OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

set message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: RE: [set] Re: Question about OASIS SET tools


Mustafa,
 
Great feedback. 
 
Just FYI for everyone working in this area - the new release of CAM upcoming will have enhanced tools for handling NIEM and NDR naming - including automated renaming and refactoring of dictionary components from schema.
 
Why is this important?   What I've discovered is that the typical enterprise data model contains a range of quality control over the years of its development.  So - we provide ability to handle renaming - including list of 200 common typos (from wikipedia research) - plus expansion of abbreviations and acronym detection.
 
Basically you cannot accurately compare things if the original dictionary has errors!  Plus by making the NDR terms uniform - now you have consistent comparisons happening based on intended use of data.
 
Also - we make the elements spellcheck accessible by splitting names into component parts.
 
This stuff is just too hard to check manually with 1,000s of elements in typical dictionary.
 
I'll let everyone know when the new release is available - plan is to have pre-release next week - and then finish regression testing for end of month.

Thanks, DW
 
 
-------- Original Message --------
Subject: [set] Re: Question about OASIS SET tools
From: Mustafa Yuksel <mustafa@srdc.metu.edu.tr>
Date: Thu, February 04, 2010 3:59 am
To: Jaewook Kim <jaewook@nist.gov>
Cc: "Dr. Asuman Dogac" <asuman@srdc.metu.edu.tr>, Erdem Alpay
<erdem@srdc.com.tr>, senan@srdc.com.tr, set@lists.oasis-open.org

Dear Jaewook,

I am copying this email to SET mail list as well. Please do so for your
further communication. Thanks.

Please find my responses below:

On 2/3/2010 9:40 PM, Jaewook Kim wrote:
> Dear Dr. Asuman,
>
> Since I contacted you a few month ago to ask about OASIS SET tools, I
> have continued my research for XML schema matching. And, I reviewed
> the OASIS SET tools and, for sample schemas to evaluate my approach, I
> used the mapping results (I could get it from
> http://www.srdc.com.tr/isurf/documents/Mappings.rar) the SET tools
> generated. I could get a few results about this.
>

Those mappings are not the final ones. We had to make some modifications
since the UBL Committee commented on the UBL 2.1 schemas and requested
some changes. The final mappings can be found in
http://www.srdc.com.tr/isurf/documents/MappingsNEW.rar.

> The question is
>
> - Can I consider the mapping results SET tools generated for 10
> messages as the correct mapping ? Is that fully automatically
> generated matching based on your approach ? Or human engineer somehow
> involved to adjust the matching result ?
>

These mappings are not direct output of the SET Tool. The XSLTs that are
generated by SET Tool are handcrafted in order to make the XSLTs stable
enough to be used in message translation in the pilot application of the
iSURF project.

> - And, I understand the matching based on ontology cannot fully cover
> the schemas' structural information. So, the OASIS SET tools proposes
> a few additional rules to analyze the XML schemas' information. But
> still it only provides a partial matching like matching between
> components of two schemas. Last time, one of your colleague I guess,
> Mustafa Yuksel mentioned the current approach presents the partially
> matched XPaths for the corresponding ontology classes to the user,
> then the matching can be done through lexical mechanisms. Do you have
> any additional investigation about this or any specific mechanism you
> are currently using ? Otherwise, I guess my approach might be helpful
> for this.
>
> My approach provides a progressive and iterative methods for XML
> Schema Matching, which can help user to easily review the matching
> results according to the iterative steps of matching algorithm and to
> input their feedback to improve the matching. Also, it provides an
> architecture to easily import and integrate the existing semantic
> information like ontology.
>

Well, SET Tools provide several mechanisms in order to enhance the
mapping process, such as rules to capture better mappings, syntactical
analysis of the element names, saving user preferences and using them
for new mappings, and so on. However, as you noticed, it is not always
possible to make 100% matching mappings.

We would be glad to have your contribution to the SET TC. Please send
any information related to your work to the SET TC mailing list.

Thank you for your interest and best regards,

Mustafa & Erdem

> We have a project related with the schema matching work, I hope I can
> find a way to collaborate with you and your project.
>
> Thank you.
> Sincerely yours,
>
> Jaewook Kim
>
> Guest Researcher
> Manufacturing Systems Integration Division
> National Institute of Standards and Technology
>
> E-mail: jaewook@nist.gov
> Phone: 1-301-975-8798
>
>

--
Mustafa Yuksel

Software Research and Development Center
Department of Computer Eng.
Middle East Technical University
06531 Ankara TURKEY

Phone: +90 (312) 2101763
Fax: +90 (312) 2101837


---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail. Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]