Loading ...
Sorry, an error occurred while loading the content.

Re: [rest-discuss] Re: Relationship between HTTP and RDF/OWL

Expand Messages
  • Roy T. Fielding
    ... That s my definition for a resource within a REST-based information system. That is because such a system creates an interface that causes that mapping to
    Message 1 of 15 , Apr 30, 2004
    • 0 Attachment
      > Roy's definition of resource:
      >
      > More precisely, a resource R is a temporally varying membership
      > function M_R(t), which for time t maps to a set of entities, or
      > values, which are equivalent. The values in the set may be resource
      > representations and/or resource identifiers.

      That's my definition for a resource within a REST-based information
      system. That is because such a system creates an interface that
      causes that mapping to be the only observable definition of a resource.
      What the resource actually *is*, within a larger scope, is a different
      topic.

      > This leads me to a conclusion I don't like: Two resources that are
      > owl:sameAs must respond to the same requests at the same time with the
      > same representations, because the resource *is* the membership
      > function.

      That depends on what you mean by "is". ;-)

      > Something smells wrong to me.

      Context.

      > Roy says that "a non-virtual object (e.g. a person)" can be a
      > resource. How is a person a membership function? I tried to
      > rationalize this by saying that the words "entity" and "value" in the
      > definition referred to the person to which the membership function
      > mapped, but the very next sentence of the definition limits the values
      > to "representations and/or identifiers".
      >
      > What am I missing?

      When you watch a movie on TV, are you watching photons, characters,
      or people playing the roles of characters? When someone makes a
      statement about something they see on TV, do they identify the
      photons, the characters, or the actors playing those characters?
      Most of the time, they try to clarify the ambiguity by setting the
      context -- by referring to the actor by their real name (or pseudonym)
      as opposed to the character's name, or by clarifying the context
      via some other part of the statement.

      So, in my disseration I was trying to describe both the set of things
      that have been named (resources) and the extent to which those named
      things can be represented within an information system (the mapping).
      It is a bit confusing because I edited it too many times.

      My definition of resource in general is in rfc2396bis.

      http://gbiv.com/protocols/uri/rev-2002/rfc2396bis.html#overview

      ....Roy
    • Roy T. Fielding
      ... We use the term resource because thing includes those that have yet to be named or described, and thus cannot be individually considered a resource by
      Message 2 of 15 , Apr 30, 2004
      • 0 Attachment
        > Yes. I'm of the opinion that the main reason we use the word "resource"
        > rather than the word "thing" is that using the former makes us seem a
        > bit smarter than if we use the latter. However the nuances of "thing"
        > are probably more apposite than the nuances of "resource".

        We use the term resource because "thing" includes those that have yet
        to be named or described, and thus cannot be individually considered
        a resource by any system.

        Short terminology may lessen universal understanding, but at least
        it saves me from typing too many phrases.

        ....Roy
      • Walden Mathews
        ... Okay, that covers the case where the HTTP + OWL scope is to be considered a larger scope than REST, and so the true definition of resource lies
        Message 3 of 15 , Apr 30, 2004
        • 0 Attachment
          : > Roy's definition of resource:
          : >
          : > More precisely, a resource R is a temporally varying membership
          : > function M_R(t), which for time t maps to a set of entities, or
          : > values, which are equivalent. The values in the set may be resource
          : > representations and/or resource identifiers.
          :
          : That's my definition for a resource within a REST-based information
          : system. That is because such a system creates an interface that
          : causes that mapping to be the only observable definition of a resource.
          : What the resource actually *is*, within a larger scope, is a different
          : topic.
          :
          : > This leads me to a conclusion I don't like: Two resources that are
          : > owl:sameAs must respond to the same requests at the same time with the
          : > same representations, because the resource *is* the membership
          : > function.
          :
          : That depends on what you mean by "is". ;-)

          Okay, that covers the case where the HTTP + OWL scope is to
          be considered a "larger scope" than REST, and so the true definition
          of "resource" lies elsewhere, perhaps.

          But even scoped to REST, where does one find in Roy's definition
          of resource a promise that two clients will ever receive identical
          representations of a single resource? The fact that they are dereferencing
          the same mapping is not enough to constitute such a promise. Hence
          how do you test the assertion </resource1> <owl:sameAs> </resource2>,
          and what good is it in the design of systems if it cannot be tested?

          Walden
        • Jon Hanna
          ... Correct in a spec or a paper, but correct in a tutorial? To many people Resource implies ownership and belonging. To many others Resource implies
          Message 4 of 15 , May 6, 2004
          • 0 Attachment
            Quoting "Roy T. Fielding" <fielding@...>:

            > > Yes. I'm of the opinion that the main reason we use the word "resource"
            > > rather than the word "thing" is that using the former makes us seem a
            > > bit smarter than if we use the latter. However the nuances of "thing"
            > > are probably more apposite than the nuances of "resource".
            >
            > We use the term resource because "thing" includes those that have yet
            > to be named or described, and thus cannot be individually considered
            > a resource by any system.

            Correct in a spec or a paper, but correct in a tutorial?

            To many people "Resource" implies ownership and belonging. To many others
            "Resource" implies membership of some sort of category. Hence we frequently get
            discussion of how <http://example.net/Ireland> could or could not be the same
            resource as <http://example.org/Ireland> when the former is "owned" by
            example.net and the latter by example.org and talk of REST resources vs. RDF
            resources vs. OWL resources vs. DAML resources.

            Since generally people don't try to say anything about things they don't have an
            identifier for I think "thing" has it's place in non-normative descriptions of
            resources (unless there are further ways in which thing is wrong).

            --
            Jon Hanna
            <http://www.hackcraft.net/>
            "…it has been truly said that hackers have even more words for
            equipment failures than Yiddish has for obnoxious people." - jargon.txt
          Your message has been successfully submitted and would be delivered to recipients shortly.