Hi Jens,
First of all, thanks for your feedback.
On 03/21/2012 12:19 PM, Jens Hübel wrote:
This sounds interesting. I need to
think more about the use cases for this. Perhaps it also
would be interesting to expose certain features of Stanbol
through a CMIS interface (acting as a CMIS server). For
example the classification hierarchy could be (a virtual)
CMIS folder structure. This might sometimes be better then
copying the information into the repository.
As far as I understand, you mean implementing some part of the CMIS
specification here together with some extensions exposing Stanbol
features. If so, I think we won't move on in this direction in the
project, at least in the near future. I mean implementing CMIS
specs. Our approach is more like to retrieve content from a CMIS
compliant repository, semantically enhance it and provide a
knowledge repository for those enhancements for future semantic
operations on the content. In short, Stanbol is designed to work
"with" a content management system.
Another way of integration could be
to somehow plugin to a CMIS query/search interface to
provide better results by making use of the semantic
information in the background. Not sure if this is in scope
for your project and also not sure that I fully understood
your current work.
From my side this seems a more feasible approach. Contenthub [1][2]
component of Stanbol indexes the content together with its semantic
metadata and provides semantic search facilities on it. It may be
interesting to this community as I suppose most of the members of
CMIS committee have their own content management system.
Let me give some usage examples. CMS Adapter component of Stanbol
interacts with CMIS compliant content management systems and
transforms the content structure into an RDF. This is a simple
ontology which exposes the hierarchy and properties within the CMIS
structure. Afterwards, if a newcoming content is submitted to
Stanbol (through Contenthub) and includes some words which match the
already extracted ontology's elements, the knowledge about this new
content is enhanced with the properties of the ontological element.
A second use case can be the following: when a user enters keyword
to perform a keyword-based search, new keyword suggestions are
presented to the user based on that ontology.
From the above perspective, I will be happy to present the use of
CMS Adapter (and other components of Apache Stanbol) to CMIS
community in-line with the use cases I mention above. This might be
a good opportunity for providing a proof of concept for automatic
ontology generation from CMIS compliant CMSs.
Regardless from the ontology generation mechanism, you can find a
demonstration which shows CMS Adapter and Contenthub integration for
ehealth domain at [3]. Note that in that demo a JCR compliant
repository is used but it is valid for CMIS compliant repositories
also.
So I hope this is not complete
garbage. By the way some of your links in your blog are
broken, perhaps you can check them again.
Absolutely not. I added comments for correct links to the posts as
they cannot be modified.
You also should check the upcoming
CMIS 1.1 capabilities. Type Mutability and Secondary Types
will not give you full flexibility but may help you to avoid
the _metadata file you mention in your blog.
Thanks for posting this!
Jens
Best regards,
Suat
[1] http://incubator.apache.org/stanbol/docs/trunk/contenthub/
[2]
http://incubator.apache.org/stanbol/docs/trunk/contenthub/contenthub5min
[3] https://www.youtube.com/watch?v=l7n6aRFcn1U
Dear Suat,
Thank you for your post. We appreciate your CMS Adapter work.
To all CMIS TC members, especially those who are interested in
ontology or RDF:
The Apache Stanbol project provides semantic capabilities to
CMIS repositories. I encourage you to take a look at Suat's
blog on model mapping (see link below) and inject your
thoughts on the proposed approach.
Best regards,
David
On Tue, Mar 20, 2012 at 4:13 AM, Suat
Gonul <suatgonul@gmail.com>
wrote:
Dear CMIS
community,
I would like to inform you about the study related with the
CMIS
specification, that we did in the scope of Apache Stanbol
project [1].
As a brief introduction, Stanbol is a software aiming to
bring semantic
capabilities to existing content management systems. One of
the
subcomponents of Stanbol is CMS Adapter[2].
In the scope of CMS Adapter, in the first step, we tried to
transform
structure of a CMIS repository into an OWL model based
ontology. In this
method, all object types registered with the repository are
transformed
to classes. Folders and documents can be transformed into
either classes
or individuals. When content repository objects are
transformed into
individuals, they are created as instances of classes which
correspond
to their object types. Furthermore, relationships are
transformed into
object properties and corresponding assertion is set between
the source
and target individuals. The aim in extracting such kind of
an OWL model
based ontology, is to enable reasoning on it (which cannot
be done on
content repository itself in our opinion). I have written a
blog post
about this feature at [3]. We have also a paper regarding
this feature.
It can be downloaded from [4].
The second approach is a bidirectional mapping approach
between RDF data
and content repositories. It basically provides populating
content
repository with RDF data in one way and transforming content
repository
structure into RDF in the other way. The mapping process is
done
according to a handful of configurations parameters. For
instance one
indicates the resources to be mapped to the content
repository, one
indicates parents of selected resources, etc. There is also
another blog
post including details about this feature at [5].
I would be glad if you have a look at the blog post and
could share your
opinions about our approaches, further ideas to improve the
existing
design.
Best regards,
Suat
[1] http://incubator.apache.org/stanbol/
[2]
http://incubator.apache.org/stanbol/docs/trunk/cmsadapter/
[3]
http://blog.iks-project.eu/stanbol-cms-adaptor-allows-you-to-transform-jcrcmis-nodes-into-semantic-web-schemas/
[4]
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.188.7723&rep=rep1&type=pdf
[5]
http://blog.iks-project.eu/adding-knowledge-to-jcrcmis-content-repositories/
--
This publicly archived list offers a means to provide input
to the
OASIS Content Management Interoperability Services (CMIS)
TC.
In order to verify user consent to the Feedback License
terms and
to minimize spam in the list archive, subscription is
required
before posting.
Subscribe: cmis-comment-subscribe@lists.oasis-open.org
Unsubscribe: cmis-comment-unsubscribe@lists.oasis-open.org
List help: cmis-comment-help@lists.oasis-open.org
List archive:
http://lists.oasis-open.org/archives/cmis-comment/
Feedback License:
http://www.oasis-open.org/who/ipr/feedback_license.pdf
List Guidelines:
http://www.oasis-open.org/maillists/guidelines.php
Committee:
http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=cmis
Join OASIS: http://www.oasis-open.org/join/
|