Re: ANN: myOntology Prototype Available

Hi Azamat,

the basic answer to both your comments and your question is that of the 
six technical effects of  ontologies [1,2]:

1. using proven philosophical distinctions for making more stable 
conceptual choices (e.g. OntoClean),
2. having a global naming scheme that provides unique identifiers for 
conceptual elements,
3. excluding unwanted interpretations by means of text and other 
informal ways,
4. excluding unwanted interpretations by means of formal logic,
5. inferring implicit facts from the formal account, and
6. spotting logical inconsistencies in the ontology itself or in 
knowledge bases referring to the ontology by means of formal logic,

I am deeply convinced that the second (providing unique identifiers for 
entities) is the most important contribution - for many of the promises 
of using ontologies for interoperability, one does not need a whole lot 
more.

Practically, "lightweight" for us means that the meta-model is very 
limited and allows only to define classes, datatype properties, object 
properties, ontologically significant instances, and a weak taxonomic 
relation for properties and classes. With "define" we mean one can 
create a unique identifier and describe the intended meaning by labels, 
short text descriptions, and multimedia elements like flickr images.

I do not share your opinion that we must first reach consensus on the 
overall conceptual framework. In my world, it is perfectly fine if there 
are two different notions of e.g. what gravity is * as long as it is 
always clear to which notion I am referring to.*. In a nutshell, having 
a dereferencable URI for meanings alone is, IMO, sufficient for making a 
large part of the Semantic Web vision a reality. We avoid all problems 
caused by synonyms and lexical variants, and the ability to retrieve a 
description of the element greatly supports the community in 
establishing and renewing consensus about the associated meaning.

Best
Martin

[1] http://flickr.com/photos/dunken69/2569034940/
[2] http://www.heppnetz.de/files/hepp-ontologies-state-of-the%20art.pdf

Azamat wrote:
> On Tuesday, July 08, 2008 4:34 PM, Martin wrote:
>  
> >The basic idea is that all elements defined in myOntology are within 
> the myOntology namespace - in fact, we assume that the whole set of 
> elements is one huge single ontology.
>  
> Since we are close to create an open global knowledge base (covering 
> both Wikipedia and Encylopedia Britannica) supported by the unifying 
> ontology framework, I can readily share that to achieve this high aim, 
> one first need to do a devil work: to develop the general structure 
> (or framework) within which domain ontologies, taxonomies, models, 
> catalogues, vocabularies, terminologies, etc. can be given the 
> unifying knowledge organization. In the information sciences, this 
> simple truth is as old as the following reference, 
> www.mitteleuropafoundation.org/Papers/RP/*Ontology*%20for%20knowledge%20*organization*.pdf 
> <http://www.mitteleuropafoundation.org/Papers/RP/Ontology%20for%20knowledge%20organization.pdf> 
>
>  
> > All ontology modules (e.g. the cell phone ontology) are actually 
> just views (proper subsets). We hope to achieve this by a very 
> lightweight meta-model, which we hope minimizes conflicts between 
> ontology modules. So our approach is the direct opposite of attempts 
> to contextualize the meaning of elements or what is being done in Cyc.>
>  
> [Up to now the term 'lightweight' has been associated with something 
> having no importance or influence, like a ''lightweight intellect'']. 
> Now wonder if the authors can go for more details about this 
> intriguing '' very lightweight meta-model'', an addition (or 
> alternative?) to the CYC ontology.
>  
> Thanks,
> Azamat Abdoullaev
>
>     PS: expect this sudden discovery might be interested for the
>     Ontolog-Forum as well.
>

-- 

-----------------------------------
martin hepp, http://www.heppnetz.de
mhepp@computer.org, skype mfhepp