Specification

All the Information Access Environment Systems evaluated in the VisEx experiments using human subjects have the architecture shown in the figure.

The user (subject of the experiments) interacts with the Information Access Environment System (IAES) Core through a web browser. One of browser's tab is used as an editor, and the user uses this editor for compiling information gathered. Firefox is used for the browser. The editor will be described in the following section.

The IAES Core retrieves information needed by the user through interactions with the Information retrieval (IR) Engine. That is, the IR engine is a server that provides document retrieval facilities and the IAES core is its client program. We use the Apache-Solr as the IR engine. The interface protocol between client programs and Apache-Solr is based on the API provided by Apache-Solr in general.

The role of the IAES Core includes getting information requests from the user, translating those into queries to the IR engine, that is Apache-Solr, and showing obtained responses, such as retrieved documents, as a well-formed presentation. It rests upon the design of the IAES Core, in what style the user expresses her information requests, how she inspects and manipulates the responses. There features are the main part contrived by the participants of the VisEx task and to be evaluated in the experiments. The IAES Core is allowed to consult any knowledge base and external procedures in addition to interaction with the IR engine.

The Editor, which is a custom made software with the name of Quick Edit, is implemented as an add-in of the firefox browser. It can keep a log of user operations on the browser. Requests to and responses from the IR engine are also logged. The behaviors of the user and IAES Core are recorded using these logging facilities in the experiments. It is strongly expected to keep a log of the IAES Core with its own mechanism for more detailed analysis, as the log of the browser cannot understand the semantics of the operations of the IAES Core.

The experimental tasks which the users will carry out in the experiments are an event collection task and/or a trend summarization task. The former is conducted in Japanese and English, and uses the Mainichi newspaper articles and the Xinhua News Agency English service articles, respectively. The latter is conducted only in Japanese, and uses the Mainichi newspaper articles. The followings are the instruction to the users (subjects) and the list of themes of each experimental task. Each user is requested to conduct experimental task of a few of themes selected from this list.

Instruction to users for the event collection task

You are supposed to investigate and report the situation of the years from 1998 to 2001. Use the information search system on the browser for investigation on news paper articles and use the editor on the Quick Edit tab for compiling obtained information. The way to use the information search system is explained in another brochure. The all content of the report should be drawn from the information search system, while visiting other sites is also allowed for getting supplementary information such as clarifying the meanings of unknown terms.

The theme of the investigation is to list events that match with the following description and happened during 1998 and 2001. Your report should include not only some fundamental information such as date and place but also event specific features that characterize each event, for example organizations involved and the number of the toll for accidents. In addition, in order to make it clear where given information comes from, an Id of the article that reports it should be attached in parentheses to your description.

Draw a horizontal line such as ===, somewhere in the middle of your editing file, and compile the report for submission above it. For compilation, you can copy and paste the contents of news papers and edit those. Use freely the space below that line for taking memos. You don't need to clean up those memos for submission.

Around 30 minutes are given for one theme. You can finish at the time when you think enough report is made. When the end of the given time comes you should submit your report even if you are still in middle of compilation. If you can find only few events under a given condition, you are encouraged to make up those by similar or related events with some comments.

Themes of the event collection task

• Report incidents where NATO has recognized cases of friendly fire including specifics about dates and locations.
• Report incidents all over the world where dinosaur fossils have been excavated including when and where they were dug up, the type of fossil, and when the dinosaur was alive.
• Report the production of the Cannes Film Festival, when it is held, and the names of the winners.
• Report airplane crashes that have happened in Asia, and include information about where and when they occurred, the airline company name, and numbers of casualties.
• Report earthquakes that have happened in Asia, including information about when and where they occurred, the scale of the earthquake, and the amount and kinds of damage caused.
• Report incidents that have occurred at Japanese Nuclear Power Plants, and an outline of the events.
• Report the attack and invasion between Palestine and Israel, the preventative measures that have been taken to oppose the conflict, and steps that foreign countries have taken to control the situation.
• Report the companies that made international alliances in the cellular phone industry and the details of the alliances.
• Report foot-and-mouth disease outbreaks across the world, the date, time, and locations of their occurrence, and any responses to the issue.
• Report nuclear weapons tests that have been carried out by various countries.

Instruction to users for the trend summarization task

You are supposed to investigate and report the situation of the years from 1998 to 2001 in Japan. Use the information search system on the browser for investigation on news paper articles and use the editor on the Quick Edit tab for compiling obtained information. The way to use the information search system is explained in another brochure. The all content of the report should be drawn from the information search system, while visiting other sites is also allowed for getting supplementary information such as clarifying the meanings of unknown terms.

The theme of the investigation is to summarize changes of the following statistics and indexes during 1998 and 2001, that reflects some economical and/or social situation. Your report should include not only what kind of changes occurred and when, but also the reason those changes took place and the impact of those changes if such information available. In addition, in order to make it clear where given information comes from, an Id of the article that reports it should be attached in parentheses to your description.

Draw a horizontal line such as ===, somewhere in the middle of your editing file, and compile the report for submission above it. For compilation, you can copy and paste the contents of news papers and edit those. Use freely the space below that line for taking memos. You don't need to clean up those memos for submission.

Around 30 minutes are given for one theme. You can finish at the time when you think enough report is made. When the end of the given time comes you should submit your report in middle of compilation.

Themes of the trend summarization task

• Report the changes in oil price and pomp price of gasoline.
• Report the changes in the numbers of subscribers of fixed telephone, cellular phone and PHS.
• Report the changes in cabinet's approval/disapproval rating.
• Report the changes in values and volumes of national shipments of personal computers, and those maker share.
• Report the changes in shipping volumes of beer and low-malt beer, those ratio, and those maker share.
• Report the changes in employment situation indexes such as number and rate of unemployment.
• Report the changes in exchange rate of the yen against the dollar and the Euro.
• Report the changes in number of users of internet services using cellular phone, such as i-mode.
• Report the changes in population and its characteristics by gender and generations
• Report the changes in the stock market rate such as the Nikkei Stock Average.

For the specification of Apache-Solr, the following reference is useful.

• Solr 1.4 Enterprise Search Server David Smiley and Eric Pugh　Packet Publishing　

The document databases can be found at the following URLs. Those are planned to use in the experiments.

http://must.c.u-tokyo.ac.jp:8984/solr/admin

• The articles of Mainichi newspaper（an Yomiuri newspaper）from1998 to 2001

http://must.c.u-tokyo.ac.jp:8985/solr/admin

• The articles of The Xinhua News Agency English Service from 1998 to 2001

You can check the retrieval of these documents from these URLs. Use the browser that can show XML data properly such as firefox, as the results return in the XML format. The schema of the document data base can be shown from this URL by choosing [schema] in the Solr section. In addition to retrievals from these URL, the following command allows you to retrieve the documents.

  curl --user name:password http://must.c.u-tokyo.ac.jp:8985/solr/select?q=*:*

These URLs are protected by BASIC authentication. You should refer to the mail the organizers sent to you for the username and password.

Only Mainichi newspaper is used for the experiment for Japanese experimental task, though both Mainichi and Yomiuri newspapers are indexed now. The schema may be changed by adding new fields based on the results of discussion among the participants.

You can download the editor and its manual here. Redistribution is not allowed, and only the participants can use this software. The organizers are not responsible to any result caused by the usage of this software.

Please ask any question you may have to the organizers without hesitation. The organizers welcome any comment you would give too.