Re: JAR's exploration of TimBL's notion of information resource

On Tue, May 26, 2009 at 8:35 AM, Jonathan Rees <jar@creativecommons.org> wrote:
> On Tue, May 26, 2009 at 2:39 AM, Alan Ruttenberg
> <alanruttenberg@gmail.com> wrote:
>> On Wed, May 13, 2009 at 10:34 AM, Jonathan Rees <jar@creativecommons.org> wrote:
>>> JAR's exploration of TimBL's notion of information resource.
>>>
>>> [The below does not constitute an endorsement of any particular theory
>>> of generic resources or information resources - especially not the one
>>> described herein.]
>>>
>>> Terminology / notation:
>>>  generic-resource = GR = 'information resource' in Tim's sense [1] as
>>>    JAR currently understands it
>>>  wa-representation = 'representation' in the REST or web architecture
>>>    (AWWW) sense (NOT in the Xiaoshu or plain-English sense that
>>>    permits, say, a rock, or a citizen, to be represented)
>>>
>>>  G = some generic resource
>>>  Z = a multidimensional parameter space
>>>    (e.g. time * language * content-type * user-agent)
>>>  P = a point in Z
>>>
>>> So far all we know formally is that there is a 3-way relation
>>>
>>>  G has wa-representation R at point P
>>>
>>> That is, for each P = tuple of parameter values, there is a set
>>> (possibly empty, or quite large) of wa-representations with the
>>> property that they are wa-representations of G for the parameters P.
>>>
>>
>> Every wa-representation is of some G at some point P, yah?
>> Just trying the get the all-somes on the classes clear.
>
> I don't know. You'd have to ask Tim as I don't think this follows
> from what he's said so far.

No, it doesn't. But  was trying to parse "G has wa-representation R at
point P". You are obviously talking about generalities, rather than
some specific G. Is this a compact way of stating domains and ranges?
Alternatively as a relation is defined as the set of tuples which
satisfy the relation are you saying that that the 3 way relation is a
subset of G x R x P where G is the universe of generic resources, R of
all representations and P of all parameters?

Do you see the question? It's not about what's true, it's about what
you are saying.

>>> In Tim's theory we know that Z has at least three axes (time,
>>> language, content-type), maybe more (user-agent, authorization,
>>> Russell 2000 index).  We know that a wa-representation can belong to
>>> the trace of more than one GR, and that a GR can have, at one point P,
>>> more than one wa-representation (as would e.g. Moby Dick).
>>
>> Another check in: that a wa-representation can belong to the trace of
>> more than one GR is a consequence of that fact that a
>> wa-representation of a more "specific" resource is also a
>> wa-representation of a more "generic" resource? I mention this because
>> if this is the reason it implies more structure than you've elaborated
>> in the above. If there's another case then it's worth noting.
>
> Well, it does follow from this yes, but also because there is
> nothing to rule it out.  Another case is two versioned documents
> where at one point in time one of them happens to have the same
> content as what the other one had at a different point in time.
> Same wa-representations, different resources.

OK, so, for example, representations of books by blocked writers
before writers become unblocked, consisting of "Once upon a time",
text/plain

>>> As determined on the call, there is nothing that formally rules out a
>>> "bottom" GR that has no wa-representations (trace is everywhere
>>> empty), or a "top" GR that has *all* wa-representations (i.e. GR
>>> has wa-representation R at P for all R and P, or trace is everywhere universal).
>>> The latter may be useless, but not nonsensical.  In fact it may be the
>>> case that given an *arbitrary* trace, there is (or could be?) a GR
>>> with that trace.  This is not essential in what follows, but it would
>>> be nice to know, if it is not true (ontologically), why it isn't -
>>> what kinds of traces *do* not have corresponding GRs?
>>
>> Assuming my speculation about the reasoning behind wa-representations
>> is wrong, and the term "generic resource" is instead to be taken to
>> name classes, then rdf:Resource is your "top" GR.
>
> No, clearly generic resource (which includes GR) is a proper subclass of
> rdf:Resource, since there exist things that aren't generic resources
> (in Tim's sense).

Apologies - I should have said "Information Resource" as TOP.

>>> Suppose that two parties, Alice and Bob, get together.  Between the
>>> two of them they somehow agree to talk about a particular
>>> generic-resource (such as Moby Dick generically, or perhaps the
>>> Penguin 2001 edition of Moby Dick, or a GR having as its sole
>>> wa-representation one whose content-type and content are those from
>>> [2] with content checksum 137aace70c30eb076407cf28bd78b884), which
>>> they between themselves call G1.  Suppose that they agree on what G1
>>> is to the extent that they can each separately distinguish
>>> wa-representations that are wa-representations of G1 from those that
>>> aren't, for any parameters P, and do so with perfect agreement - that
>>> is, they both have full knowledge of the trace of G1.
>>
>> Careful. You are giving two scenarios. One in which they know what G
>> is (e.g. because they meet and point at it) and another in which they
>> only know the trace. The latter doesn't imply the former unless you
>> assert that a trace is unique to a resource. Which you don't/
>
> No, the assumption of shared understanding of the trace was an
> additional assumption, not an alternative assumption.

OK.

>>> We don't yet know what additional
>>> characteristics would be sufficient (essential) for determining
>>> sameness, and since these characteristics must be message-conveyable
>>> (according to AWWW), it will be very interesting to learn what they
>>> are....
>>
>> What you say above is asserted to be true only of Information
>> Resources, if I follow the argument, since it is the n&s condition for
>> being an information resource. So at this point I am assuming that we
>> are only talking about IRs. Is it obvious that generic resources are
>> information resources? (might be but I don't believe it's stated in
>> the docs)
>
> Tim has said that it was his intent that generic resource = information resource.
> If this is true then questions about conveyableness arise. But these are not
> important right now.

OK.

>>> OK, now suppose that S is an HTTP server, and G is a generic-resource,
>>> and U is a URI.  Define "S is consistent with G at U" as follows:
>>>
>>>  if whenever S receives an HTTP GET request with request-URI U and
>>>    responds with a 200 response is received,
>> *and* ?
>>>  the RFC2616-entity in the 200 response is a wa-representation of G,
>>>  then S is consistent with G at U.
>>
>> S is a ___ ?
>
> Server
>
>> Server? Whatever it is, it can "receive" and it can "respond".
>> I'm not sure that "consistent" is adding anything here. By the
>> architecture and what I've heard if the server responds 200 then the
>> entity *is* the representation, so the antecedent is a tautology.
>
> Hardly!  If for U the server responds with a wa-representation X that is
> *not* a wa-representation of the resource in question G, then it is *not*
> consistent with G at U. That is, some *other* resource might be
> served by S at U.

This assumes that we (or the server) can know what resource a server
is interacting about. Yes?
Is that obvious?

> Another way of saying this: Is there any evidence from S against the hypothesis
> that U wa-identifies G?

Right, and I have no way to assess (so far) what is or isn't evidence
other than the response code being 200.

> Obviously different servers can produce different wa-representations
> at a given URI, and only some of these will be wa-representations
> of any given generic resource G.

This needs to be an axiom afaict. Something like the axiom of choice.
Here you are saying it is possible that servers generate
wa-representations of resources other than the one that is requested.
I don't think that this can be concluded from anything that has been
said so far.

> In fact different a server S1 can be
> faithful to G at U, while a different server S2 can be *unfaithful* to
> (or inconsistent with) G at the same URI U.  (Consider an evil proxy server
> intent on giving you misinformation about some generic resource.)
>
> This may be trivial, but it is not tautological.

Don't know. All I know is that faithful(x) can seemingly only be
asserted by an oracle. It sounds like this is some sort of assessment
we are able to make within the framework you are proposing and I don't
see it.

>> Well, this suggests an actual link between the representation and the
>> resource, (sort of a kr-representing kind of thing) that hasn't been
>> part of your account as of yet. So I don't know how you form any
>> hypothesis. On what basis?
>
> I don't get this. The whole account is premised on there being
> a relationship between wa-representations and generic resources,
> as in geometry one posits axiomatically an incidence relation
> between points and lines.
> Remember the goal in this exercise was only to consider the generic
> resource framework
> axiomatically, since an ontological treatment has repeatedly failed. So
> the kind of analysis you like to do simply can't apply yet.

I guess it hinges on the status of "faithful", "consistent".

>>> The nice thing
>>> about requiring only consistency is that it is equivalent to merely
>>> adding triples asserting that the wa-representations observed from S
>>> are representations of whatever U is interpreted to be, which seems
>>> almost tractable.)
>>
>> What kind of representations of U? What is a criterion for such a
>> representation. Seems you've moved to kr-representation at this point.
>
> Nope, still totally in the axiomatic realm, per program. I don't
> follow your reasoning.

OK. Might be useful to have some triples at this point.

-Alan

Received on Tuesday, 26 May 2009 13:51:48 UTC