Re: Another round of subject identity crisis c.2011

Hi Lars,

On 9 Mar 2011, at 2:27 PM, Lars Heuer wrote:

> Hi Andrew,
> 
> [...]
>> Hmmm.... I understand what you're saying, but I'm not sure I get
>> it.  Maybe it's just personal perspective warpage going on here, but
>> in this case, why is the extra level of redirection necessary?  This
>> is probably old ground, but why wouldn't you just put the metadata
>> on the thing itself ant treat the thing as a topic?
> 
> Let's try another, concrete example. :) Imagine I have created a PSI
> for the city "Bremen":
> 
>    http://psi.example.org/Bremen - "Freie Hansestadt Bremen".
> 
> 
> Now I like to attach information to the PSI, like when I created it,
> if it is stable etc. If I'd state:
> 
>    http://psi.example.org/Bremen
>        created-by(Lars-Heuer);
>        psi:status(psi:experimental);
>        created-at(2011-03-09);
>    .
> 
> I'd effectively say that the *city* Bremen was created by me and has
> an experimental status. That's obviously wrong.
> 
> The usage of a subject locator
> 
>    = http://psi.example.org/Bremen
>        created-by(Lars-Heuer);
>        psi:status(psi:experimental);
>        created-at(2011-03-09);
>    .
> 
> makes clear that the statements created-by etc. refer to the resource
> and not to the subject "Bremen".
> 
> I have two topics which describe different aspects of the IRI
> <http://psi.example.org/Bremen> and the topics wouldn't automatically
> merge (since one uses the address as subject identifier and the other
> as subject locator).
> 
> [...]

This isn't the type of example I was talking about, and yes, I agree with you that the only way to do this is via PSI's and the normal subject indicator model.

What I'm talking about are topic instances that, yes, I did in fact define or at least create to stand for a particular semantic meaning.  Alternatively, someone else did, but it's basically the same thing in my view.  I'm not talking about subjects which aren't locators, I'm talking about subjects which *are* locators.  Concrete representations of abstract concepts which may or may not have any other description of what they mean within a particular context.

The above examples don't capture this scenario at all, and they're back in what I would consider the "semantic happy place" of subject identity.  I'm trying to talk about the boundary cases (at least as I see them).

>> If I want to make type-based statements about the content from said
>> resource, I do something like this:
> 
>> dublin = http://en.wikipedia.org/wiki/Dublin
>>        isa city;
>>        population: 506211;
>>        ... ;
>>        .
> 
> Here you're saying that the *webpage* is an instance of "city" and has
> a population. I think you want to use a subject identifier here:

Am I?  How?  My understanding is that I'm saying that the item identifier dublin is an instance of "city" and that instance reifies the resource in a particular manner (in this case as the properties relating to the type "city").  Granted, the last bit is a slight extrapolation of my understanding of the former, but I don't see why it's invalid.

>  dublin http://en.wikipedia.org/wiki/Dublin
>         isa city;
>         population: 506211;
>         ... ;
>         .
> 
> would say: The subject described by
> <http://en.wikipedia.org/wiki/Dublin> is an instance of "city" and has
> a population of ... .

This is really back to my discussion a few months ago.  Your version tells me useful things, but it doesn't tell me anything about the particular assertions made by *that* page having to do with Dublin the city.  I want to capture the notion of reifying a statement (or set of statements) made about a particular subject by a particular resource as a single proxy instance.

Again, you're recast the statement back to the known, "happy path" scenario that I understand very well... ;)  But it doesn't go far enough, in my view, and it also subtly assumes you have a TMDM view of merging in play.

If I wanted to talk about a proxy that reifies an image of Dublin, who defines the rules by which the type and properties are determined?  This picture is from 1942 and this one is from 2011 of the same street.  There has to be a way to reference a collection of properties about a subject (so maybe I should've included <http://.../Dublin> as well) and tie those efficiently to the resource from which they were derived.

In my view, this is what's done by the "obvious" interpretations of saying reification of an image with a subject locator should be limited to the physical properties of the current electronic representation of that image.  It's just so obvious, we just don't think about it.  What happens with the information that representation conveys?

[however, as an aside, this part of the discussion is tending towards a tangent at this stage more relevant to my earlier emails on this issue]

>> What you're effectively saying above is that there's no way to
>> publish networked topic maps, right?  If both locators and
>> identifiers are treated as opaque, this must be true.
> 
> I haven't said that, I hope. Why shouldn't it possible to publish
> publish networked topic maps?
> 
> Let's assume you have information about Dublin published at
> <http://psi.example.andrew/Dublin>.
> 
> I have published information about Bremen at
> <http://psi.example.lars/Bremen>.
> 
> A bus tour operator uses the PSIs in it topic map and publishes
> somewhere the following topic map:
> 
>    distance(location: <http://psi.example.andrew/Dublin>,
>             location: <http://psi.example.lars/Bremen>,
>             km: km-1300)
> 
>    km-1300 - "1300".
> 
> I an agent reads the topic map, it can easily hop to your server and
> receive more information about Dublin and to my server to fetch more
> information about Bremen (assuming that we both are smart guys which
> deliver not rubbish HTML but topic maps if we're asked for the right
> media type).
> 
> [...]

This isn't really what I meant by "networked", but it does indeed use the network.  For the purposes of the map in question, though, the PSIs are opaque identifiers that MAY be resolved by the application to do either merging or additional discovery.  They MAY NOT be used by the local map as anything other than opaque, item identifiers that happen to theoretically (as far as the map is concerned) denote a subject, meaning (as far as the map is concerned) that if it sees the same byte sequence elsewhere, they should be deemed to identify the same subject.

I'm talking more about being able to extend the same concepts and rules governing item identifiers within what I would consider a "local" map to a map that was treated as a "local" map, but which actually was able to leverage remote subject locators as non-opaque identifiers - complete with full introspection capabilities - without merging (at least in the TMDM sense).

...not sure about the "smart guys" being applicable this direction, but hey, ya gotta have goals! ;)

>> Therefore, if subject locators are addressable, and, by definition,
>> they're subjects - but only until the address resolution takes
>> place.  Once you have this, you have a "something".  In my view, the
>> only way networked topic maps will work is if you can attempt to
>> introspect that something, at some level,
> 
> Yes that's possible.
> 
> [...]
>> t1 isa =http://psi.example.org/v1/mytype .
> 
>> is that legal?
> 
> Yes, that would be legal. If that's useful is another question.
> 
> 
> [...]
> 
> Sorry, I have to stop here. I'll try to come back to the rest of your
> e-mail this evening.


No worries.  Take your time. :)

Cheers,

ast
--
Andrew S. Townley <[email protected]>
http://atownley.org