Re: NADF and other pilots [ was Re: UK Academic...]

[[e-mail]]

[Pardon the presumption in inviting myself into this discussion,
 but I think Erik raises an important concern.]

> > It was not an easy discussion and we did not resolve much, except to
> > (more or less) note that the NADF pilot and the White Pages Pilot
> > (including the Internet Pilot) will need to be kept separate without
> > any interworking, at least in the beginning ...
> 
> As I am a very firm believer in the "The only viable directory service is the
> global directory service" this worries me very much ...
> It therefore worries me that the NADF has chosen to keep it's pilot separate
> from the internet/paradise/WPP ...

In the short term, this decision was made for purely pragmatic reasons:
the NADF pilot is just starting up, and it was just too difficult
to start up the pilot and at the same time reconcile the structure/technology
being used in it with the structure/technology of the Internet pilots.

In the long term, I don't see any opposition to the idea of combining
the pilots from either community. Of course, that may be due to neither
community knowing very much about what the other is doing ... :-)

> In short (I do not want to break Wengyik's record) Stef (or Marshall) could
> you elaborate a bit on the arguments for not joining the pilots...

Again, I don't think anybody is opposed to the idea of combining the pilots.
However, there is major work involved in reconciling just the technological
basis of the two pilots, never mind the nontechnical stuff. For that
reason (and I know of no other; Stef, Marshall and the other NADF
people on osi-ds feel free to correct me), it was considered expedient
to keep the pilots separate, at least initially.

To give you an idea of the differences between the pilots, the
approaches of the NADF and the Internet differ in the following
key (to me) areas:

Naming: 
   The NADF naming scheme is listing-centric, the (current)
   Internet scheme is registration-centric. I actually disagree
   somewhat with Stef here in that I believe that reconciling
   the disparate naming schemes is the easiest problem to resolve
   (from a purely technical standpoint, anyway).

   The NADF and the Internet have different schemas. 
   [Actually, the NADF schema is effectively a subset of the
    Internet's by virtue of the fact that most everybody in the
    Internet is using quipu, and the NADF schema additions are
    in the 7.0 quipu oidtables.]

Information-sharing:
   The NADF and the Internet differ very fundamentally in that
   the former allows authority for an entry to be distributed
   across multiple DMDs, whereas the Internet requires that
   an authority for any single entry reside in one, and only
   one, DMD. In fact, with the exception of France, the Internet
   effectively requires that a single DSA be authoritative for
   all the immediate subordinates of a nonleaf node in the DIT
   (ie., EDBs).

   The NADF and the Internet also differ in that for the most
   part (again, excluding France, sorry don't mean to pick on
   the French), the Internet manages knowledge (DSA entries)
   using the Directory itself, whereas the NADF manages Directory
   knowledge by means outside of the Directory.

   The Internet has a common DIT which is shared across all
   participating DITs. The NADF allows each DMD to have its
   own copy (view) of the DIT (this is an oversimplification,
   but so is this entire message).

Directory Operations:
   The Internet mostly (excepting France again) synchronizes
   the DIT across DSAs automatically (using the EDB replication
   protocol). The NADF has a different scheme involving a
   exchanges between a central clearing house and various DMDs
   that require human involvement.

   NADF Directory operations are DMD-based, Internet Directory
   operations are DSA-based. This is not just a matter of
   different words, but affects the way responsibility for the
   operation of various parts of the Directory is apportioned.

... and this is just the differences off the top of my head.
There are millions of annoying little things that are going
to have to be reconciled for the two pilots to combine.

Ultimately, both communities are going to have to do some serious
compromising if the pilots are ever combined.


Wengyik