Re: [ontology-summit] OS-2012 Problem Space

[Ali SH <asaegyn+out@xxxxxxxxx>]

Dear Jack,

On Tue, Feb 7, 2012 at 2:45 PM, Jack Ring <jring7@xxxxxxxxx> wrote:

Why not start by showing us the similarities and differences among modeling, analysis, problem, codification and transformation. 

Then was heist 'patterns'?

I definitely do not follow re "heist patterns".

I'm not sure I understand. Are you asking how modeling, analysis, problems etc. are all similar and/or different? At a high level, one identifies a problem, analyses it, models it, perhaps codifies said model in a particular language, and if they need to communicate with others, then perhaps provide a transformation of said codification to another. Not necessarily a linear activity, but easy to conceptualize as such. 

Obviously, they are all connected to one another. 

My interpretation of these terms is that a problem pattern pertains to descriptions of commonly occurring issues with regards to various aspects of systems. I would suggest that the problem themes in Exhibit 3 in [1] could all be candidate problem patterns.

An analysis pattern might then refer to different ways of characterizing how the problem can be addressed in the systems context. It would refers to the methodologies, perspectives or other that are useful in analyzing the domain. Though it need not to, it could take as input problem patterns and produce analyses of the problems as they apply to different systems contexts.

A modeling pattern would refer to particular commitments for the domain description. These patterns might take as input analysis patterns, and provide as output suggestions / possible ways of modeling the analysis - whether in natural language, FOL, category theory, but a way of ontologizing aspects of the analysis.  For example, formalizing / articulating the notion of "component" as a particular set of axioms could be a possible modeling pattern. 

The distinction between modeling and codification patterns is a bit less clear to me, though perhaps Giancarlo can clarify his intent with these terms. Hazarding a guess, I suspect that a codification pattern take as input a modeling pattern, and produce as output a representation of said pattern in a specific coding scheme / language. Though I am not happy with this treatment. But say the modeling pattern suggested that a 3D design choice was useful for characterizing the notion of role. The codification pattern would take this 3D design choice and explicate how it would be represented in say, CommonLogic or OWL...

Lastly, transformation patterns would be commonly used techniques / translations / transformations for converting a model that is stored in one formalism to another. For example, http://www.ihmc.us/users/phayes/IKL/GUIDE/GUIDE.html#TranslatingInto these could all be transformation patterns for OWL-CL.

So let's take the notion of roles. MatthewWest presented a use case which I believe contained a problem pattern for the role of roles within a systems appilcation. This spawned a series of discussions regarding how to treat roles, each perspective contributing an analysis pattern of varying quality. Some further discussions elaborated on this analysis and went so far as to tie it to an upper ontology or ways of formally representing roles in an ontology - these would be modeling patterns. Finally, though we haven't quite gotten there, perhaps someone would suggest how to incorporate such a modeling pattern into SysML or formalize it as a set of Common Logic axioms -- these would be the codification patterns.

Finally, perhaps we want to demonstrate how the notion of Role as formalized and codified in SysML can be mapped to roles as they are used in another modeling paradigm. Such an output would be a transformation pattern.

Is this clear?

Best,

Ali

Modeling Patterns 

Analysis Patterns 

Problem Patterns 

Codification Patterns

Transformation Patterns 

I think the various discussions / email threads on roles, parts and components can all be effectively turned into a variety of modeling and/or analysis patterns. It would be nice to show how they factor in various specific systems problems - where might one treatment of role be more appropriate than another?

Such an approach helps me grok the contributions far more efficiently and helps somewhat alleviate the tremendous complexity of the summit topic.

Best,

Ali

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012  
Community Portal: http://ontolog.cim3.net/wiki/

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012
Community Portal: http://ontolog.cim3.net/wiki/

--


(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,

_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/   
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/  
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012  
Community Portal: http://ontolog.cim3.net/wiki/     (01)