25.09.2012 WP10 - X-Search

Agenda

Time: Tue, September 25, 2012, 11:10 - 11:50 CEST

• Project review
• Coordination for 3rd TCom meeting
• Current status of integration
• XSearch use cases

Participants

• Alex Antoniadis (NKUA)
• Gerasimos Farantatos (NKUA)
• John Gerbesiotis (NKUA)
• Rena Tsantouli (NKUA)
• Yannis Marketakis (FORTH)
• Yannis Kitsos (FORTH)
• Anton Ellenbroek (FAO)
• Claudio Baldassarre (FAO)

Discussion Summary

• Regarding the plan for the project review
  • NKUA expressed the importance of including the XSearch portlet with the functionality achieved thus far in the material under demonstration, in order for the outcome of the integration activity to be shown, and the importance of not reducing the effort put in the coming month. FORTH agreed, and expressed the significance of also demonstrating the full set of features that the implemented web apps offer.
  • The topic can be further discussed during the alotted time at the TCom meeting.

• Coordination for 3rd TCom meeting
  • NKUA and FORTH will present their contribution to the integration activity in two separate presentations in the "Search and XSearch" time slot. NKUA will present a brief outline of the functionality implemented during the previous periods, followed by the functionality and tests implemented since the previous TCom meeting, mainly ResultSetConsumer performance results and ranking. FORTH's structure will be similar.
  • FORTH feels that since "Search and XSearch" and "Semantic Data Analysis" occupy contiguous slots, the available time can be managed as if they were one unified time slot. NKUA agrees with that.

• Current status of integration
  • NKUA has performed some changes in ASL which have increased the performance of ResultSetConsumer. First concrete results show that reading of 100 results through the full stack of gCube Search and ASL takes 250 milliseconds. More results are going to be gathered this week and presented at the TCom meeting.
  • NKUA is putting effort in providing ranked results for multi-collection queries. A significant part of the functionality is already implemented and will be presented at the TCom meeting.
  • FORTH currently has no new version of the XSearch portlet available. Some changes are pending waiting for the performance results of ResultSetConsumer. NKUA requested any new version containing changes other than these to be made available for deployment in the dev infrastructure as soon as it's ready.

• XSearch use cases
  • Use cases will show the motivation behind XSearch, as well as how and why it's useful for communitites.
  • FAO feels that it's important to have a good "story" for the review, but of equal importance is to show how XSearch will be used by the communities.
  • FORTH intends to present some good use cases. The "Semantic Data Analysis" time slot during the TCom meeting is a good opportunity for such discussions.
  • FAO feels that collaboration is important to agree on a made up story for the project review, that will however not compromise subsequent work. FAO suggests that once the portlet is integrated, we should see which features should be developed for the FAO use cases. This discussion can take place either before or during the TCom meeting. FORTH proposes that the discussion be held during the TCOM meeting, as it's better to have something concrete in hand to begin discussions, and also because the discussion can benefit from the involvement of IRD, as Julien Barde will also be present at the meeting.
  • On the categorization of main entities performed by XSearch, FAO feels that the fact that the XSearch webapp recognises entities that have already been selected for FAO is no real result. What probably can be a good thing to show, and an added-value, is the feature of mining the entities into some web page FORTH states that this is more or less provided by the infrastructure: FIGIS is linked with entities provided by FLOD, therefore this is provided by the webapp. FAO feels that mining the entities and making a connection is only part of the work. Results from the search engine can perhaps be enriched from the search engine with data coming from RDF relationships. Since there is available a search point at FIGIS + semantic web knowledge, these two can be connected and the result be seamless to user, transpartent. FAO suggests that now the integration of XSearch is being finalized, FORTH can put more effort into developing feature. FORTH feels that this scenario and will be a good starting point and should be discussed at the TCom, but having also in mind to be careful not to end up with an overly ambitious plan that cannot be realized.
  • FORTH argues that the development plan for the future of XSearch should be based on requirements, which are in turn based on use cases. FAO agrees with that, therefore the discussion on what can be done will follow the discussion on use cases. FAO argues that the purpose of XSearch is to be used by the communities, we should therefore think carefully about the plan. FAO's contribution will be to provide use cases of XSearch, along with IRD, and ways of it being useful for the communities.
  • FAO will produce cases so that FORTH can have them available in advance before the TCom, in order to be able to discuss the during the TCom.
  • FAO and FORTH agree that the TCom will provide a good opportunity for everyone involved to discuss about XSearch and the top-level ontology.On Wednesday morning, during the semantic cluster there can be further discussion on use cases (XSearch+ontology) and in the afternoon we could organize a parallel session.

Actions

• NKUA and FORTH will continue effort on integration activity
• FAO will provide FORTH with use cases before the 3rd TCom meeting
• Discussion on XSearch demonstration during the review to be held at 3rd TCom meeting
• Discussion on XSearch use cases to be held at 3rd TCom meeting