- From: Ron Daniel Jr. <rdaniel@acl.lanl.gov>
- Date: Thu, 22 Jun 1995 16:10:02 -0600
- To: "Ronald E. Daniel" <rdaniel@acl.lanl.gov>, Darren New <dnew@sgf.fv.com>
- Cc: Martijn Koster <m.koster@nexor.co.uk>, Nathaniel Borenstein <nsb@nsb.fv.com>, rating@junction.net, uri@bunyip.com, www-talk@www10.w3.org
Earlier this week I promised to revise part of the URC Scenarios and requirements document to accomodate information filtering ala KidCode. That one section of the document is appended. Regards, Ron Daniel ----------- \subsection{Annotations, Seals of Approval, and Information Filtering.} \label{sec:soap} Many projects already use the web as a means for describing, recording, and coordinating the work of the project members. URCs might be used as the basis for even more collaborative work. Workflow systems, and enterprise library systems are a couple of models that might be constructed on top of a URC service. A very interesting class of services falls under the general notion of ``annotation systems''. In the early days of Mosaic, a global annotation system was provided. However, that system did not have the necessary scalability, and we are unaware of any good proposals for global annotations. However, a very good proposal for group annotations has come out of the Stanford Digital Library effort \cite{prdm}. Its architecture would enable a variety of interesting scenarios, including several intriguing business opportunities. Many scenarios could be constructed using the models mentioned above. In this paper we will only consider one generic scenario, then describe how it can be specialized to meet a variety of information filtering needs. The generic scenario will have one group annotate resources with a Seal Of APpproval (SOAP), and other groups using such seals as information filters in order to assure appropriate and/or high quality information. SOAPs are one of the interesting products of the Interpedia effort \cite{rhine}. A SOAP is just a very short review of a resource indicating approval or disapproval according to the tastes of the person or organization issuing the review. Being short, it is easy to query in order to make view / don't-view decisions. It may have a link to another resource in order to provide a more comprehensive, human-readable, review. For the truely paranoid among us, digital signature techniques could be used to ensure their veracity, as outlined in earlier scenarios. However, this is not a necessary feature for all SOAPs. The first part of the scenario requires a group of people to run a rating service. The group might be a professional organization doing peer review of scholorly publications, a local school board looking for materials deemed inappropriate for children in the community, a lone critic reviewing Internet resources, a review group at a facility that did potentially classified research, or any of a wide number of other reviewers. When they find a resource that they wish to rate, they build a URC for it. This may be a whole new URC, or it may be strongly based on information provided by the publisher or another third party. The new URC would contain their review information in whatever format they believe best meets their needs. Clients of the rating service bear the responsibility for being able to understand the rating. Even resources without URNs could have such a 3'rd party URC established if the URC server can be queried according to URL or other information. For information filtering, the browser would have a set of active annotation servers. When the user attempts to access a new resource, its URN or URL is sent to the annotation servers. If they have any information on that resource, it could be sent back in the form of URCs. The browser could then examine the URCs, decide if the content ratings did or did not meet the criteria for information display, and either show the resource or not. In an information discovery scenario, a physics grad student wants to find high-quality information in a specialized subdiscipline of string theory. The student accesses a search page run by the American Physical Society and puts "string theory" into the subject field and 9 into the "minimum rating" field. (We assume the APS grades accepted papers on a scale of 1 to 10, a totally unrealistic assumption). The query is submitted, which searches the APS collection of URCs on papers they believed were good enough to merit their seal of approval. A list of papers is returned to the student, who then decides which ones to look at. Because of the likely popularity of such services, SOAPs will become the valuable intellectual property of the reviewing organizations. Almost all review organizations will probably charge some money to become a client of their service. The scenarios above impose two requirements on the URC. First, it must be possible to extend the URC by adding arbitrary elements. In the case above, the SOAP is the new element. Second, it must be possible to ignore elements that you do not understand. For this to be possible it must be possible to determine where any particular element ends, even if you know nothing about the structure of information inside the element. Note that there is an interaction between ignorability and having a consistent representation for the purposes of digital signatures. Digital signatures are computed over the {\it external} representation, which can include experimental elements. Ignorability is a feature of the conversion from the external to the {\it internal} representation, where if we do not understand an element we are free to discard it while we are parsing the URC. -- Ron Daniel Jr. email: rdaniel@lanl.gov Advanced Computing Lab voice: (505) 665 0597 MS B287 fax: (505) 665 4939 Los Alamos National Laboratory http://www.acl.lanl.gov/~rdaniel/ Los Alamos, NM 87545 tautology:"Conformity is very popular"
Received on Thursday, 22 June 1995 18:06:12 UTC