Search Mailing List Archives

Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-discussion] Help required on ontology project.

awatif Mubarak awalisco at
Tue Jan 13 02:47:10 PST 2009

Please, I need help on my project .

From: jonathan.carter at e-asolutions.comTo: protege-discussion at lists.stanford.eduDate: Tue, 13 Jan 2009 10:38:58 +0000Subject: Re: [protege-discussion] Help required on ontology project.Hi Rishi,

Good to hear that you're progressing. Feel free to send on your project.
Happy to help.



__________________________________________Jonathan Carter - Head of Technical ArchitectureEnterprise Architecture Solutions Ltd__________________________________________Assess your EA maturity

On 12 Jan 2009, at 23:07, saptha rishi wrote:
Thank You very much sir,
                           These were the details i wanted to know , now i have started proceeding in my project,as soon as i finish i will send you sir can you review it? once again thanks a lot sir.
On Mon, Jan 12, 2009 at 5:15 PM, Jonathan Carter <jonathan.carter at> wrote:

Hi Rishi,

__________________________________________Jonathan Carter - Head of Technical ArchitectureEnterprise Architecture Solutions Ltd__________________________________________Assess your EA maturity

On 10 Jan 2009, at 12:11, saptha rishi wrote:

Thank you Mr.Jonathan,
                   Very kind of you sir,i went through your mail and got a better idea of what to do and what not.I understood that we must run protege in server mode and java api in client mode and use it to interogate with protege.
                   I have some doubts sir can you please clarify or correct me ,
> Can i have values inserted at each class (sub-class ,individual etc) or should i use a database to store the value ,do you feel what i did was correct or please tell how to do it sir.
Yes, these are what are called Instances in Protege. You define your class hierarchy and then create instances of the classes, e.g. if you have a class "Person" then you would have instances for individuals of that class, e.g. an instance representing yourself as a person. 

>Sir let me explain briefly about my project
I want to develop a ontology based search engine for geospatial information.
My modules:-
1)User Interface,
2)Ontology Representation,
Now i have created my user interface as an jsp page using netbeans sir
and have also developed the ontology by using protege.
After this i am having lots of cofusion sir,
1)Since i am doing a search engine on geospatial information i want to store maps of countries and continents,where can i store it 
I would suggest that you store what you need to in the Protege ontology, using the instances. So, you'll have a class for country and have instances, e.g. UK, India, USA etc, and a class for continent with the instances for the continents in there. If you're not clear about how to create instances, have a look at the Protege tutorials. They cover this - and also have a look at the demo projects shipped with Protege as they all have classes and instances defined.
I'm not sure what sort of geospacial information you need to hold against the countries etc - that is what level of granularity you need in your ontology. I don't think I would recommend trying to hold the maps in the ontology. Rather, hold these in relational database - use the MySQL server that you're already using for your Protege database, but create a new logical DB in there to hold you maps etc. - and cross-reference these with a relevent identifier in the ontology and pull them out when you need to display them.

2)How to interrogate the ontology using my jsp page as the front end.
I think there are 2 options. The one to take depends on how frequently your ontology is updated. The solution I have used with a lot of success is to export the ontology as XML once the changes have been 'approved' and then interrogate the ontology using the XML export file.
I've written a plugin to send this XML to a web application that then simply performs XSLT on this XML to query the ontology and render the results. In many ways, this makes the querying of the ontology simpler because of the structure of the XML. 
In this approach, you can write simple JSPs that call an XSLT engine to transform the XML with a specific XSL file to do the query and results rendering.

The other option is to connect to the Protege server directly and interrogate the knowledge base via the Java API.
To do this, in your JSP you're going to need to make a connection to the Protege server (see the wiki) and this is proper Java code as a scriptlet, not JSP tags. You may want to consider writing some sort of servlet that does all this and manages the connection to the Protege server and loads up beans that your JSP can use to render the results.

3)Where to write the search code,
With the XSLT approach, the search code is in the XSL.
With the "live connection to Protege" approach, to some extent, I think that this depends on the nature of your searches. If they are simple you can code these as scriplets in your JSP, e.g.:

1. connect to protege
2. interrogate the knowledge base
3. render results
Step 2 will varying in complexity depending on the searches you need to make.

I'm afraid I haven't had time to review your ontology but hopefully these less specific comments will help you move forward.



protege-discussion mailing listprotege-discussion at lists.stanford.edu for unsubscribing: _______________________________________________protege-discussion mailing listprotege-discussion at lists.stanford.edu for unsubscribing: mailing listprotege-discussion at lists.stanford.edu for unsubscribing: 
Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the protege-discussion mailing list