Re: ldap: ou=group versus ou=groups
some more (and partially general) thoughts ...
On Tue, Feb 15, 2011 at 12:40:31PM +0100, Christian Kuelker wrote:
> On 02/15/2011 11:18 AM, Petter Reinholdtsen wrote:
> > I believe we should leave it unchanged unless we have a good reason to
> > change it.
> Every change in an LDAP DIT causes drain of human man power. Admins
> or maintenance contractors have to work more for using continuously
> Skolelinux. Migration scripts have to written any way, but the "s"
> add some extra minutes of writing, testing, verifying ... Which
> leads to demotivation and less acceptance.
I think the best way to do the migration is completely independent of
all changes I proposed:
* Prepare a list (csv) of all user for every category you use:
students, teachers, etc.
* Prepare a (GOsa-) template for each category.
* Mass-create all users from the lists. For each category use the
I cannot imagine a more efficient way to do that, and if we want to
avoid that way and have it simpler we need to revert all the 'new
stuff' (Kerberos, GOsa) which has been developed since lenny.
> If a change is nessessary due to technical reasons, this unavoidable
> drain of man power is mostly accepted.
> However if the cause is just a normative rule (that plural looks
> better) it is hardly to justify to use man power for a Debian Pure
> Blend that is not respecting the time of others.
Well, where do you draw the line? It is now the chance to make these
changes (and in my opinion without extra minutes for the 's'). This
chance will not come again soon (hopefully). The missing 's' will be
missing "forever". If every second school in the world uses debian-edu
;-) it will be too late, but the missing 's' will be still annoying
(at least to some).
It's clear that backwards compatibility is important. You have to
compare what you gain with the work you create (especially for
others). My point of view is for sure the one of a developer not being
the one who has to do the migration (but maybe this changes soon...).
But I think (and made the experience when working on debian-edu), that
after quite some years since the beginning of skolelinux, here and
there cruft has built up. It's time to refurbish some things. This may
cause a bit more work for now (not the 's'), but will in the end
lead to a more attractive and better maintainable system. And this is
true for maintainers, developers as well as for our users in the
If you are too conservative, the "next generation" will one day