Are there any public domain XBRL Ontologies derived from XBRL Taxonomies out in the wild?

[kidehen]

All,

After the initial effort to commence the process of producing OWL
ontologies for a number of key XBRL ontologies, we find ourselves at
an impasse. From my vantage point, the issue remains the contribution
of a base ontology for community management and development.

I am once again seeking clarity about the availability of any initial
ontology efforts so that the community effort can commence.

Regards,

Kingsley

[GKJ]

Thank you, Kingsley, for your persistent efforts.

Since the translation of any publically used XBRL taxonomy into OWL
ontology is not trivial, I would suggest you may call for any ontology
used or to be used with XBRL (of course including direct translation
of XBRL taxonomy). These ontologies can be used, e.g. for the
integration and conversation between XBRL data and database, or the
convergence and comparison among XBRL data of different taxonomies
(e.g. between the US GAAP and IFRS). By nature, ontology differs from
an XML Schema (on which XBRL taxonomy is based) in that it is a
knowledge representation, not a message format. This combined use of
ontology/semantic web with XBRL can enhance the adoption of XBRL and
make XBRL data more value-added.

I hope more ontology examples will come with this broader scope.

Regards,
Graham Rong

[kidehen]

Graham,

On Oct 10, 3:26 pm, GKJ <gkj12...@gmail.com> wrote:
> Thank you, Kingsley, for your persistent efforts.
>
> Since the translation of any publically used XBRL taxonomy into OWL
> ontology is not trivial, I would suggest you may call for any ontology
> used or to be used with XBRL (of course including direct translation
> of XBRL taxonomy).

Yes, of course :-)
Hopefully your comment above aids in the qualification of my callout
i.e. the ontology doesn't have to be derived specifically from an XBRL
taxonomy, it simple needs to be a viable ontology for the mapping part
of RDF instance Data Generation by RDF Middleware (RDFizers).

>These ontologies can be used, e.g. for the
> integration and conversation between XBRL data and database, or the
> convergence and comparison among XBRL data of different taxonomies
> (e.g. between the US GAAP and IFRS).

Of course, the Semantic Data Web is ultimately about meshing disparate
and heterogeneous data sources :-)

>By nature, ontology differs from
> an XML Schema (on which XBRL taxonomy is based) in that it is a
> knowledge representation, not a message format.

Preacing to choir :-)

>This combined use of
> ontology/semantic web with XBRL can enhance the adoption of XBRL and
> make XBRL data more value-added.
>

Naturally :-)

> I hope more ontology examples will come with this broader scope.
>

So do I.

Kingsley

[Rene van Egmond [rene@c3ismc.com]]

Dear Kingsley;

I have given this combination of XBRL and RDF/OWL quite some thought in the past (a first outline for a paper in Jan'06) but due to busy schedules haven't paid any further attention to it.


So my 2cents;

            In my opinion the translation from XBRL taxonomies into OWL ontologies is not the way forward.


This is possibly a little too string of a statement so let me qualify this a bit; it is based on the objectives I had for my efforts and these might be different from your intentions. I wanted to "tie XBRL(*) into an RDF/OWL based Semantic Web". Now it is important to note that in this sentence, with XBRL(*) I mean the language XBRL and not XBRL taxonomies. Here lies a very important - dare I say, crucial - difference. Now it could be that in your case you want to represent XBRL taxonomies as RDF/OWL ontologies in which case our objectives are different and this email bears no relevance to your email ;-) .


The language XBRL, as does any language, has two aspects;

1) the language semantics;
what are the language constructs and their constraints (taxonomy-linkbase; a set of relation objects of a specific type with specific rules, the concept-relation; a single connection between two or more concepts that has specific attributes and constraints, a taxonomy-schema: a dictionary of concept definitions, concept: a simple or complex information object with standard properties, the instance: a collection of facts,... etc, etc.) which are defined in the XBRL v2.1 language specification as well as a number satellite specifications such as the XBRL versioning spec, the XBRL dimensions spec, the XBRL formula spec...

2) the representation (and processing);
how do we persist (and enforce?) these language constructs. In the case of XBRL the community has chosen for XML, XSchema, XLink, XPointer, XPath, etc. and the corresponding parser technologies.


In this sense XBRL is more XSchema on steroids then an RDF/OWL ontology. The first being a generic markup language and the second being a representation of knowledge about a domain, or put a little different XBRL is more like 'the RDF/OWL language itself' then 'the ontologies represented with RDF/OWL'. And I think here lies the way forward to integrate XBRL into the RDF/OWL domain - if that is what one is aiming for.

What would emerge from this approach if carried forward is a persistence of the XBRL language semantics in terms of RDF/OWL. After an RDF/OWL ontology about the XBRL Language is created defining a standard way to represent the XBRL language - the language constructs and their constraints. One could then design a converter/writer that takes the XBRL language object model (which would underly most XBRL editors etc as the internal language representation) and in stead of writing out the current "XBRL angle brackets", write out "RDF/OWL-based angle brackets".

This would achieve a number of very interesting and helpful things;
1) It would create a formal model of the XBRL language (something that doesn't exist today) with its semantic constraints enforced through an OWL ontology!
2) Ground the language in a single and very flexible XML representation language (RDF/OWL) that supports many of the language semantics needed rather then many different XML technologies XSchema, XLink, XPointer etc (OK stretching it a bit here :-)
3) ...?


This would get us to persistence of XBRL Language semantics in RDF/OWL, and thus the representation of XBRL taxonomies (which themselves represent the domain about which the knowledge is written down).

So bring back the parallels:
          look at the XBRL language  as RDW/OWL language
          look at the XBRL taxonomy  as RDW/OWL ontology

          look at the XBRL processor as the interpreter of the domain knowledge represented with the XBRL Ontology(**)

The IT solutions would still need to have the XBRL processors to interpret the information represented as an XBRL ontology(**) (domain knowledge info defined in XBRL language constructs persisted as RDF/OWL). These processors are the various applications as they exist now they write/read their internal XBRL application object models for taxonomies and instances in an RDF/OWL-based representation of XBRL.

This still leaves open the handling of information in terms (UoD) of the XBRL Ontology (the values for the UoD concepts) but this too could be represented in RDF. We could possibly create an XBRL-Business Rules ontology (the RDF/OWL representation of the constraints on the values and relations between domain concepts to constrain the values for the domain concepts... representing the language semantics of the XBRL Formula specification in RDF/OWL.

I have never gotten to the stage of actually completing a prototype for this RDF/OWL-based XBRL language and never started on the XBRL Business Ontology so I may very well still miss some important points... but I personally haven't given up on this track yet. Maybe one day when I have cycles to spare I can pick this all up again :-D


So... a little different perspective on the XBRL-Ontologies topic here?
at least I hope to have provided some new food for thought


Best regards;

         --- René ---

---

          Raynier A. van Egmond

          C3i Strategic Mgt. Consulting  LLC.
          10104 SW 153rd Street
          Vashon (WA) 98070
          USA.

          phone (Office):    +1 (206) 567 5627
          email:                  re...@c3ismc.com
          alternative:         c3i...@yahoo.com

[kidehen]

Rene,

On Oct 11, 4:21 am, "Rene van Egmond [r...@c3ismc.com]"

<r...@c3ismc.com> wrote:
> Dear Kingsley;
>
> I have given this combination of XBRL and RDF/OWL quite some thought in
> the past (a first outline for a paper in Jan'06) but due to busy
> schedules haven't paid any further attention to it.
>
> So my 2cents;
>

> *In my opinion the translation from XBRL taxonomies into OWL

> ontologies is not the way forward.

I think I should have been a little clearer in my earlier
communications :-)

I am seeking the construction of an Ontology that enables the
production of RDF instance data from XBRL docs aligned to a given
Taxonomy. The XML Schema based Taxonomy provides enough insights into
the Entity, Entity Types, and Associations in the many Finance domains
covered by XBRL.

As a first step, we will be releasing an Ontology (along the lines
described above) contributed by the University of Laval, Quebec,
Canada.

>
> *

> This is possibly a little too string of a statement so let me qualify
> this a bit; it is based on the objectives I had for my efforts and these

> might be different from your intentions. I wanted to "_/*tie XBRL(*)
> into an RDF/OWL based Semantic Web*/_"*.* Now it is important to note

> that in this sentence, with XBRL(*) I mean the language XBRL and not
> XBRL taxonomies. Here lies a very important - dare I say, crucial -

> difference. Now it could be that in your case you _*want*_ to represent

> XBRL taxonomies as RDF/OWL ontologies in which case our objectives are
> different and this email bears no relevance to your email ;-) .
>

Our objectives are the same. I simply see the Taxonomies as source
material for domain entities (concepts) and associations across these
entities.

> The language XBRL, as does any language, has two aspects;
>

> 1) /*the language semantics*/;

> what are the language constructs and their constraints
> (taxonomy-linkbase; a set of relation objects of a specific type with
> specific rules, the concept-relation; a single connection between two or
> more concepts that has specific attributes and constraints, a
> taxonomy-schema: a dictionary of concept definitions, concept: a simple
> or complex information object with standard properties, the instance: a
> collection of facts,... etc, etc.) which are defined in the XBRL v2.1
> language specification as well as a number satellite specifications such
> as the XBRL versioning spec, the XBRL dimensions spec, the XBRL formula
> spec...
>

> 2) /*the representation (and processing);
> */how do we persist (/and enforce/?) these language constructs. In the

> case of XBRL the community has chosen for XML, XSchema, XLink, XPointer,
> XPath, etc. and the corresponding parser technologies.
>

> In this sense _*XBRL is more XSchema on steroids*_ then an RDF/OWL

> ontology. The first being a generic markup language and the second being
> a representation of knowledge about a domain, or put a little different

> XBRL is more like '_the RDF/OWL language itself_' then '_the ontologies
> represented _with RDF/OWL'. And I think here lies the way forward to

> integrate XBRL into the RDF/OWL domain - if that is what one is aiming for.
>
> What would emerge from this approach if carried forward is a persistence
> of the XBRL language semantics in terms of RDF/OWL. After an RDF/OWL
> ontology about the XBRL Language is created defining a standard way to
> represent the XBRL language - the language constructs and their
> constraints. One could then design a converter/writer that takes the
> XBRL language object model (which would underly most XBRL editors etc as
> the internal language representation) and in stead of writing out the
> current "XBRL angle brackets", write out "RDF/OWL-based angle brackets".
>
> This would achieve a number of very interesting and helpful things;
> 1) It would create a formal model of the XBRL language (something that
> doesn't exist today) with its semantic constraints enforced through an
> OWL ontology!

> 2) Ground the language in a _*single and very flexible XML
> representation language *_(RDF/OWL) that supports many of the language

> semantics needed rather then many different XML technologies XSchema,
> XLink, XPointer etc (OK stretching it a bit here :-)
> 3) ...?
>
> This would get us to persistence of XBRL Language semantics in RDF/OWL,
> and thus the representation of XBRL taxonomies (which themselves
> represent the domain about which the knowledge is written down).
>

Yes.

> So bring back the parallels:
> look at the XBRL language as RDW/OWL language
> look at the XBRL taxonomy as RDW/OWL ontology
>
> look at the XBRL processor as the interpreter of the domain
> knowledge represented with the XBRL Ontology(**)
>
> The IT solutions would still need to have the XBRL processors to
> interpret the information represented as an XBRL ontology(**) (domain
> knowledge info defined in XBRL language constructs persisted as
> RDF/OWL). These processors are the various applications as they exist
> now they write/read their internal XBRL application object models for
> taxonomies and instances in an RDF/OWL-based representation of XBRL.
>
> This still leaves open the handling of information in terms (UoD) of the
> XBRL Ontology (the values for the UoD concepts) but this too could be
> represented in RDF. We could possibly create an XBRL-Business Rules
> ontology (the RDF/OWL representation of the constraints on the values

> and _*relations between domain concepts*_ to constrain the values for

> the domain concepts... representing the language semantics of the XBRL
> Formula specification in RDF/OWL.

> _*
> *_I have never gotten to the stage of actually completing a prototype

> for this RDF/OWL-based XBRL language and never started on the XBRL
> Business Ontology so I may very well still miss some important points...
> but I personally haven't given up on this track yet. Maybe one day when
> I have cycles to spare I can pick this all up again :-D
>
> So... a little different perspective on the XBRL-Ontologies topic here?
> at least I hope to have provided some new food for thought

This is in line with our overall goals.

The initial ontology from Laval will at least enable us commence
production of RDF from XBRL content re. Financial Data available on
Stock Exchanges, for instance (e.g. SEC submissions).

Kingsley

>
> Best regards;
>
> --- René ---
>

> ------------------------------------------------------------------------

>
> Raynier A. van Egmond
>
> C3i Strategic Mgt. Consulting LLC.
> 10104 SW 153rd Street
> Vashon (WA) 98070
> USA.
>
> phone (Office): +1 (206) 567 5627

> email: r...@c3ismc.com

> rene.vcf
> 1KDownload