Lists Home |
Date Index |
Yes and no. That is what they want: perfect reporting, interoperation
and portability. It may not work even if they build their registry.
SGML lessons learned from CALS:
1. The politics of registries and repositories are far harder than
setting them up. This is not just a people problem. Semantics drift,
names collide, worlds, collide and so on. On the other hand, a library
is a useful thing but the politics of who calls what when are local.
2. Because item one means that the chance of all parties converging
on perfect(n) is not good, transformation systems were applied. These
are usually local because of the semantic drift.
3. The Tower of Babel was not the failure of men. It was the ambition
of one king to be the only intercessor to a Deity. That Deity chose to
destroy the tower, not the people. It is a lesson in plurality. Deities
don't need intercessors. This is a command and control issue; not
an intelligence and computers issue. The primary is pattern recognition
based on requests by topic, then local policy sets operational objectives.
So a registry will be a good thing as long as it isn't the only thing
and the locals can still make local decisions. This goes for fusion
projects as well. Fusion is fine until one uses it to dispatch local
assets from centralized services without regard to local conditions.
Fusion systems should make Calls For Service and let the locals handle
the dispatching of assets as well as most command and control decisions.
Information should inform. Operational control and functional control
are not the same thing. Be wise about this. Much is at stake. Much.
From: Joshua Allen [mailto:email@example.com]
> Do read this article. In the realm of things the Beltway
> has tried before and failed last time for reasons that have
> absolutely nothing to to with the technology, this one
"If everyone develops their own data, schemas and definitions, we have a
Tower of Babel and we haven't gotten what we wanted."
That pretty much sums it up.