[tei-council] the release conundrum

[Syd Bauman]

Sebastian's view of the process (reproduced below) sounds good to me,
too. However, I think 6 months is an enormously long time to wait
between releases. I'm not suggesting we change the process to 3 or 4
months, but rather that if 6 months have gone by and nothing
"interesting" has happened, we need to re-examine either our system
for making things happen or our criteria for what is and isn't
"interesting". :-)  (Personally, I'm betting this won't be a problem
at all.)

I also think it is vitally important, Sebastian, that you write up
the procedure for making a release, so that someone else could be the
release manager without fuss.


LB> This is all fine by me, though just for the record I think I
LB> should point out that the editors are right out of acolytes.

I agree with Lou, and furthermore my ads in the personals for
"acolyte wanted -- no pay, minimal reward" have gone unanswered for
weeks. On the other hand, we need to decide at some point who else,
if anyone, gets R/W access to the CVS tree. E.g., Laurent is working
on both the dictionary and the terminological databases chapter --
should he be doing this directly in CVS? David Durand is working on
the linking chapter -- should he?

---------
I have tried to streamline our release offerings. We now have
 a) CVS
 b) a release which consists of:
    - debian and RPM packages
    - SF file releases
    - a duplicate of a Debian release on the TEI web site. the
      existing copies of P4 and P5 DTDs and Guidelines have now been
      removed, and the URLs are now silently redirected to the
      "release" tree.
 c) packages for Knoppix and Windows Emacs (not in sync yet)
 d) distribution by oXygen

we are now at release 0.1.10.

My view of the process is as follows:

1) the editors or their acolytes change CVS as they need to. it is
   not guarenteed to "work" at any given time.

2) when a consensus emerges that something interesting has happened,
   or that 6 months has passed, the editors will instruct the release
   manager (currently me) to prepare all the outputs in b) above and
   release them as close together as possible

3) the Knoppix and Emacs releases are a project of TEI Oxford which
   will follow a timetable of their own

4) we will all do our best to assist the oXygen people stay in sycn
   and be forewarned of important changes

5) the editors do not need permission from the Council to make minimo
   releases (ie 0.1.13), but the council will discuss making a minor
   release (0.2.0), and the council and board will both discuss
   making a major release (1.0.0)

I would propose now that we plan to make the 0.2.0 release after the
class and attribute changes have been implemented.

How does that sound?