Devel models [was: Setting Deafult Font Size in X]

["Stephen J. Turnbull" <turnbull@example.com>]

This is a little OT, so tell us to take it off line if you want, but I
think most people will be interested.

>>>>> "SN" == SN Diamond <Norman.Diamond@example.com> writes:

    SN> That was accomplished with vim 6.0.  (Present alpha version
    SN> 6.0ac, considered not ready for beta yet, but nonetheless this
    SN> has been done.

Hm.  As a matter of professional interest, how does the vim community
decide what's "alpha", "beta", or what?  What other categories do they
have (stable, I suppose)?

At XEmacs we've basically given up entirely on "alpha" versions
(although we have automated "alpha testing", just do "make check").
All development code is available to those who wish to update via CVS.
So there's no good reason not to call it "beta".

We do make tarballs (and CVS tags) which are explicitly labelled "beta
releases."  That just means that the release engineer successfully
built the source code in question (which means that XEmacs is
successfully running batch as a Lisp interpreter, since it compiles
its own Lisp, but there are no guarantees about UI stuff).

We've recently also given up on trying to release a "stable" XEmacs on
a specific date.  "Point oh" releases suck, there's no way around it.
So what we've done is introduce a "gamma" testing stage, which means
(a) the developers all use the "gamma" code base in their personal
activities, (b) enough beta testing has been done that we are
confident that there are very few critical bugs left undiscovered, (c)
feature development is complete in the sense that the developers are
not going to add more code to them, and (d) no destabilizing code will
be added -- even non-critical bugfixes will be tested in the
development branch before being added to the gamma code base.

The rationale is that (1) beta coverage isn't going to be complete, so
we are going to see more bugs, and it's only fair to admit that to the
users, (2) a large fraction of users is willing to accept that minimal
risk, especially Debian unstable/Red Hat Rawhide/Mandrake Cooker etc
users, and (3) the gamma code base will be monotonically more stable
over time.  After a certain amount of successful use, we'll declare it
"stable" and retire the previous stable version from active maintenance.

I think this matches the reality of usage in the open source OS
community pretty well.  Comments appreciated, and information about
other models as well.

    SN> No "j" version will be needed as far as I can tell, and I'm
    SN> sure the Japanese participants are making sure of that.)

Congratulations!  Major milestone.  We still can't justify making the
Mule configuration default.  Maybe for 22.0....

    SN> Is Torvalds-san American already?  I know he got a job there
    SN> by doing something for pay instead of for free, but didn't
    SN> hear about a change of citizenship.

Not just him -- XEmacs (temporarily) lost the services of five
developers over the last year due to adjustment to new jobs -- three
of them moving from Europe to the U.S.

So I think the American influence among non-Americans is perhaps
larger than you acknowledge.

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
_________________  _________________  _________________  _________________
What are those straight lines for?  "XEmacs rules."