|
RDF (and RSS?) for the "future Web" (Was: Welcome back!)
|
Silvio Porcellana
|
Jan 18, 2001 07:52 PST
|
Aaron Swartz ha scritto:
| |
Disclosure: I am on the RSS WG.
|
Hi Aaron!
First of all, as an RSS-DEV lurker, I know your involvment in RSS and I
would like to thank you (and all the other WG members) for the great job
you're doing.
| |
I think that RSS is probably the best way to use RDF -- but only in a
specific situation. RSS is only useful for describing plain old websites,
but in this situation is has a number of benefits:
- it's simple
- it's popular
- it works with non-RDF (plain old XML) systems also
- it has a WG to standardize the most popular uses
However, RDF really shines when it comes to the "future Web" -- what the W3C
is calling the Semantic Web -- where instead of describing plain old
websites, we describe concepts and ideas, etc. It also works a lot better
for all of the more complicated things: logic, trust, etc.
|
As you say, RSS is a good protocol for describing existing resources or
for syndicating content and - thank's to the fact that it's really light
- for distributing these descriptions around the Web.
I do also agree with you on the fact that RDF is the tool for the
"future web" adding a semantic meaning to resources, and so transforming
them in "ideas", "concepts" and more general "things" than just a
collection of web pages.
Now, my question is: how are we going to build this "Web of the future"?
I mean, basically: what are we going to do with the tons of resources
that are already out there? If we were able to re-create the Internet
from scratch, then we will for sure start adding a semantic context to
all the resources from the beginning, but this is not the world we have
to face. Instead, we have to cope with billions of web pages, most of
them already catalogued (in search engines or directories) using quite
useless (or false) keywords and with no or little care for the context.
Another example: we have now thousands of RSS feeds and channels, and
navigating thru all these resources has become already quite difficult.
Should we start thinking about a meta-RSS? And then we will need a
meta-meta-RSS, and so on...
I don't know if RSS can be used for such a task, remembering also all
the problems that the WG had to face when RDF was introduced and the
purpose for what it was created: still, RSS looks like the only real
application that uses RDF at the moment (and all the other tools
intended to provide a bit of "semantic", like DC) and so, I think, it
should function at least as the foundation for a more complex work in
this direction. Maybe adding a trust mechanism and a bit of logic to an
existing (and working) system is easier and quicker than building a new
system from scratch. Or not?
Sorry I've been so verbose... Probably I need to work a little bit more
on the semantic meaning of what *I* write ;)
Cheers!
|
|
 |
|
