Jörg Henne wrote:
Alex,
although I don't really aspire to become a AD committer, I'd like to
emphasize how important your observation is.
Thanks Jörg. You are one of those bright people on the periphery that I
would like coming closer to the core as a committer. We need more
people especially on things like DHCP where we only have a single
committer at this point in time.
I've been working with AD
for some time now, albeit very intermittendly.
My experience is that the learning curve is quite steep and the barrier
of entry somewhat high
Yes I agree whole heartedly with you on this point. Not only is the
documentation virtually nil it's aging rapidly with changes to the code
base.
The strategy also needs a fresh new update. But to sum it up quickly
here we're trying to implement a directory server however we want it to
have plugins that allow it to do other protocols needed to replace
Active Directory with a free Apache solution.
So ApacheDS is much more than LDAP as you have observed. However we
need to make this perfectly clear on our website.
o The lack of documentation. The AD documentation seems quite
fragmented. In places where there is documentation, it is rather good
(e.g. authentication), but in other places it is severely lacking. In
particular with respect to what I would call the "big picture" or
"strategy" and the core design. Therefore I'm looking forward to your
effort of creating educational material about this.
Even some code has virtually no internal documentation. A lurker just
pointed this out to me a while back regarding the Kerberos code. Any
kind of help even with documentation would be appreciated. BTW
committers need not just be code committers IMO.
o This strategy, big picture or whatever you want to call it, is the
second problem area I see. The documentation and web site doesn't
clearly convey where AD wants to go. This may be due to the fact that it
doesn't want to go anywhere particular (i.e. live happily in the "base
technology" area).
Yes this is a problem but we do in fact want to go somewhere. I think
our goal is to provide a solid infrastructure server that makes using
Windows Server obsolete without reimplementing Samba. Our approach is
from a protocol angle.
However this is perhaps material for another thread.
...
A few days ago I posted a question (without getting an answer) about
storing the partition configuration on the system partition.
I'm really sorry about this. I've been slipping a lot lately with my
responsiveness to this list. This is why building up the community with
some cross instruction on the core internals of the server is so important.
This has
clearly been part of the vision once
(ou=partitions,ou=configuration,ou=system), but has never been
implemented as far as I can see, and has never been documented as
missing.
Funny I did this a while back intending to put almost all the
configuration into the system partition minus some smart defaults.
However some of the OSGi work will replicate some of this so I did not
prioritize this high enough to complete it.
So, why did I ask this question? In order to be able to ship a
self-contained AD-SAR for JBoss, there has to be a way of configuring
it. Currently the configuration lives within the SAR which is quite
inconvenient, since I want the SAR to be a ZIP, not a directory.
Right I follow you here. Well let's see where this discussion goes and
if it leads to the lowering of the barrier to entry for new committers.
I'd like to see you get involved in making sure the SAR is working
properly (we had some issues here) and that the DHCP server is being
developed actively.
or go straight for
what I think would be the "nice" solution, i.e. store the configuration
in the system partition.
This is our aim. Would be nice to have you help us get there before the
next decade :).
What I can do, is to offer some help in two specific areas: the DHCP
server and the SAR packaging:
This is great to hear. You and others lurking and contributing as of
late are all perspective candidates as far as I am concerned. I think
together we can revitalize this project so it is moving forward much faster.
o With respect to to the SAR packaging I already outlined a few problems
I see with its current state. I made some changes to it in order to make
it work at all, but I think that a lot more changes whould be necessary
in order to make in genuinely and generally useful.
o I have worked quite a bit on the DHCP server. The problem is that we
currently don't really need a full DHCP server in our yet, but just a
PXE proxy DHCP server. Nevertheless, a full implementation may be one of
our future goals and my code already contains quite a lot of what would
be needed for the full implementation.
Sounds to me like we can really use those changes of yours in the DHCP
server.
Thanks,
Alex
