[protege-owl] RDF/RDFS/OWL XML-based validation

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

[protege-owl] RDF/RDFS/OWL XML-based validation

Fermín Galán Márquez
Hello,

Given that RDF/RDFS/OWL ontology files are XML files, I wonder if the W3C
(or any other) has provided/standarized DTD/XML Schemas (or any other
XML-based validation document), to valided RDF/RDFS/OWL files without using
"ontology layer" validation mechanisms.

Anybdoy knows such DTD/XML Schemas (or any on-progress initiative)?

Please, any information about this topic is really welcome.

I don't actually know if this question has been already dealt with in the
list in the past (although I've been subscribed for a while and I didn't see
nothing related :) or even if this is the right place to discuss it. In that
case, please, forgive me! :)

Best regards,

--------------------
Fermín Galán Márquez
CTTC - Centre Tecnològic de Telecomunicacions de Catalunya
Parc Mediterrani de la Tecnologia, Av. del Canal Olímpic s/n, 08860
Castelldefels, Spain
Room 1.02
Tel : +34 93 645 29 12
Fax : +34 93 645 29 01
Email address: [hidden email]

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
Reply | Threaded
Open this post in threaded view
|

Re: [protege-owl] RDF/RDFS/OWL XML-based validation

Olivier Dameron
On Mon, 2 Oct 2006 18:45:07 +0200, Fermín Galán Márquez
<[hidden email]> wrote:

> Given that RDF/RDFS/OWL ontology files are XML files, I wonder if the
> W3C (or any other) has provided/standarized DTD/XML Schemas (or any
> other XML-based validation document), to valided RDF/RDFS/OWL files
> without using "ontology layer" validation mechanisms.
>
> Anybdoy knows such DTD/XML Schemas (or any on-progress initiative)?

My friend google sure does:
http://www.google.fr/search?hl=en&q=dtd+validator&btnG=Google+Search
http://www.google.fr/search?hl=en&q=xsd+validator&btnG=Search
http://www.google.fr/search?hl=en&q=rdf+validator&btnG=Search

Some of these tools are online, other are applications that can be run
locally. Choosing one depends on what it is you want to do.

I hope this helps
Olivier
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
Reply | Threaded
Open this post in threaded view
|

Re: [protege-owl] RDF/RDFS/OWL XML-based validation

Fermín Galán Márquez
Dear Olivier,

>> Given that RDF/RDFS/OWL ontology files are XML files, I wonder if the
>> W3C (or any other) has provided/standarized DTD/XML Schemas (or any
>> other XML-based validation document), to valided RDF/RDFS/OWL files
>> without using "ontology layer" validation mechanisms.
>>
>> Anybdoy knows such DTD/XML Schemas (or any on-progress initiative)?
>
> My friend google sure does:
> http://www.google.fr/search?hl=en&q=dtd+validator&btnG=Google+Search
> http://www.google.fr/search?hl=en&q=xsd+validator&btnG=Search
> http://www.google.fr/search?hl=en&q=rdf+validator&btnG=Search
>
> Some of these tools are online, other are applications that can be run
locally. Choosing one depends on what it is
> you want to do.

I'm afraid I didn't express myself with enough clarity...

I'm not looking for DTD/XML Schema validation tools: I'm looking the DTD
document or the XML Schema document that would be needed to use with those
validators. For example, in the case of OWL, some kind of owl.dtd or
owl.xsd, to validate arbitrary OWL definitions.

> I hope this helps

Thank you anyway! :)

Best regards,

--------------------
Fermín Galán Márquez
CTTC - Centre Tecnològic de Telecomunicacions de Catalunya
Parc Mediterrani de la Tecnologia, Av. del Canal Olímpic s/n, 08860
Castelldefels, Spain
Room 1.02
Tel : +34 93 645 29 12
Fax : +34 93 645 29 01
Email address: [hidden email]

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
Reply | Threaded
Open this post in threaded view
|

Re: [protege-owl] RDF/RDFS/OWL XML-based validation

Olivier Dameron
Hello Fermin,

On Tue, 3 Oct 2006 10:53:40 +0200, Fermín Galán Márquez

> I'm not looking for DTD/XML Schema validation tools: I'm looking the
> DTD document or the XML Schema document that would be needed to use
> with those validators. For example, in the case of OWL, some kind of
> owl.dtd or owl.xsd, to validate arbitrary OWL definitions.

Sorry I got you wrong.

It turns out that some folks have taken upon themselves to provide such
as dtd:
http://www.google.fr/search?hl=en&q=owl+dtd&btnG=Google+Search

HOWEVER, this is not (my point of view) such a good idea, because:
1. OWL-* are RDF languages. Therefore, there are several constructs
that are perfectly valid for representing the same thing (see example
bellow). Writing a dtd or schema validator for checking these will be a
dreadful task for little added value (see point 3). If you have access
to the production of the owl code, you may coerce it so that only one
form of construct is used extensively across the file. Then you may use
a dtd or schema validator
2. because of the previous point, if you want to apply your validation
on any owl file, you will get many false positive, i.e. an owl file
that would fail your validation may still be perfectly valid.
3. as owl-* are rdf languages, using a rdf validator will tell you if
they are syntactically and lexically valid, regardless of the
combination of rdf constructs that are used. Writing a dtd for checking
these will eventually amount to wrtiting a rdf validator from scratch.
4. side note to be confirmed or corrected: it seems to me that
libraries such as jena can read owl and rewrite it in a consistent and
non-optimised syntax that could be used for writing stylesheets or
playing with dtd.
5. finally, if your document is valid rdf, then you know that it is
correct xml, and I don't see the point of running an additional dtd
validation just to learn... that it is correct.

For representing that eltA is related to eltB through a relation
someRelation in RDF, the following three forms are valid and equivalent:
1.
<rdf:Description rdf:about="eltA">
  <rdf:type rdf:resource="#SomeType"/>
  <someRelation rdf:resource="#eltB" />
</rdf:Description>

2.
<someType rdf:about="eltA">
  <someRelation rdf:resource="#eltB" />
</someType>

3.
<someType rdf:about="eltA" someRelation="#eltB"/>

Hope this helps
olivier
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
Reply | Threaded
Open this post in threaded view
|

Re: [protege-owl] RDF/RDFS/OWL XML-based validation

Fermín Galán Márquez
Dear Olivier,

After reading your arguments, I'm agree in that DTD/XML Schema aren't
actually a good idea, so I should give up :)

However, I'd like to do some comments, just to follow this interesting
conversation :)

> 1. OWL-* are RDF languages. Therefore, there are several constructs
> that are perfectly valid for representing the same thing (see example
> bellow). Writing a dtd or schema validator for checking these will be a
> dreadful task for little added value (see point 3). If you have access
> to the production of the owl code, you may coerce it so that only one
> form of construct is used extensively across the file. Then you may use
> a dtd or schema validator
> 2. because of the previous point, if you want to apply your validation
> on any owl file, you will get many false positive, i.e. an owl file
> that would fail your validation may still be perfectly valid.
> 3. as owl-* are rdf languages, using a rdf validator will tell you if
> they are syntactically and lexically valid, regardless of the
> combination of rdf constructs that are used. Writing a dtd for checking
> these will eventually amount to wrtiting a rdf validator from scratch.

Yes, that's the entire point: the OWL syntax is quite rich (I mean, as you
show with your samples, the same thing can be told using different syntaxis)
and syntax richness is against DTD/XML Schemas (which works well when the
syntax is quite strict). Of course, a DTD/XMLSchema adapted to a particular
kind of OWL could be defined, but this DTD/XMLSchema would break when
appliying to general OWLs files that are equivalent but written using
different constructs.

As you point, to use a RDF validator seems definitely the best idea.

> 4. side note to be confirmed or corrected: it seems to me that
> libraries such as jena can read owl and rewrite it in a consistent and
> non-optimised syntax that could be used for writing stylesheets or
> playing with dtd.

Yes, this is some kind of "OWL shaper". I mean, a process that could read
any input OWL and output a equivalent OWL, but with a strict syntax that
works fine with a given DTD/XMLSchema. This is an alternative to RDF,
although of course it is less efficient and smart.

You mentioned stylesheet (I guess you referst to XSLT, right?). As a matter
of fact, my real goal was not a DTD/XMLSchema for OWL, but to build
translators from OWL to other XML vocabularies. The idea is to go from
instances in the "ontological layer" to particular XML vocabularies (e.g,
XHTML) in the "application layer". For example, a translator able to
transform a OWL description of instances in the wine ontology (e.g,
particular instances of wines) to a description of thoses instances in XHTML
(e.g., a XTHML-formatted description of each wine and its propierties o
something like that).

Knowing that OWL and XHTML are both XML, my first thought was to use XSLT.
But, building a XSLT without DTD/XMLSchema is quite difficult, and, given
that building DTD/XMLSchemas for OWL is dreadful (as you has shown me :),
the XSLT approach for the translator doesn't seems feabible.

That brings to the following question: are there some kind of XSLT eqivalent
approach in the "OWL world"? I mean, a way of describing rules to transfor a
OWL file (I mean, instances, classes, etc.) in a different kind of document
(e.g, XML). If this XSLT-like system doesn't exists, it seems the only
solution is to build the translator from the scratch (using libraries such
as jena).

Best regards,

--------------------
Fermín Galán Márquez
CTTC - Centre Tecnològic de Telecomunicacions de Catalunya
Parc Mediterrani de la Tecnologia, Av. del Canal Olímpic s/n, 08860
Castelldefels, Spain
Room 1.02
Tel : +34 93 645 29 12
Fax : +34 93 645 29 01
Email address: [hidden email]

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
Reply | Threaded
Open this post in threaded view
|

Re: [protege-owl] RDF/RDFS/OWL XML-based validation

adam.saltiel
Transforms and tool kits:-
http://www.harmo-ten.info/index.php?option=content&task=view&id=5&Itemid=29

HarmoSuite

The Harmonise Suite is the software platform which implements the Harmonise solution to the data interoperability problem and allows tourist organisations to effectively exchange information. It is composed of the following three subsystems:

  1. Ontology Mangement
    The Ontology Management is the sub-system responsible for defining the IMHO. It exports the IMHO (as an RDFS file) which is loaded in the THNCC.
  2. THNCC (Tourism Harmonise Network Control Center)
    The THNCC is the central service for the THN (Tourism Harmonisation Network). It handles central Actor registration, allows download of the COSS and provides technical support and help for users (all using standard web interface).
  3. COSS (Co-Operation Subsystem)
    The COSS is the "distributable" part of the HarmoSuite which is installed and customized for each Actor. The COSS will act with other COSSs for information sharing.

The Harmonise project at SourceForge:


Model management:-

Phil Bernstein

http://research.microsoft.com/~philbe/
You:-

but this DTD/XMLSchema would break when
appliying to general OWLs files that are equivalent but written using
different constructs.


[Block] Matching for Ontologies
e.g.

XOBjects Group

http://66.249.93.104/search?q=cache:yjou1ZOXM74J:xobjects.seu.edu.cn/+xobjects&hl=en&lr=&client=firefox-a&strip=1
or
http://www.falcons.com.cn/
and

COMA++ | Abteilung Datenbanken Leipzig

http://dbs.uni-leipzig.de/Research/coma.html

You:-
The idea is to go from
instances in the "ontological layer" to particular XML vocabularies (e.g,
XHTML) in the "application layer". For example, a translator able to
transform a OWL description of instances in the wine ontology (e.g,
particular instances of wines) to a description of thoses instances in XHTML
I would have thought you are talking about the representation layer, not the application layer.
You would build your own, because only you know how you want to represent these things.
BUT there is, I think, a case for the judicious use of ontologies of the representation layer that would lead to a more generic approach to these problems, but my research is incomplete on this matter.
An ontology of the view that is separate from the way in which the view is rendered - this would only work with those technologies that allow themselves to be sufficiently seperated, but because this is work half done in every web app, web app technology seems a contender, e.g. AJAX + web app framework + injection of ontology specifically for this purpose + ability to consume XML (by transforming to an ontology?) for purpose of display.

You:-
As you point, to use a RDF validator seems definitely the best idea.

Why? Surely the process of producing OWL entails that of validating the RDF AND the ability to reason over the constructs. Isn't the ability to reason what is wanted? BTW, the point about OWL is that it is open, so chances are that a. it your app will have interchangable elements with external apps and b. you may find others who have done substantial work in your area, or have expertise, that you can use.
BUT, there is a proliferation of languages in this area which is a (slight) problem in that transformation techniques are needed, e.g. from OWL to WSMO.

You:-
I mean, a way of describing rules to transfor a
OWL file (I mean, instances, classes, etc.) in a different kind of document
(e.g, XML). If this XSLT-like system doesn't exists, it seems the only
solution is to build the translator from the scratch (using libraries such
as jena).

Which rules? The OWL is XML anyway, so you want to abstract the rules that determine the way OWL is formed, but to what end? And in what form would you represent those rules, as I understand it, they already exist in OWL-S. Model management, again. BTW there are implementations of transformation tools in XSLT, but I think XSLT is a bit fragile and is, anyway, still a step away from any persisted form. Once again, surely the aim would be to work with the model as object, in memory, and read/write to a database, so JENA, SESAME and attendant technologies, e.g. SAIL.
It seems that some of this functionality is also present in TopBraid Composer.

Adam Saltiel

On 05/10/06, Fermín Galán Márquez <[hidden email]> wrote:
Dear Olivier,

After reading your arguments, I'm agree in that DTD/XML Schema aren't
actually a good idea, so I should give up :)

However, I'd like to do some comments, just to follow this interesting
conversation :)

> 1. OWL-* are RDF languages. Therefore, there are several constructs
> that are perfectly valid for representing the same thing (see example
> bellow). Writing a dtd or schema validator for checking these will be a
> dreadful task for little added value (see point 3). If you have access
> to the production of the owl code, you may coerce it so that only one
> form of construct is used extensively across the file. Then you may use
> a dtd or schema validator
> 2. because of the previous point, if you want to apply your validation
> on any owl file, you will get many false positive, i.e. an owl file
> that would fail your validation may still be perfectly valid.
> 3. as owl-* are rdf languages, using a rdf validator will tell you if
> they are syntactically and lexically valid, regardless of the
> combination of rdf constructs that are used. Writing a dtd for checking
> these will eventually amount to wrtiting a rdf validator from scratch.

Yes, that's the entire point: the OWL syntax is quite rich (I mean, as you
show with your samples, the same thing can be told using different syntaxis)
and syntax richness is against DTD/XML Schemas (which works well when the
syntax is quite strict). Of course, a DTD/XMLSchema adapted to a particular
kind of OWL could be defined, but this DTD/XMLSchema would break when
appliying to general OWLs files that are equivalent but written using
different constructs.

As you point, to use a RDF validator seems definitely the best idea.

> 4. side note to be confirmed or corrected: it seems to me that
> libraries such as jena can read owl and rewrite it in a consistent and
> non-optimised syntax that could be used for writing stylesheets or
> playing with dtd.

Yes, this is some kind of "OWL shaper". I mean, a process that could read
any input OWL and output a equivalent OWL, but with a strict syntax that
works fine with a given DTD/XMLSchema. This is an alternative to RDF,
although of course it is less efficient and smart.

You mentioned stylesheet (I guess you referst to XSLT, right?). As a matter
of fact, my real goal was not a DTD/XMLSchema for OWL, but to build
translators from OWL to other XML vocabularies. The idea is to go from
instances in the "ontological layer" to particular XML vocabularies ( e.g,
XHTML) in the "application layer". For example, a translator able to
transform a OWL description of instances in the wine ontology (e.g,
particular instances of wines) to a description of thoses instances in XHTML
(e.g., a XTHML-formatted description of each wine and its propierties o
something like that).

Knowing that OWL and XHTML are both XML, my first thought was to use XSLT.
But, building a XSLT without DTD/XMLSchema is quite difficult, and, given
that building DTD/XMLSchemas for OWL is dreadful (as you has shown me :),
the XSLT approach for the translator doesn't seems feabible.

That brings to the following question: are there some kind of XSLT eqivalent
approach in the "OWL world"? I mean, a way of describing rules to transfor a
OWL file (I mean, instances, classes, etc.) in a different kind of document
(e.g, XML). If this XSLT-like system doesn't exists, it seems the only
solution is to build the translator from the scratch (using libraries such
as jena).

Best regards,

--------------------
Fermín Galán Márquez
CTTC - Centre Tecnològic de Telecomunicacions de Catalunya
Parc Mediterrani de la Tecnologia, Av. del Canal Olímpic s/n, 08860
Castelldefels, Spain
Room 1.02
Tel : +34 93 645 29 12
Fax : +34 93 645 29 01
Email address: [hidden email]

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03


_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03