We implemented Topic Maps besed web applications which use the Topic Maps web service. We are publishing the applications on the internet. By using the service, according to identified subjects the web applications can get richer information from other topic map web applications easily and effectively. In this paper, we describe usable components for the service. We report the case examples of topic map web applications and Topic Maps web services which we implemented. They use the PSIs to identify subjects among applications. They also use TMRAP (Topic Maps Remote Access Protocol) which is a Web API to exchange Topic Maps fragment. We also consider the general structure of Topic Maps web service.

1. TMRA 2009
Topic Maps web service:
Case examples and general structure
2009/11/13, Leipzig, Germany
Motomu_Naito (motom@green.ocn.ne.jp)
Knowledge Synergy Inc.
Lars Marius Garshol (larsga@bouvet.no)
Bouvet ASA

2. Table of Contents
1. Introduction
2. Existing components for the Topic Maps web service
3. Case examples of Topic Maps web service
4. General structure
5. Conclusion and Future work

3. 1. Introduction
Background
- Many topic maps (tms) and web applications based on
them already exist, and more and more tms and their
applications are expected to appear
- The same topics (subjects) sometimes appear in
different tms in different applications and organizations
- Topic characteristics (i.e. topic name, occurrence
and role) are different in each tm
- We think tms become richer and more useful if they can
exchange and share the characteristics
Purpose
- To report the case examples of Topic Maps web service
- To consider the general structure of the service

4. 2. Existing components
1. ONTOPEDIA’s PSI Server
(source: http://psi.ontopedia.net/ )
2. subj3ct
Subj3ct is a registry and clearinghouse service for subject
identifiers for the Semantic Web.
(source: https://subj3ct.com/about )
3. TMRAP : Topic Maps Remote Access Protocol
4. Topic Maps web application
- Lars Marius’s photo (tmphoto)
- Topic Maps case example (tmcase1)
- Topic Maps tools (tmtools)
- Everyday Physics on Web (EPW)
- Larsblog engine (larsblog)

5. 3. Case examples of TM web service
The get-illustration web service among
Client existing TM web applications
(tmtools)
TMRAP requests
TMRAP requests
Server/ TMRAP requests
Client Client Server
(tmcase1) (tmphoto) (Larsblog)
TMRAP requests
Client More information about
(EPW) the get-illustration web service
http://www.garshol.priv.no/blog/183.html

6. 3. Case examples of TM web service
- TM fragment exchange between TM applications using TMRAP
- Using the same PSI

7. The get-illustration web service
Using the same
PSI Server PSIs for persons
Tmphoto owns persons
tmcase1 requests photo
photos and provide their
to tmphoto and get url
url According to request
and displays it
Lars Marius’s photo
TM case example (tmcase1) (tmphoto)
Request
Return

8. ONTOPEDIA’s PSI Server
- Managed by ONTOPEDIA ( http://psi.ontopedia.net/ )
- Managing PSI & PSD (Published Subject Descriptor)
8

9. TMRAP
・ TMRAP (Topic Maps Remote Access Protocol) is a
web service interface
( http://www.ontopia.net/topicmaps/tmrap.html )
・ It makes possible to retrieve and modify Topic Maps
fragments from a remote Topic Maps server
・ Two protocols (HTTP or SOAP) are usable
・ Consist of the following methods
- get-topic
- get-topic-page
- get-tolog
- add-fragment
- delete-topic

10. Lars Marius’s photo TM web application
(tmphoto)
Created by Lars Marius Garshol
・ It manages 13,564 photos (at the time of 2009.11.5)
・ It’s ontology consists of topic types such as photo, person,
event, location and category and association type among them
・ User can navigate from the point of topic type view
・ It also has the following functions
- Filtering
- Rating the photos and showing the best photos
- Full text search
- Access control ( log in)
- The get-illustration web service
・ Enjoy it at http://www.garshol.priv.no/tmphoto/

11. Topic Maps case examples TM web application
(tmcase1)
Created by Motomu Naito
・ It manages 67 presentation (at the time of 2009.11.5)
・ It will be added more presentations very soon
・ It’s ontology consists of topic types such as presentation,
activity, event, session, person country, organization, purpose,
domain etc. and association types among them
・ User can navigate from the point of topic type view
・ It also has the following functions
- Client of Lars Marius photo
- Full text search
- Graphic display
- tolog query
・ Enjoy it at http://www.garshol.priv.no/tmcase1/

12. 4. General structure
Identifying subjects
・ It is necessary to attach PSIs to ontology topics and
instance topics
・ It is necessary to use commonly recognized PSIs
・ We suggest subj3ct.com as the place to seek PSIs for subject
・ It has 15,661,381 subjects at the time of 2009.11.5
・ If you can’t find PSI for your subject, create a new PSI and
register it on subj3ct.com

13. 4. General structure
Hub service
・ Hub server switches client’s TMRAP request to appropriate
TMRAP server
・ TMRAP servers can be registered with Hub server
・ Then clients can simply request
to the Hub server, instead of
requesting each server
・ The Hub would implement
the TMRAP get-topic-page
request to each server
・ The return from each server
is topic map
・ The Hub server merge
the return from server then
return the result to client

14. 4. General structure
Big picture: Identified subjects networking

15. 4. General structure
Big picture: Identified subjects networking
- The network consists of PSI server/clearinghouse, Hub server,
TM web applications, and something
- TM web applications play the role TMRAP client or/and server
- In the network we can link identified subjects and gather
information related to the subject
- Those subjects and information are still owned and managed by
their original applications
- In the network, we can enjoy rich, pure and high quality
information

16. 4. General structure
More information will be required for
Identified Subjects Networking
(Required information to use web service)
・Who and how select only appropriate information?
Human or computer?
・To enable to select only appropriate information,
the following information is needed
- Existence of other topic maps web applications
- topic characteristics
- context, domain, etc.
・Do we need Service Description language like a WSDL?
・We could use topic maps for TMWSDL

17. 5. Conclusion and Future work
Conclusion
・ We showed already existing components for TM web service
・ We explained examples of TM web services we implemented
・ We realized they are very useful
・ We considered general structure and required components
・ And we propose Identified Subjects Networking

18. 5. Conclusion and Future work
Conclusion
・ It is impossible to gather all information about one subject in one
topic map
・ It is inevitable to make topic maps distributed manner and
complement each other
・ To do so, each topic map and its web application can be simplified
and become easy to develop and maintain
・ TM web service is a very encouraging approach to realize the
Identified Subjects Networking
・ Today’s web search engines are very wasteful and inefficient
because they have to spider the entire internet continuously
to update their indexes
・ In contrast, to make it possible to link only suitable web
applications is very sensible and effective but a big challenge

19. 5. Conclusion and Future work
Future work
- To make it possible to link only suitable applications,
according to contexts, situations, etc.
- To make it possible to filter in only appropriate information and
filter out irrelevant information
- We will continue to work out web applications and required
components to realize generalized TM web service
i.e. the Identified Subjects Networking

20. ありがとう
ございました。
Takk!
20