ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] quadruples talk

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Patrick Durusau <patrick@xxxxxxxxxxx>
Date: Sun, 16 Sep 2007 16:02:53 -0400
Message-id: <46ED8BED.9040705@xxxxxxxxxxx>
Peter,    (01)

Peter F Brown wrote:
> Patrick, couldn't agree more regarding the tooling. Where I was going
> with this was precisely the idea that a specific "context" can be
> represented as a "bundle" of scope elements but which, although
> individually each scope element can be reified (ie represented as a
> topic) cannot collectively be reified as "context" as such - at least
> not in an easily manageable way (cf Matthew West's earlier post
> regarding quadruples and reifying one of the arities...).
>
>   
Oh, ok, now I understand. I haven't been following closely due to other 
obligations.    (02)

Well, if you want to create a context, that is a set of scopes that 
together represent a "context," what difficulty do you see with 
representing the set of scopes with an association and then reifying 
that association so it can be used as a role player in an association?    (03)

Admittedly that is kinda convoluted and perhaps easier to simply 
directly consider the "context" as a subject that is represented by a 
topic.    (04)

Or does your use case presume that there are separate scopes that 
collectively represent a "context"? That is you have no control over the 
"context" representation that you are going to encounter and so can't do 
something cleaner like representing the "context" with a single topic in 
an association that represents the scope?    (05)

Hope you are at the start of a great week!    (06)

Patrick
> Peter
>
> -----Original Message-----
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Patrick
> Durusau
> Sent: 16 September 2007 17:21
> To: [ontolog-forum]
> Subject: Re: [ontolog-forum] quadruples talk
>
> Peter,
>
> Sorry, I missed your original post where you said that scope in topic 
> maps cannot be reified.
>
> Recall that in topic maps to be "reified" (as far as the data model) 
> means to be represented by a topic.
>
> So far as I know, all scopes are represented by topics and therefore 
> "reified" in the sense the term is used in topic maps.
>
> So, is your question how to determine when two or more associations 
> share the same scope?
>
> While it is true that the data model does not provide a mechanism by 
> which you can determine from a topic that represents a scope all the 
> associations in which it appears (your suggestion of querying the entire
>
> dataset), that is more of a question of how you want to represent scope 
> than anything else.
>
> You could quite legitimately have scope playing a role in the 
> association, in which case you could determine from that topic (which 
> represents a scope) all the associations where it plays that role. Not 
> the usual representation but certainly doable.
>
> As far as CL being "king of the hill" I would only offer the observation
>
> that if project requirements require its use, by all means use it. One 
> can shred paper with a chain saw but I would not recommend it. Choose 
> the tool that is appropriate to what you are trying to do, not on the 
> basis of it being a great tool.
>
> Hope you are having a great weekend!
>
> Patrick
>
> Peter F Brown wrote:
>   
>> John:
>> By "complete", I meant more in the sense that computationally it is
>> impossible to determine whether every possible "situation" *can* be
>> expressed by the language.
>>
>> I'm sure that you're right and that every Topic Map statement can be
>> expressed in CL. My concern is rather that not everything in Topic
>>     
> Maps
>   
>> can be reified: for example, scope. One can make a statement with a
>> determined scope .... but one cannot identify all the statements for
>> which a given scope is valid, short of running a query over a very
>>     
> large
>   
>> - potentially undeterminably sized - dataset.
>>
>> As an aside, if CL can perform all the functions you indicate, do we
>> need OMG's Ontology Definition Metamodel? Is CL king of the hill?
>>
>> Best regards,
>> Peter
>>
>> -----Original Message-----
>> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F.
>> Sowa
>> Sent: 08 September 2007 22:21
>> To: [ontolog-forum]
>> Subject: Re: [ontolog-forum] quadruples talk
>>
>> Peter,
>>
>> I was just giving a few obvious examples, and the most
>> obvious ones are from prepositions and verbs.  I normally
>> use a more flexible mapping of natural languages to logic.
>>
>>  > Mapping parts of speech to arities is only part of the issue...
>>
>> The issues with NLs are immensely complicated. I just wanted
>> to cite a few examples to show that NLs very commonly represent
>> complex relationships with many more than two participants.
>> When you add further complications, that just emphasizes
>> the points I was trying to make.
>>
>>  > RDF uses pure triples, and require many joins;
>>  > Topic Maps goes one better and uses triples with, additionally,
>>  > scope (0 to n times) assigned to the association arc and role(s)
>>  > assigned to the topic nodes connected by the arc; but still is
>>  > inadequate to express context - and although the arc type
>>  > (association type) can be reified, the scope cannot.
>>  >
>>  > Can CL really do this? It may be consistent in handling
>>  > expressions (and not assuming or prejudicing any specific number
>>  > of arities) but it can never claim to be complete.
>>
>> I have no idea what you mean by "complete".  If you mean
>> complete with respect to NLs, that is still a research issue,
>> since nobody has a clue about the limitations of NLs.
>>
>> But if you mean, can CL represent anything in Topic Maps,
>> the answer is yes.  If you have any example you don't know
>> how to map to CL, we'd be happy to show you how.  Just two
>> conditions:
>>
>>   1. Give the TM example.
>>
>>   2. Explain in English every aspect of the significant information
>>      that is represented by each feature of the TM.
>>
>>  From that we can give you a CL representation.
>>
>> John
>>
>>
>>  
>> _________________________________________________________________
>> 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
>>  
>>  
>> _________________________________________________________________
>> 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
>>  
>>
>>
>>   
>>     
>
>       (07)

-- 
Patrick Durusau
patrick@xxxxxxxxxxx
Chair, V1 - US TAG to JTC 1/SC 34
Acting Convener, JTC 1/SC 34/WG 3 (Topic Maps)
Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)
Co-Editor, OpenDocument Format (OASIS, ISO/IEC 26300)    (08)


_________________________________________________________________
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    (09)

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