- From: Kevin Smathers <kevin.smathers@hp.com>
- Date: Tue, 17 Jun 2003 11:33:24 -0700
- To: "Butler, Mark" <Mark_Butler@hplb.hpl.hp.com>
- Cc: www-rdf-dspace <www-rdf-dspace@w3.org>
- Message-ID: <3EEF5EF4.3060706@hp.com>
More changes for technologies.tex and simile.bib are attached.
Cheers,
-kls
Butler, Mark wrote:
>Hi Kevin
>
>I have incorporated your changes, except for this line:
>
>
>
>
>>-----Original Message-----
>>From: Kevin Smathers [mailto:kevin.smathers@hp.com]
>>Sent: 17 June 2003 00:03
>>To: www-rdf-dspace
>>Subject: Comments on Motivating problems
>>
>>
>>Hi all,
>>
>>More to follow, but here are my first set of comments from a quick
>>review of the current document status.
>>
>>
>>--
>>========================================================
>> Kevin Smathers kevin.smathers@hp.com
>> Hewlett-Packard kevin@ank.com
>> Palo Alto Research Lab
>> 1501 Page Mill Rd. 650-857-4477 work
>> M/S 1135 650-852-8186 fax
>> Palo Alto, CA 94304 510-247-1031 home
>>========================================================
>>use "Standard::Disclaimer";
>>carp("This message was printed on 100% recycled bits.");
>>
>>
>>
>>
--
========================================================
Kevin Smathers kevin.smathers@hp.com
Hewlett-Packard kevin@ank.com
Palo Alto Research Lab
1501 Page Mill Rd. 650-857-4477 work
M/S 1135 650-852-8186 fax
Palo Alto, CA 94304 510-247-1031 home
========================================================
use "Standard::Disclaimer";
carp("This message was printed on 100% recycled bits.");
? patch2.txt
? relevantTechnologies/foo
Index: simile.bib
===================================================================
RCS file: /cvs/simile/docs/simile.bib,v
retrieving revision 1.26
diff -u -r1.26 simile.bib
--- simile.bib 17 Jun 2003 13:32:55 -0000 1.26
+++ simile.bib 17 Jun 2003 18:31:34 -0000
@@ -796,3 +796,13 @@
organization="MIT",
howpublished="\url{http://ocw.mit.edu}"}
+@MISC{shibboleth,
+ TITLE="{Shibboleth Initiative}",
+ organization="Internet2",
+ howpublished="\url{http://shibboleth.internet2.edu/shib-intro.html}"}
+
+@MISC{Damianou,
+ TITLE="{A Policy Framework for Management of Distributed Systems}",
+ author="Nicodemos Damianou",
+ organization="Imperial College's Policy Research Group",
+ howpublished="\url{http://www-dse.doc.ic.ac.uk/Research/policies/ponder/thesis-ncd.pdf}"}
Index: relevantTechnologies/technologies.tex
===================================================================
RCS file: /cvs/simile/docs/relevantTechnologies/technologies.tex,v
retrieving revision 1.3
diff -u -r1.3 technologies.tex
--- relevantTechnologies/technologies.tex 17 Jun 2003 13:33:09 -0000 1.3
+++ relevantTechnologies/technologies.tex 17 Jun 2003 18:31:34 -0000
@@ -496,7 +496,8 @@
The other naming problem is when we are using URLs to describe documents
and their subcomponents i.e.
identifying resources smaller than the atomic document.
-Doing this with a URL is arguably convenient, in that it permanently
+Doing this within a URL (according to some specification of the URL
+semantics) is arguably convenient, in that it permanently
binds the smaller object to its containing object, giving you the
semantics that if you are looking for the smaller object it is a good
subgoal to look for the containing object.
@@ -515,53 +516,42 @@
that depict violence, and remove them during playback of the movie.
Obviously the metadata read by the DVD player will have to include
-data that identifies the parts of the overal movie that represent the
-selected content. Using a URL to represent the content is insufficient
--- we can't create new URL's for every possible subregion of a movie,
+data that identifies the parts of the overall movie that represent the
+selected content. Using an opaque URL to represent the content is
+insufficient -- we can't create new URL's for every possible subregion
+of a movie,
and even if we did so, such an approach wouldn't help in finding an
playing back parts of the movie that do not correspond to that URL.
-
-Naming, as is being described in section 3.2.7, has nothing to do with
-the URL for the asset. The purpose of naming is to create a linkage
+The purpose of naming in this context is to create a linkage
between the metadata and the movie subregion.
-Stepping out of our example, the purpose of Naming in this document
-is to represent other assets in ways that URLs cannot. Such linkages
-are neccessarily specific to the type of data being indexed so they
-cannot be generalized to a single technology, but that doesn't mean
-that we can't create a pattern around them.
+Stepping out of our example, the purpose of Naming
+is to represent other assets in ways that opaque URLs cannot. Such linkages
+are semantically tied to the type of data that they index, so
+so they cannot be generalized to a single technology, but that doesn't mean
+that we can't create a usage pattern around them.
While using URLs with semantics is one option, an alternative way to
-specify a particular subpart of the movie is with a blob of RDF eg,
-there is a resource foo (no semantics) and assertions
-"foo fragment-of the-lord-of-the-rings", "foo start-offset 300", and
-"foo end-offset 500". Whatever semantics I intend to place in the URL, I
-can instead, without any loss of expressive power, place in a blob of
-RDF statements. This leaves me with URLs containing no semantics at
-all, which has a consistency I like.
-
-There are many different ways to represent the subgraph in question. You
-have broken it up into three statements (and an implied statement of the
-schema type), another implementation might use more statements or fewer.
-In addition there are many other types of documents that could be named,
-in whole or in part.
-
-The point of the Naming discussion is to map those statements to their
-meaning, where the meaning is a subindex into a document. This makes
-Name a specialization of Class.
-
-The issue here is not so much whether or not URNs are appropriate for
-each of the names, but rather:
-
-by what mechanism are the names generated and assigned?
-which of the URNs are URI's, and which are URL's? How can I tell?
-\begin{itemize}
-\item Here "A" could be a URL, but if I wish it to be location-independent
-I may assign a URN and use some mapping service (PURLs, Handles).
- "A" is a URL - do an http:get
-\item "B" is probably a URN, not useful to attempt to resolve it.
- I must map to some query on the contents of the graph represented by contents of "A".
-\item "C" could be either a URN or a URL. How do I find the schema? Not sure
+specify a particular subpart of the movie is with an RDF subgraph. Suppose
+there is a resource some-uri (no semantics) and assertions
+"some-uri fragment-of the-lord-of-the-rings", "some-uri start-offset 300", and
+"some-uri end-offset 500". Whatever semantics I intend to place in the URL, I
+can instead, without any loss of expressive power, place in an
+RDF subgraph. This leaves the URL free of semantics and thereby confers the
+benefit of restricting semantic data to RDF as the solitary format.
+
+Another issue here is by what mechanism are the names generated and assigned.
+Which of the URNs are URI's, and which are URL's, and how can I tell?
+\begin{itemize}
+\item Here some-uri could be a URL, but if I wish it to be location-independent
+I may assign a URN and use some mapping service (PURLs, CNRI Handles).
+ If some-uri is a URL do an http GET to retrieve the metadata needed to
+ locate the subregion of film.
+\item The predicates fragment-of, start-offset, and end-offset are
+probably URN's, it is not useful to attempt to resolve them.
+ Collectively these predicates map to some query on the contents of the
+ graph represented by contents of some-uri.
+\item The object the-lord-of-the-rings could be either a URN or a URL. Either way there must be some means of discovering the schema or type by which the object together with the semantic subgraph should be interpreted.
\end{itemize}
\subsection{Processing Models}
@@ -613,7 +603,7 @@
support automated discovery.
\end{itemize}
-It is possible to higlight this with some other processing models:
+It is possible to highlight this with some other processing models:
\subsubsection{Resource directory discovery via namespace processing model}
@@ -780,7 +770,16 @@
\subsection{Classification}
-One important issue is classification, but it has several different axes:
+One important issue is classification, that is the specification of
+the type of a referenced resource, and also of the role of that resource in
+the current context. In RDF graphs the classification role of a
+resource is obtained from the predicate that links that resource
+to the current graph. When the linked resource is itself an RDF
+graph then the type can be inferred from RDFS statements in the
+referenced graph. When the linked resoruce is not an RDF graph there
+must be some other mechanism for describing type. However, even
+when type is clearly specified, the classification of a resource
+has several different axes:
\begin{description}
\item [Metadata versus original versus abstract object] Classification
@@ -809,6 +808,12 @@
\caption{\label{dissemninationdiagram}Dissemination}
\end{figure}
+Dissemination implies a requirement for transformational repurposing
+of the data and metadata stored within the system. A graph might be
+represented to a web browser using a directory metaphor, or it might
+be transformed into embedded HTML to start a multimedia player to turn
+MP3 data into sound.
+
\subsection{To Humans}
Current thinking is to have an ontology describing how metadata is
@@ -818,7 +823,63 @@
\subsection{To software / agents}
-Policy-compliant dissemination
+The most successful strategy for defining application to application
+communication formats, permissions, and policies, has been application
+specific. If a web site wants to share a service with another web site
+then the authors of the respective sites agree on a set of ad-hoc
+interfaces. Web services standards such as SOAP and XML-RPC are starting
+to make a dent in these custom interfaces, but the policy and security
+implementations have remained the custom work of each service provider.
+
+\subsubsection{Security and Policy}
+
+One approach to a universal view of distributed security is to express
+permissions or policies in terms of split capabilities (see Alan Karp's
+tech report). By communicating in terms of these capabilities cooperating
+systems can enforce access controls that have been compiled from policies
+which they don't themselves understand.
+
+A more pressing issue is not how to implement access control in general,
+but rather, how to express and enforce dynamic access control policies over
+RDF graphs.
+
+Security must be tightly coupled with Simile's access policy and event
+management mechanisms. That is, access policies, event mechanisms,
+and security must be interdependent. In particular Simile must define
+how one can use policies to restrict access to (RDF) subgraphs. For
+example, given a set of statement identifiers (e.g., reified statements),
+how could one use a domain expression to specify a subset of the set
+to be the target of a security policy?
+
+By extension, Simile may need to express and enforce ``implied''
+policies. For example, a new statement
+being added to a store could cause a policy to now apply to existing
+statements.
+
+Note that the question of how to express access policies
+for RDF is different from (and more interesting than) the question of
+how to express access control policies in RDF. The latter is simply
+a presentation issue, and is common to all uses of RDF, while the former
+strikes at the heart of what security means as applied to RDF.
+
+Simile has not yet committed to a model for its security, or for that matter for its policy mechanisms. In particular, the following issues need to be explored in the context of Simile:
+
+\begin{itemize}
+\item How should we express policies over RDF subgraphs? (This seems a particularly rich area.)
+
+\item How can we enforce policies over RDF subgraphs? (This seems a particularly rich area.)
+\item What should be the granularity of access policies? (e.g., resource, subgraph, model?)
+\item What should be the scope of policies ?
+\end{itemize}
+
+We all agreed that from Genesis's perspective, the granularity could be model-based. That is, that if Genesis provided a means of controlling access to a given model, then Simile could use it to implement its own finer-grained mechanisms.
+
+There are to two specific pieces of related work that are interesting:
+
+\begin{itemize}
+\item The first is A Policy Framework for Management of Distributed Systems \cit{Damianou}, the thesis of Nicodemos Damianou, whose advisor at Imperial College, University of London, is Morris Sloman (and who has also collaborated with Dr. Emil Lupu). Daminanou is a member of the Imperial College's Policy Research Group, which has proposed Ponder, an object-oriented, declarative, programming language for specifying distributed system management and security policies. The Policy Research Group also provides associated tools for editing, compiling and managing policies in a distributed system. (Note that Francisco Garcia from Agilent is a program co-chair for the upcoming IEEE workshop on Policies for Distributed Systems and Networks, where last summer's SEED Lalana Kagal has a paper.)
+\item Also of interest is the Internet2 Shibboleth initiative \cite{shibboleth}. The goal of Shibboleth is to develop an open, standards-based solution for organizations to exchange information about their users in a secure and privacy-preserving manner.
+\end{itemize}
\section{Distributed Resources}
@@ -991,12 +1052,15 @@
For example it may hide the syntax used to express the schema, such as
RDFS or OWL from the user and instead present the schema graphically.
+Codified best practices, whether in the form of processes or wizards,
+will go a long way toward making the schema system easy to use.
+
\subsection{Simplify}
Applying complex classification schemes on resources could negatively
impact users' ability to search for resources. It
-is important to hide unnecessary detailS until userS need it. This
-may be done in several ways:
+is important to hide unnecessary details, and then to introduce complex
+operations gradually as users need it. This may be done in several ways:
\begin{itemize}
\item By providing default behaviors that allow users to carry out typical
@@ -1016,7 +1080,7 @@
for a user to repeat the same query multiple times on different systems.
\item It may be desirable to provide mechanism to allow users to update
several records simultaneously as a set rather than individually when
-performing instance versioning.
+performing schema versioning.
\item If possible, tasks like merging of records or mapping between schemas
and vocabularies should be automatic and only require user intervention
when absolutely necessary.
@@ -1039,17 +1103,29 @@
other uses and then making recommendations based on the items other
users have searched for.
-There are some limitations with the current versions of such systems.
-Most notably they have no way for a user to denote the context for
-their search: therefore on Amazon a user may search for very different
+Such systems do have limitations. Most notably it can be difficult to
+infer the context for a search: therefore on Amazon a user may search
+for very different
items if they are purchasing an item for a relative compared to when
-they are purchasing items for themselves. Therefore making recommendations
+they are purchasing items for themselves. Making recommendations
based on the entire users history may not be as effective as making
-recommendations based on recent search terms from the user. Also there
-are potential privacy issues that need to be addressed when recording
-user behavior, whether it is occurring with or without their knowledge.
+recommendations based on recent search terms from the user.
+
+Also when combining user behaviors into groups it is important for
+many applications to protect privacy. When recording
+user behavior, whether it is occurring with or without their knowledge,
+the system must be able to remove any personally identifying information
+from the collective behavior predictors.
+
+Other factors include whether to have users self-categorize, or to
+infer categorization from a best fit of behavior. There must also be
+mechanisms for defining and refining these categories.
\subsection{Policy Expression}
+
+In user interfaces policy expressions need to be translated into human
+readable text that simply and concisely informs the users of their
+rights and limitations with respect to accessed data.
\subsection{Misc}
Received on Tuesday, 17 June 2003 14:34:19 UTC