>>>> "Jan" == Jan Vroonhof
<vroonhof(a)math.ethz.ch> writes:
> people who make critical decisions. That person or small group
> of people is going to be the 'core' group that keeps the site
> going, and CVS makes their lives *harder*.
Jan> I can guess why having multiple people hack on it randomly
Jan> will cause problems. But how does CVS do so?
Doesn't that depend on how many people have CVS commit access?
We're not talking about making the Web site publically writable. I
think a reasonable thing to do would be to include the Webmaster on
the Review Board, add general responsibility for documentation to his
portfolio, and help him find volunteers to do those parts of the
documentation he didn't volunteer to do. (The reason for doing this
is to have somebody who will yell about changes to the docs that screw
his stuff up.) And he can add a couple of team members as he likes as
Web staff (who would not be review board members).
And maybe the "last patch wins" rule should be removed for the Web
site ;-)
Jan> First of all CVS will make your access problem easier: You
Jan> just edit the stuff locally and CVS will take care of what
Jan> needs go to the site. More importantly, CVS is designed for
Jan> group collaboration, that means that you can see _what_ is
Jan> changed and by who.
> Not to stomp on toes, or anything, but in the long run,
> *CONTENT* *IS* *EASY*. And CVS only simplies content addition.
Yes, I wonder about whether it will cause drag on design operations.
Remember, XEmacs's directory structure changes only slowly, but we can
expect that the Web site's directory structure will change fairly
often, at least for a while.
I have tried moving directories in CVS, and I've discovered that (for
me) past a certain, fairly small, point deleting the current module
and creating a new one is the easiest way to go. That's a lot of
extra work. Is this not true?
Jan> Yes, but our current problem is (or better was) one of
Jan> content, I can write all the READMEs , installation notes,
Jan> little items telling people there is know problem with
Jan> version so-and-so of the mule-base package but I have no way
Jan> of getting it to the site.
That's true for review board members; but the rest of us will still be
limited to xemacs-patches (or something equivalent?
webmaster(a)xemacs.org?), I presume.
> as we're working!) Give us a bit of time (like a couple
weeks
> or so) to make some changes.
One of the points of CVS is that if you (as Webmaster) do move a file,
CVS will tell other authorized users, mostly the Review Board, that
you have moved the file. A CVS wizard could set it up so that nobody
but the Web maintainer could move files, I believe. Only the Web
maintainer should be allowed to do that kind of thing, of course.
> SUMMARY MESSAGE FROM A WEB SITE VOLUNTEER: Implementing a whole
> bunch of cruft (like CVS, IMHO) is only going to make the whole
> thing messier in the long run.
Jan> Well, I respect your opinion and I promise to leave you alone
Jan> while you do your work. However when your work has stabilized
Jan> I would like you to consider a solution that allows multiple
Jan> people to work on the website, if need be[1]. A solution that
Jan> no longer makes the maintenance[2] of the website dependent
Jan> on one (or a few) persons time. For me such a solution is
Jan> CVS, it isn't perfect but we already have the infra structure
Jan> in place.
Jan> Footnotes: [1] I am not claiming this should be the normal
Jan> mode of operation.
I would go the opposite direction. Under normal circumstances,
everybody with CVS commit access to the sources should have equivalent
access to the Web site. Eg, someone who maintains packaged Lisp
should have check-in access to a set of pages (possibly a
subdirectory, depending on how tight artistic control the Webmaster
wants to maintain) devoted to the package they maintain. Maintainers
of binary kits might also have their own pages.
But when a design or organization overhaul is being done, the web
maintenance team should be left alone until they're done (or done in
for taking too long...).
Jan> [2] Note: That is maintenance, _not_ design. I think we are
Jan> disciplined enough to keep a make minor (but possibly crucial
Jan> information wise) changes in accordance with your style
Jan> guides if you are on that round the world trip.
I think CVS will help enforce this, as mentioned above.
--
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 two straight lines for? "Free software rules."