[Top] [All Lists]

Re: [ontolog-forum] Visual Complexity

To: paola.dimaio@xxxxxxxxx
Cc: ontolog-forum@xxxxxxxxxxxxxxxx
From: Pat Hayes <phayes@xxxxxxx>
Date: Mon, 5 Feb 2007 18:03:10 -0600
Message-id: <p06230902c1ed45537bb4@[]>

Highly snipped reply, still long though :-)
...    (02)

>>>An ontology is first formed at conceptual level (design)  being
>>>inmplemented is the last step.
>>What do you mean by "implemented"?
>I mean that the LOGICAL THEORY does not have a single  language attached to it
>You can express a logica theory using a language, but the theory IS NOT
>the language.    (03)

The theory is not the language, but it is (and now I am quoting an 
almost universally accepted technical usage of 'logical theory') a 
set of sentences in some formal language. So any such logical theory 
has a single language associated with it, to wit, the formal language 
it is written in. So the idea of a theory which is somehow 
independent of a formal language, is (literally) meaningless. If you 
mean something else by "logical theory", it would help if you could 
explain what you mean.    (04)

>>>A  formalization (eg, in OWL, or IKL if you prefer a
>>more exotic notation) is not an implementation of
>>anything in the usual sense. Ontologies are not
>>software! (BTW, I think that this is not widely
>>recognized enough. Thinking of ontologies as
>>software is one of the recurrent motivations for
>>applying software engineering principles to
>>ontology design, which may be a mistake. Or not,
>>of course.)  <<There is a wider argument there, later-PDM>>
>'implemented'  is the physical dimension of a conceptual/knowledge formalism
>a formalisation is still implementation independent
>I think we learn this when we study systems  engineering
>a system has a functional desgin, a logical design and an implementation
>I can implement a formalisation using different languages, owl being just one    (05)

Please explain how your "formalisation" is written or specified 
without using a formal language.    (06)

It sounds as though you may be suffering from the misapprehension I 
tried to describe in my last message. An OWL ontology (for example) 
is not an OWL implementation of something else. It *is* the 
formalisation. It is not itself an implementation of anything, as it 
does not "run" or specify any kind of algorithm or behavior. It is 
data, intended to be used by things that run (and themselves are 
written in LISP, Java, Perl, etc.)
...    (07)

>>I see (reading ahead) that you are using the
>>third sense, as in "terms of the treaty", where
>>it means roughly the propositions contained in
>>the treaty (or ontology). That is not
>>conventional usage in the field, so you run the
>>risk (as we are here illustrating) of being
>>seriously misunderstood; but OK, yes of course if
>>you include the entire apparatus of the ontology
>>under the phrase "terms of the ontology" then
>>defining the terms indeed amounts to creating the
>yes I intended the entire apparatus, (was trying to use generally understood
>'term' but onvious that does not work with the very clever people)    (08)

Your sarcasm or irony here is misplaced. This usage of "term" is not 
widely used in the field (and in fact, not even in my native English 
dialect, to tell you the truth: I can recall my English teacher 
shouting at me to "clarify our terms", and he meant the word in the 
first sense.)    (09)

>>OK, I see you were using 'source of knowledge'
>>far more broadly. OK, I will agree then with your
>>requirement. But we must recognize that many
>>ontologies have no such 'source' in this sense,
>>or may represent a distillation from many such
>>sources. I look forward to a future in which
>>ontologies themselves are considered to be
>>definitive sources of knowledge, so that your
>>'broad' view and my 'narrow' view may become
>>closer in scope.
>well, I think that is the general, overall scope of an ontology
>If I want to create a system for the red cross to use, then I need to
>develop an ontology that reflects the red cross view of the world
>but if I want to create a system that works for 'any' emergency
>response, then I have to
>create a new, more neutral ontology. my problem is to reconcile
>different points of views
>to create a common language, in terms of conceptual  as well a
>semantic reference    (010)

OK, I see. We are using 'language' differently. I am using it refer 
to the notation in which the ontology is written; you are using it 
more to indicate the concepts, presumptions etc,. which the ontology 
sets out to describe and define, the 'conceptual framework'. Again, 
apparently a miscommunication.
...    (011)

>>Im really not sure what would count as "declaring
>for example:
>I mean that today, when we (attempt) to design a system that can be synched up
>with emergency providers, we are told by the experts (some experts ,
>the most experts that we can put our hands on) that the best knowledge
>source to date is the red cross, which has the most complete set of
>conceptual and sematic definitions in the world    (012)

OK.    (013)

>But I have a problem with that    (014)

Well, I might well have a problem with the 'in the world' part, but I 
see no problem with at least starting with those concepts, and 
modifying them if necessary.    (015)

>1) the red cross ontology is not publicly accessible, and if is, it is
>not visible (could not find it online)    (016)

Which is exactly the situation that the SWeb may in future prevent, 
one hopes, or at least discourage. Have you asked the RC if you can 
see their ontology?    (017)

>2) assuming I can find it, the red cross operations are not smooth,
>and not transparent, and not  necessarily efficient .    (018)

Do you KNOW this, or are you extrapolating from some experience or data?    (019)

>This is (I
>argue) also because their ontology has  been developed top down, nor
>allows anyone to provide feedback. there could be intrisic bias,
>and knowledge misrepresentation due to the point of view represented
>not being a collectivee one, of diverse communities, but a 'standard'
>one, that may not rflect the reality of an emergency.    (020)

Well, possibly: but on the other hand, one could say the same for any 
ontology or indeed almost any repository of knowledge in any form. 
Has the RC behaved in a way which would lead you to think that its 
ontologies are broken?    (021)

>I am studying this a little, and I have reason to believe that what I
>say above is true, aldthough do have results to share as such.
>So, the red cross ontology may well be the best ontology in emergency
>today, but we dont know on what assumptions it was developed (racial,
>gender, age and clas discrimination for  example? may all be built
>into the system    (022)

Indeed. But now, ask yourself: suppose this is true; is it likely 
that there will be a declaration at the top, along the lines of 
"Advised by the KKK", or "based on non-PC eugenic prejudices"? Of 
course not.    (023)

>, and the people would never know. why on eearth FEMA
>and Red Cross operate they way they do) It could be because their
>information system is designed to reflects very partial knowledge.    (024)

Again, is this likely to be part of a declaration by the authors? 
"Based on incomplete information" ??    (025)

>I think that in order to be useful and widely adopted, an ontology
>should be accessible
>visible and transparent in the sense of declaring explicitly what
>assumptions it is based on
>I hope we are still talking about the same thing at this stage    (026)

I'm really not sure. I am still very unclear quite what it is that 
you have in mind. I can't think of a more explicit declaration of the 
assumptions on which an ontology is based, than the actual ontology 
itself. That, after all, *is* the assumptions made by the ontology.    (027)

>>Declare how? What would one say? Would a
>>reference to a cultural tradition do? Or are you
>>asking for a formalized logic to be used on the
>>ontology (as I was presuming)? Or something in
>>between? What?
>havent worked it out yet
>you tell me    (028)

Hey, you are the one who wants it to be a requirement. I didn't know 
what you meant, and it seems that you don't know what you meant 
either. I suggest we just drop this requirement, as nobody seems to 
know what it means.    (029)

>>What is the purpose of this declaration? Will it
>>influence how the ontology is to be processed by
>>machines? (I presume not.) Knowing the purpose
>>might help answer the above questions.
>also knowing the hidden agendas of an organisation    (030)

But again: if an organization has *hidden* agendas, is it likely that 
it will publish them in an open declaration?    (031)

...    (032)

>"Who is"  simply tells Google to use a specialized set of
>>criteria in its next retrieval. And what you get
>>back certainly isn't natural language.
>I consider my knowledge queries on search engines a good example
>I type everyday  define:natural language or what is: bla bla
>I get a set of documents written in natural language
>by  natural language I mean 'not code' am wrong?    (033)

OK, I guess I was taking NL to mean that one can have something like 
a genuine conversation with the machine all conducted in, say, 
English. This is often cited as the goal of a NL interface/query 
service. It amounts to passing a limited form of the famous Turing 
Test. And this is what we aren't likely to have in the forseeable 
future. If you count Google-style predefined-phrase style 
interactions (and indeed considerably better, eg a smallish grammar 
for query sentences) then yes, that is feasible. But IMO it is often 
not as useful an interface as is often assumed (especially for lay 
users), because users have great trouble staying within the limited 
grammar, and these systems tend to fail suddenly, which is very 
destructive for the naturalness of the interaction.    (034)

>Okay, to learn OWL you need to have some prior knowledge (be skilled
>at some things)  or undergo the expensive training at stanford (2500
>usd) .    (035)

No, nothing like. I'm not sure, but I bet one can teach the core 
notions of OWL (which account for about 90% of the published OWL 
content) to, say, bright middle school kids in about a day, and they 
would have fun learning it. My granddaughter (age 7) understands it 
reasonably well, though she has trouble understanding why anyone 
would bother, and I havn't tried to explain 'subproperty' to her. 
Now, the OWL/RDF/XML *syntax* is another matter altogether, but 
nobody, even its designers, intended that to be for human use. For a 
(still imperfect, but) much easier representation, take a look for 
example at the OWL CMaps that COE produces. Download it from 
http://coe.ihmc.us/.    (036)

For a very simple brief survey of OWL, try reading the COE manual , 
section "COE notation and conventions", page 12 et. seq. . Or have 
you tried the OWL tutorial, available on the W3C website?    (037)

>>But my point is that any particular ontology is
>>going to be represented in SOMETHING: it might be
>>OWL or CL or GOFOL or Prolog or RDF or Concept
>>Graphs or CGIF or who knows what. But it can't be
>>represented in nothing, and it can't be
>>represented in some kind of supervening
>>?bernotation, because there is no such universal
>my point is that knowledge in any domain is first represented by concepts and
>words    (038)

Not always. While this may be true for many cases, there is a great 
deal of knowledge which is necessary for use by reasoners but is 
rarely put into words in NL, because all adult human speakers already 
know it, so they never have to say it to one another. And there are 
distinctions which are central to ontology engineering decisions 
which are almost never put into words, except by rather obscure 
philosophers.    (039)

>  -  I think grammars,  and logical diagrams, e/r notation, plus
>controlled vocabularies
>do the trick to represent ontology from the KR point of view
>choice of formalism is personal    (040)

Not if we wish to have our systems interact. The formalism in which 
ontologies are written is in fact better thought of as an interchange 
language than a personal, local "interior" choice. It is how 
ontologies are transmitted from place to place on the Web, to be used 
at the point of reception.    (041)

>, and knowledge representation shold be
>as as independent as possble from formalisms to be widely usable    (042)

Well, we really do disagree about this. First, I don't even accept 
that KR without formalism is really worth being called KR. Second, I 
don't think unformalized knowledge is actually much use in building 
ontologies. So-called "common sense' knowledge is supposed to be 
universally shared among competent adult human beings, but even if 
that is true, it has not been of much use in the task of formalizing 
it. And third, I don't think it is even meaningful to speak of 
representing knowledge independently of any formalism or notation for 
representing it. That is, unrepresented representation is an oxymoron.
...    (043)

>I dont have any feelings towards OWL, other than I am looking for
>someone who can teach me ohw to use it and I have not succeeded yet.    (044)

I'm sure it would not be hard to get you to the level of an OWL user 
reasonably quickly. If you like I can try to do it by email (though 
off-list, I think :-) You might find the experience frustrating, but 
join the club.    (045)

>can learn chinese over here, but OWL
>cant. I have also tried to use the protege tutorial, and havent 
>gotten anywhere
>also asked around 'can you teach me owl' no luck yet. I can go to the
>protege training at stanford this spring but do you know what that
>means? Long distance travel, plus tuition fees, plus a new passport
>and the risk of being sent back on entry cause I am threat to national
>But I can understand any/most knowledge in plain language, so maybe
>owl should not be a reuquirement for knowledge representation on the
>web. I guess thats the point of that requirement.    (046)

If by KR you mean represented in a form humans can read, then the Web 
does that already. The idea of the SWeb (and of ontologies more 
generally) is to provide KR which *machines* can utilize without 
human intervention. For that, it must be formalized, and it must use 
publicly agreed formal conventions.    (047)

>I am advocating 'clear abstraction'.
>I can understant if abstraction can be alienating for some, and that something
>abstract does not mean anything to you, and that you cannot visualise
>'abstract knowledge'.    (048)

I would certainly welcome some exposition of what you mean by this.    (049)

>>BTW, if I may blow a different trumpet for a
>>second, the best candidate so far for a single
>>overarching KR notation is I think the IKL
>>language ..
>let me look at it, references? we can always work on it    (050)

You can find a slideshow with many further pointers in it at 
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2006_10_26    (051)

The most useful is probably the guide, obtainable at    (052)

http://www.ihmc.us:16080/users/phayes/IKL/GUIDE/guide.html    (053)

Pat    (054)

IHMC            (850)434 8903 or (650)494 3973   home
40 South Alcaniz St.    (850)202 4416   office
Pensacola                       (850)202 4440   fax
FL 32502                        (850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes    (055)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (056)

<Prev in Thread] Current Thread [Next in Thread>