That is true. But those are not the only criteria, not until
 we have a strong policy about moving from unstable to stable. And
 even then, I think it is a bad thing. I think we do need finer
 granularity than a released version and a work in progress version,
 we need an area for innovation.

	We also need a policy for moving from the innovative stage to
 the released stage. Asking people if a package is viable just before
 release is prone to error, seeing as how people can be incommunicado
 for months on end.

	The testing group is well enough, but they are not infallible,
 and the opinion of the author/maintainer also carries wieght.

	Package that are ndagerous, or truly different, or even plain
 alpha should go in experimental. When they are ready to join the set
 of packages that can be called stable, they are moved, by explicit
 action, out of experimental into unstable.

	This can be done on the schedule of the package development,
 not on debian's release schedules. It is also fail safe; an
 experimental package can in no way move to a release stage
 semi-automatically. So a package can't slip into stable because I was
 not around to stop it; and is you reverse the query at release time,
 we won't have perfectly fine packages being held up because of people
 being on vacation.

	Maybe it _is_ time to make experimental a more robust and
 accesible set; I would certainly welcome more innovation in Linux and
 Debian. Why not?

Guy> An alpha package can go to unstable.

	I disagree. Unstable is perhaps a wrong word -- what we have
 is a release in making. The name unstable was choosen to deter
 over-eager CD distributors; let us not fall prey to our own

	Developers are supposed to be (nay, required to be) running
 the so called unstable distribution. Alpha code should be kept out of
 it. Beta code, yes, alpha code, no.

	Unlike people at universities, some of us use our development
 machines as our primary machine; loosing tiamat shall paralyze me.

 Abstainer: A weak person who yields to the temptation of denying
 himself a pleasure.  A total abstainer is one who abstains from
 everything but abstention, and especially from inactivity in the
 affairs of others. Ambrose Bierce
Manoj Srivastava  <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05  CC 2D 27 12 1D F5 E8 6E

TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: