[topicmapmail] Not just Temporality in Topic Maps, but a whole approach

[Murray Altheim]

Quoting Simon Grant <asimong at btinternet.com>:

> Thanks to Murray for more copious, helpful and interesting comments.

You're welcome. I'm thankful to hear they were at all helpful.

> A few things I'd like to pick up on below; but first a question:
> what sense of identity does this list have? What it is, who it is for,
> what kind of people belong to it, what are the topics which are
> discussed here, as opposed to those which are not?

As Bernard says, there is no "list identity" that I know of. It's a list
devoted to discussion of Topic Maps, that's about it. There's another
list for the ISO committee to do their work; this one is for general
discussion about Topic Maps. Full stop.

> I am still sorry that I wrote as I did earlier, which was interpreted
> as a (not necessarily complementary) view of the list members - of
> course it was not intended that way at all. But it might be helpful to
> consider what may have contributed to my mistake. I'll tell you
> straight what would have avoided it altogether: a public topic map of
> the subjects of interest to people here.

That would require that someone build that Topic Map and the service
that provides it online. We are (so far as I know) mostly industry
professionals, and few have the kind of time or energy to build such
things.

> In the absence of any such
> "face", it is very easy to project one thing or another onto the list
> and its concerns. I was imagining things, not in any prejudicial way,
> but just externalising too much of what people will inevitably go
> through when confronted by a blank sheet. They go on impressions that
> may be quite misleading. What I was expecting back was not to be
> labelled as a potential threat, but a simple correction, either along
> the lines of - actually, there are people who are interested in this  -
>  or alternatively - not here, find them elsewhere. After all, that's
> quite a common thing for a list - to know their own boundaries and
> neighbours, to route people on to the places where they can interact
> most fruitfully.

Well, not knowing you we can only interpret what you write. What you
wrote was full of unwarranted assumptions. I was merely trying to
provide my opinion on the matter. It wasn't meant as a threat, any
more than yours was meant prejudicially. We are both operating in
ignorance of the other person and their personality. Email is... etc.

> At present I am still left with questions unresolved in my mind - what
> questions is it permissible to ask about appropriate topics? Do people
> subscribe to the idea that the best way to interact with this (or any)
> list is to raise a topic, and simply go away if no one joins in? No
> doubt it is one way of doing things.

I've been on mailing lists since the mid 1980s and I see little
difference between this list and any other. It's about Topic Maps.
Questions off-topic will be generally ignored unless they perk up
someone's ears, and will be answered when and if people have the
time and energy. Whether one chooses to stay or leave is up to
them. No different than anywhere else.

> At 00:35 2006-06-12, Murray Altheim wrote:
>> But both he and Kal are correct -- there is no
>> free lunch, you don't get anything for nothing, well, not quite
>> anyway.
>
> I wasn't asking whether there was a free lunch, I intended to ask
> 1. whether people shared the view that TMs are in some way closer to
> human mental representation (which, incidentally, is used as a
> marketing point by Ontopia)

I don't read Ontopia's marketing materials. As Jack and others have
already answered, Topic Maps are just another way of expressing a
graph containing identifiers for real world subjects. They aren't
magic, despite marketing language. How they are applied is entirely
up to who uses the technology. While it has a foundation, it's in
the end just a technology, not something like FOL or Conceptual
Graphs. It's open to being used in pretty much any way.

> 2. whether people were interested in discussing which ways of using TMs
> are closer to, and which less close, to the human way of thinking about
> things.

I think this is a non-starter. Can you provide any kind of metric
that would demonstrate that any technology that could do that? I
don't myself know how to measure what "close" means. If you mean
using natural language rather than a symbolic language, then no,
Topic Maps provide no easy solution. Again, something like Cyc is
probably your ticket since it includes natural language processing
features.

> Put like this, is there anything wrong or strange or inappropriate with
> my questions? If it's a matter of "find out more first", I'm perfectly
> happy to learn what is needed to get deeper into exploring the
> questions myself (indeed I am progressing that right now). It's always
> helpful to be directed to just what needs to be learned.

No, I don't think there's anything wrong with your questions, but they
were couched in a way that made a number of a priori assumptions both
about the group and about its response to your questions. While there
are mailing lists that have a significant number of people willing and
able to act as mentors to people new to the community, of the people
that I know on this list, nobody likely is going to have the time to
tutor anyone. There was a nicely designed wiki for this a few years
ago but it got spammed out of existence. The Topic Map community is
small and due to internal strife is unlikely to get much larger. There
simply aren't people who aren't often *extremely* busy with their
existing workloads to spend a lot of time on the list.

> One of the things that was clarified for me today is that in TMs, the
> concept of "internal occurrence" is (or at least can be) used to attach
> any data to a topic that is neither an external occurrence nor
> represented as an association. I think that is rather confusing, as the
> concept of "internal occurrence" diverges quite sharply from a natural
> language understanding of the term "occurrence", compared with the very
> natural way it is used in external occurrences. I wondered what
> justification there is for using the same term. Whatever else, that
> partly clears up my own questions on temporality.

There is no "internal" and "external" occurrence in the spec.
Occurrences can be either URI references to resources or actual
String resources embedded in the XTM syntax. Noting that the
external reference could be to a String as well. That's all
there is to that. Since Occurrences are sometimes used as the
way to assign properties to Topics, it's nice to be able to
store them in the XTM rather than be required to have a separate
file to contain them.

>> As to being "aligned" to human thinking, I think (without trying to
>> appear obtuse) that it's an unanswerable question. Whose "human
>> thinking" do you mean? (and pardon me, but I think "human" is redundant
>> here) And how would one determine any alignment? Epistemologically
>> it's impossible.
>
> I disagree on both points. Certainly, higher apes have some similar
> cognitive processes, so "human" is not redundant. And anyway I'd not
> like to prejudice questions on AI in general. Easier to be specific. As
> far as different people go (indeed, different people tend to structure
> their knowledge differently), well, it could be seen as an interesting
> research area.

In the scope of this discussion the ability for apes, whales, dolphins
or ants to think is not pertinent. I was speaking epistemologically,
not zoologically.

As to "alignment with human thinking", while it may be an interesting
research area (and one that has had a great deal of activity for at
least 2500 years), it's demonstrably impossible (or should I say
intractable) to deal with in a computer system. Most of the work I've
seen done in the field of AI is surprising ignorant of anything to do
with epistemology, which is one of the reasons I'm no longer pursuing
a doctorate in the field. Engineering without thought as to how it
applies to thought seems a bit hard to swallow, but given that actually
dealing with the epistemological issues head-on would basically shut
down the research this isn't surprising. IMO.

> Determining alignment: well, I can easily design a thought experiment
> whereby people are presented with different structures, and the time
> taken to learn something sufficiently to answer questions on the
> subject is measured. Is there any reason why that shouldn't be the
> basis of a respectable experimental design? That kind of experiment
> might have interesting implications in the field of learning material
> design - but I agree there are no quick, simple, or obvious solutions.

That thought experiment would only provide a statistic, not anything
to do with any theory of "alignment to human thought" from an
epistemological standpoint, i.e., it doesn't bear whatsoever on what
is actually happening, only the interpretation of the interpretations
of your test subjects. You don't get past that impasse using statistics.
Given your interest/background in HCI this seems to be the kind of
approach you're comfortable with, and that approach is a common basis
for respectable experimental design, and it might have interesting
implications. I'm not sure what all that means though. Again, I'm not
trying to be obtuse, I just don't buy into the notion that such
deductive HCI research is getting us any closer to any answers of
questions like "alignment to human thought." I don't know of anything
that does.

>> If you mean most closely associated with what some people call "common
>> sense" reasoning, the informal way that people generally view and
>> express ideas about reality (not implying here that there is one way),
>> then what you're leaning towards is in the area of my own research,
>> i.e., the cross between formal and informal modeling, where one might
>> build an application that has enough features of formal models to be
>> useful but not stray so far into formality (at least on the surface)
>> as to make it difficult for lay people to use.
>
> Sounds like a very interesting research area. As with all of these
> things, it's difficult to make a judgement about overlap in research
> interests without a much deeper discussion.

Knowledge representation, ontological engineering, computational
linguistics, etc. are complex and interesting fields that have
their own journals, conference schedules, and communities. They
also overlap a great deal.

>> Again, here there is no free lunch.
>
> I'm not clear what the free lunch would have been, had it existed. Has
> anyone suggested there is a free lunch here?

My repeated use of the phrase was meant to imply that no simple
system produces complex results, at least in this field. I used
Cyc as an example because it is an extremely large and complex
system that has taken decades and hundreds of thousands of hours
to build, and it doesn't approach an "alignment to human thought,"
unless you are willing to believe the marketing fluff. It's an
interesting and valuable project but even to use it in a Topic
Map system would require a substantial amount of work.

>> I hope your application
>> isn't air traffic control, nuclear energy,
>
> I've been closely involved with research in the issues underlying
> complex system control and human error, but I've never developed a
> close interest in Cyc.
>
>> HCI doesn't address modeling directly, but it requires an underlying
>> model.
>
> What kind of model? Chapter 2 of my PhD thesis is rather old now, but
> did address general questions about "mental models" in some depth. I
> wonder how far the field has moved on. I did write some relevant papers
> in subsequent years... :-) Perhaps you could send me links to any
> articles you have written around the topic, so I have a better idea
> where you're coming from.

I have not written any papers in HCI, nor am I pursuing a career as
an academic.

>> If that model is weak, the interface is similarly weak. If you
>> choose to have no formal, consistent model, you'll have no consistent
>> interface, just confusion.
>
> Formality and consistency are not what humans engaged in interaction
> (as opposed to theorising) are best known for. So I wonder about what
> is being imagined here.

That is precisely what I was asking you. I don't know what you
are imagining.

>>> So really, what I would be interested in knowing is, how many people
>>> around here are interested primarily in semantics and human
>>> comprehensibility of TMs, and would like (for instance) to explore the
>>> ways of using TMs to represent human knowledge in ways that *feel* most
>>> comfortable to the people involved. My guess is that the majority here
>>> are interested in the formalities, the logic, the proofs - maybe in
>>> something close to mathematical elegance. Again, I have nothing against
>>> that in itself. Just that, for pragmatic applications, for
>>> communication, it is really useful to have representations that are as
>>> intuitive as possible. That is a different kind of constraint to the
>>> set of formal and logical ones.
>>
>> I have to agree with previous comments. It's an inappropriate
>> question;
>
> Why?

Can you provide a metric of human intuition or a way to provide
an "intuitive constraint"? If you can I'd quickly patent that
if I were you.

>> your presuppositions are wrong.
>
> What do you assume those presuppositions to be? It would be very
> helpful to spell this out - perhaps then we can really clear the air.

This conversation is going in circles.

>> This is a diverse
>> community and speaking with some experience of the matter, there
>> are as many different views here as there are people. Even given
>> that, I don't think that the majority are that interested in
>> formality and logic per se, they are more interesting in modeling
>> in an applications (rather than theoretical) setting. It's not a
>> group of mathematicians, though they are represented here too.
>
> Fine, I can accept that the list has no clear positive sense of
> identity, if that is what you are implying.

As above, it's simply a mailing list for discussion of Topic Maps.

>> What you're calling "pragmatic applications" strikes me as a rather
>> strange notion, particularly as the term "pragmatic" rings with a
>> lot more meaning than you probably intended.
>
> Maybe, but I can hardly be taken to account for other people's meanings
> unless that is part of common understanding. I suggest just taking the
> concept relatively flatly - yes, one that does something useful. I
> intend no more.

It's difficult to have a meaningful conversation without shared
understanding of terms. I was referring to concepts of pragmatism
and pragmaticism of James and C.S. Peirce, where the term derives.

>> What is a "pragmatic
>> application"? One that does something useful? One that performs
>> some sort of logic-based inferencing? Or one that does some sort
>> of intuitive magic, making machine inferences lacking any machine-
>> interpretable statements? The latter has been a broken promise of
>> AI for many decades, though I still see jokers trying to sell this
>> notion to the ignorant*.
>
> Hmm, an interesting view.

I'd be interested in seeing an AI application that fulfills the
promise, more than simply a Turing application (i.e., the ability
to fool).

>> As to intuition, there is no one generalized human intuition, so there
>> can be no one generalized model. That idea died about 2500 years ago.
>> I'd be interested in hearing your ideas on intuitive constraints.
>
> Not sure - I don't fully understand what is being asked.

Can you provide a pragmatic example of an intuitive constraint,
one that can be programmed into a computer? Even hypothetically?

>>> Is there anywhere else where people of the kind I am talking about  
>>>  hang out?
>>
>> No, sorry. You can choose a community to hang around with, but I've
>> never seen one that was of (relatively speaking) one like mind,
>> except religious cults. The closest you might find to the kind of
>> informality-without-baggage you seem to seek is probably the
>> tagging and folksonomy types such as at del.icio.us. But those
>> "systems" (and I use the term loosely) are simple classification
>> systems, not ontological models, and they raise a lot more problems
>> and issues than they actually solve, a free lunch perhaps, but
>> likely to upset your stomach if you look to them as nourishment
>> rather than simple entertainment...  like the McDonalds of KR.
>
> Ah, here I agree... :-)
>
> But how are people meant to make a reasoned choice of community
> to hang out with?

I've never known anyone to make a "reasoned" choice. I've always
done it based on knowing or enjoying the company of one or two
people, or because I was trying to learn something. Maybe that
is "reasonable." This list has a lot of lurkers and few active
participants; it's a relatively quiet list. It is either of
interest or it is not to any given person based on what I assume
are a myriad of reasons -- I don't assume any consistency. I'm not
sure what I derive from being a member, except that it maintains
a connection with some people I know and value as both professional
associates and friends. What anyone else uses to make such choices
is up to them. I don't think it's the responsibility of "the list"
to provide any rationale.

Murray

...........................................................................
Murray Altheim <murray06 at altheim.com>                              ===  = =
http://www.altheim.com/murray/                                     = =  ===
SGML Grease Monkey, Banjo Player, Wantanabe Zen Monk               = =  = =

       In the evening
       The rice leaves in the garden
       Rustle in the autumn wind
       That blows through my reed hut.  -- Minamoto no Tsunenobu