Re: [xtm-wg] parallel development of syntax and concept models

[Jim Farrugia <jim@spatial.maine.edu>]

Ben,

Thanks for this.  I include your comments, as well as a paragraph
from Michel, to provide a context for further questions...  Jim


BEN SAID...

>This took me a while to grasp too.  Basically, topics are nodes in
>topic-map space which contain links to things *outside* of
>topic-map space.  They are the endpoints of association links, which
>cannot point outside the topic map.


MICHEL SAID

The Hyperlink Perspective
=====================
The "Hyperlink perspective", also (abusively) called "syntax", says
basically: Topic maps are made of 2 hyperlinks: one is called topic,
another one is called an association. The topic hyperlink has a
supplementary property: it has name(s). The association hyperlink has
a supplementary constraint: it can only connect topic links. Both
hyperlinks can be scoped. That's basically it.


JIM ASKS....

So, from the Hyperlink Perspective then, there are only three types of 
animals being talked about?

things, qualifications of things, and links (also, I guess, a certain kind 
of thing).

         1) Things are either topics (verbalizations of ineffable subjects)
         or (topic) occurrences - the real-world "realizations" that are
         relevant to a given topic;

         2) Qualifications of things are statements or rules or constraints
         that apply to things or to links.  So, for instance, the 
supplemental property
         that a topic (usually) has a name is a qualification of a thing.
         So, too, would be the supplemental constraint that an association (a
         certain kind of link, see below) must link a topic with a topic.
         And the ability of topics and association to be "scoped" is also
         perhaps a qualification.

         3) A link is, for lack of a better description, that which connects
         two things.  The two things can be (topic, topic), which are
         connected by a link called an association, or the two things can
         be (topic, occurrence ?)  Is there a link that connects 
occurrences with        occurrences?  If not a direct one, can this done 
indirectly?
         Does it make sense to link occurrence with occurrence?

         4) According to what Ben said above, it seems accurate to speak
         of a topic (possibly) "having" two kinds of links: a link to 
another topic
         inside the Topic Maps space (that is, an association), or a link from
         a topic (inside the Topic Map space) to something, not a topic, that
         lives outside the Topic Maps space.  It seems that this other end
         of this second kind of link is an occurrence, since the standard says
         at one point, "Occurrences are the anchors of the topic link."


THE STANDARD SAYS ....

The standard gives two definitions for topic:

         An aggregate of topic characteristics, including zero or more
         names, occurrences, and roles played in associations with other
         topics, whose organizing principle is a single subject.
also...
         A topic link element.

The standard further says:

         The topic link (topic) element form is used to assign topic name
         characteristics and topic occurrence characteristics to a topic.
and ...
         A valid topic link must have at least one of the following: a topic
         name, a topic occurrence, or a role played in an association
         with at least one other valid topic.


So, ...

JIM ALSO COMMENTS AND ASKS ...

         5) There seem to be two particular ways to talk about just what
         a topic is:  one is the familiar refrain that a topic must have at 
least one    of the following: a name, an occurrence, or a role played in 
association with   another topic. The other is the notion that a topic is 
in some sense both
         a "node" and a "link" between nodes, which is what Michel and Ben 
seemed
         to be suggesting today in their responses to my original questions.
         (Below I refer to the former as "the first way" of understanding 
what a topic   is, and to the latter as "the second way" ....)  I think it 
is worthwhile to    keep these two ways of understanding a topic in mind, 
and perhaps come up       with an elegant synthesis of these two ways of 
looking at things.  Just a       thought.

         6) Is it fair to say that, in some sense, the topic link element form
         is, informally speaking, the abstract box that houses only one of 
the three     pieces that make up a topic (in the sense of the first way of
         understanding mentioned above)?  Specifically, one can find a 
topic name
         in the topic link element form.

         7) Also, in considering the dual nature of topic as both node and 
link,
         (the second way of understanding what a topic is), is it fair to 
say that the   topic link element form, by itself, deals with only the 
notion of "topic as     node", and that it deals with this via the topic id 
or name ?  That is,
         within the topic link element form, there is no mention of linking 
(between     topics and topics or between topics and occurrences).  Instead,
         the two ways to think of "topic as link" are actually specified 
elsewhere.
         Specifically, the link that connects topics with topics is 
specified in the     association link element form, and the link that 
connects topics with   occurrences is specified within the topic occurrence 
architectural form.  Is    this accurate? If so, it helps explain why this 
seems confusing (that there is  no linking done or specified within the 
topic link element form).

         8) And so, in summarizing these last thoughts, we can consider a topic
         either as being made up of at least one of name, occurrence, role 
played ....   or as a "node/link" duality.   When we look in the spec at 
the topic link       element form, we can see just one piece of each of 
these ways of        understanding topic:  we can see the (optional) name 
in the first way of        understanding, or we can see, say, the topic id 
as representing the node in     the node/link way of understanding 
topic.  But we can't see, *in the topic link         element form itself* 
any of these other pieces of these two ways of     understanding. 
Specifically, in the first way of understanding, we don't find 
in        the topic link element form any mention of "topic as occurrence 
or role         played ..."; and in the second way of understanding, we 
don't find the notion
         of topic as link.  To complete the picture of both understandings we
         need to look outside the topic link element form itself.   Is all this
         accurate?

         9)  The notion of "topic as link" in the "node/link" duality way
         of understanding topics, seems to be partly dealt with
         in the topic occurrence [element form?].  It is here where one
         kind of link is treated - the link between a topic and its occurrence.
         This same topic occurrence [element form] seems also to
         deal with the second of the three pieces (i.e., occurrence) in the
         first way of understanding a topic - that a topic consists of a 
name, an        occurrence, or a role played .....   Is this accurate?

         10) The rest of the "topic as link" in the "node/link" duality way
         of understanding topics, seems to be dealt with in the
         association link element form.  (I have some questions about
         just how such a link is established between topics using this
         form, but ...)

         What is still dangling for me is the missing piece of
         "role played in association with other topics" in the first way
         of understanding what a topic is.

         11) What is the key insight behind the association role that makes
         it, in the first way of understanding what a topic is, such a key
         component?  And how does an association role relate to the
         "node/link" way of understanding topics, if at all?


So, in thinking of a topic as both a node and a (topic) link (element form),
we note the following (this should all be understood as a question;
please correct me where I'm wrong):

         *) a topic as node can be seen as an id or name, via the topic 
link element     form

         *) a topic as link can be seen partly in the topic occurrence 
[element  form??]. This link connects topics (inside the topic map) to 
occurrences        (things outside the topic map).  A topic as link can 
also been seen
         partly in the association link element form.  This kind of link 
connects
         topics to topics.

         *) The topic's name will live in the topic link element form, but the
         (description of         a) topic occurrence will be treated in its own
         area, in the topic occurrence [element form?].  The "role played
         in association with other topics" also lives outside the topic link
         element form, in the association role element form.


For those who've read this far, ... first, thanks.  Second, ...

I'm trying to understand things.  Please comment on whether what
I wrote makes sense and whether it actually sheds any light on things,
or whether there are better ways of getting a grip on just what a topic
is.  Finally, does any of this help the discussion of "one model, two
perspectives?"  I suspect it may not contribute directly to that, since
it seems preoccupied with the Hyperlink perspective.  But maybe those
who know, will see some distinctions worth making in the Foundation 
Perspective.


Thanks for any light you care to shed.

Jim


-------------------------- eGroups Sponsor -------------------------~-~>
GET A NEXTCARD VISA, in 30 seconds!  Get rates
of 2.9% Intro or 9.9% Ongoing APR* and no annual fee!
Apply NOW!
http://click.egroups.com/1/7872/4/_/337252/_/967501234/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com