[moby] [MOBY-dev] Reserved word... perhaps unnecessary...

[Andrew D. Farmer]

Hi again-

>    One last comment regarding rootedness: If you interpres a message, you
> always need to know when to start, so anyway you define a root node. This
> is the same for RDBS: you need to know which table to start with. So the
> rootedness is not about "not having a root" but about "having more roots
> than just one". But that is very probably what you meant anyway.

Well, not exactly. Certainly for RDBMS the theory is that the starting point(s)
are irrelevant : which table to start with for a query involving joins 
is an implementation detail determined not by your query message
but by the optimizer; similarly the row in a table that is processed first
is immaterial to the sense of the query. In terms of serialization yes,
you must pick a first element (otherwise it wouldn't be serial!), but 
RDF as a data model (not a concrete syntax) seems to be inherently about
"random access"; e.g. the ability to ask for all resources having a 
given property (regardless of what class they belong to) is perhaps even
more "unrooted" than RDBMs in which you clearly specify the tables
to which your desired instances belong-
but it is probably a wrong analogy to pair tables with classes- better would
be to think of each property as its own two-element "relation".

If pressed to comment on the actual practical implication of taking one
attitude rather than another, my thoughts are not entirely clear, but
I would claim that it has to do with thinking about "service descriptions"
in part as defining queries that extract the data that they will process from a 
given dataset rather than defining message structures that they expect a 
client to formulate and send to them. The former approach is along the lines 
of the data-driven "dynamic discovery" idea of ISYS. This is also the 
direction I think semantic MOBY is in some ways trying to take.



>    I am not against RDF. I am just against jumping to it without saying
> why.

I am in whole-hearted agreement with that attitude. I feel the same way
about XML ;) In all seriousness though, I think you are absolutely right to
insist on conceptual clarity in these decisions and that there are areas
in the MOBY problem space where RDF is NOT the right approach. 

Regards,

-- 

Andrew Farmer
adf at ncgr.org
(505) 995-4464
Database Administrator/Software Developer
National Center for Genome Resources

---
"To live in the presence of great truths and eternal laws,
to be led by permanent ideals-
that is what keeps a man patient when the world ignores him,
and calm and unspoiled when the world praises him."
-Balzac
---