- From: Sandro Hawke <sandro@w3.org>
- Date: Thu, 05 Dec 2002 18:35:14 -0500
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- Cc: www-archive@w3.org, sw-team@w3.org
> Background: I'm doing a talk at euroweb on semantic web and the theme of
> the conference is sw, web services and grid and how they all relate.
Thinking a little more about that, I come to....
The web brings two significant technologies to the task of
constructing distributed computing systems.
First, it demonstrates communication through shared memory on a
massive scale. The utility of building distributed and parallel
systems on a shared-data abstraction has been explored in various
applications, from LINDA's tuple-spaces to large-scale commercial
database systems. The web's usual access primitives are, in fact,
much more primitive, but they are extensible within the current
infrastructure. (See webDAV and SOAP for two different approaches.)
Second, the web has a pattern of language modularity, reminicent of
protocol layering, which allows for data to remain usable as software
evolves without central control. This is an essential characteristic
of massively distributed systems, where for both technical and social
reasons nodes will not run the same software. The approach to
modularity began with HTML's rule saying "ignore tags you don't know
how to handle" and has matured sporadically until being directly
addressed with RDF.
These technologies complement each other well. Efforts to federate
databases traditionally stumble on the vocabulary problem, which is
addresses by RDF in combination with OWL. In effect the semantic web
technologies, combined with existing web technologies, directly
address the infrastructure needed for scalable and manageable
distributed systems.
... or something like that!
-- sandro
Received on Thursday, 5 December 2002 18:38:52 UTC