SOAPs as URCs [was: HTTP Futures]
Fri, 2 Dec 1994 06:02:54 -0800

[URI wg added to the distribution, since I discuss URC requirements herein.]

I have seen some productive discussion of SOAP-related issues recently on
the www-talk list. I've been somewhat confused, however, by the sometimes
conflicting ideas of the purpose and intended implementation (from an
Interpedia point of view) of SOAPs. There has been some discussion of
"central servers" and other types of centralized enforcement. This is not
the intention at all.

SOAPs are intended to be maintained by a URC service. The URC service
should be able to store meta-information created by an entity other than the
original publisher, so that I can rate articles as I come across them, and
others can use my ratings to help their searches. I simply ask the URC
service for any SOAPs from a list of SOAP-publishers that may be attached to
a particular resource that I'm considering retrieving. If I'm an
parental-controlled browser, I never allow links to X-rated resources;
otherwise, I simply note the type of SOAPs that have been attached for the
user's amusement. No central authority is appropriate, since anyone can
claim that a resource is X-rated; it is simply a matter of who's word I want
to trust.

As you can see, it is not necessary that rate their own
resources; some right-wing group would be more than happy to do that for
them. A parental-controlled browser would simply pick an appropriate group
of SOAP-publishers to consult for each resource. It would be unlikely to
trust Playboy's word, but it would probably trust the American Parental
Association or some-such.

I've been keeping a watchful eye on URC issues since this kind of wide
distribution is difficult to achieve. Obviously, some sort of partial
lookup facilities are necessary. Doing an exhaustive lookup for _all_ the
meta-information (or SOAPs) that have been attached to a particular resource
may take quite a while, and I may not want to wait; I should be able to do a
fairly good search without retrieving _all_ the URC information for a given
resource. It would be nice to be able to incrementally continue a URC
lookup, so that if my initial attempts are not satisfactory, I can go lookup
more information. Obviously, this kind of state information would be best
kept on the client.

I believe Christian Mogensen has come closest to the Interpedia vision of
SOAPs when he says:

CM> ObWWW: for Meta information - if we want meta information to implement
CM> SOAPs, then the architecture must allow meta information to be
CM> distributed -i.e. I may have SOAPs about your documents, and browsers
CM> which are aware of the fact need to integrate the two sources before
CM> presenting the result.

I would note that the current URC scenario document has quite a different
version of the SOAP vision. Further comments on that document should be
coming to the URI wg list soon.

-- | Harvey Mudd College |

"Remember: Once you pull the pin, Mr. Grenade is no longer your friend." -- Zippy the Pinhead