In recent artificial intelligence research, one of the fundamental problems is to handle uncertainty in symbolic data. To handle such uncertainty, one promising way is to construct a probabilistic model that represents the data well, and then to make probabilistic inferences based on the model. Furthermore, as the data get more complicated than traditional data matrices, like sequence data or relational databases, the importance of the expressivity of the modeling language has been increased. From such a background, there have been proposed plenty of formalisms that attempt to integrate first-order logic and probability. In this paper, we present a probabilistic logic programming system called PRISM, which is expected as an efficient tool for probabilistic modeling which features declarative semantics, high expressivity originated from logic programs, and built-in fast routines for various probabilistic inferences. We describe PRISM as an implemented tool with program examples, its functionalities, and the result of a benchmark evaluation of computing performance.

View full abstract

Mobile IPv6 and NEMO BS are IETF standard mobility protocols for IPv6. It is said that freely available implementations play a big role in the deployment of new protocols. To accelerate the deployment of IPv6 mobility, we implemented the different mobility protocol stacks for the BSD operating systems. During the development, we made two different implementations based on this different design policies. The first one was an in-kernel implementation and the other a user space implementation. The former design makes it easy to use kernel information necessary for mobility operation, but it is difficult to implement and to extend features than the latter. The latter design needs to have extra mechanisms to retrieve or inject kernel information from user space, but in most cases developing user space programs is easier than developing in the kernel. In this paper, we discuss the design policies and implementation details of these two stacks.

View full abstract

To build a Japanese TEX system on UNIX, we need not only teTEX which is popular in Europe and the United States, but also many files which are developed in Japan and released one by one. It is annoying to handle these files because their installation process is not standardized. So we develop a Japanese TEX distribution ‘ptetex3’: we gather such files, modify them to work together, fold them in a file, and release it as an open source software. Recently more and more people begin to use ‘ptetex3’ and some Japanese Linux distributions have adopted as their TEX system. In this paper, we present our mission statement, report our activities, and examine why ‘ptetex3’ comes to be general.

View full abstract

Lego MindStorms is a robot development kit which makes it possible to control one's own robot by attaching various sensors and motors to a programmable block, called RCX, with an 8-bit CPU. In this paper, we present a Lisp system XS which runs on an RCX block to control robots. Unlike previous Lisp/Scheme implementations for the MindStorms, the evaluator of XS runs autonomously on the RCX, with its own runtime stack and garbage-collected heap. It communicates with the front-end subsystem on a PC, to provide an interactive programming environment with features such as backtrace, function trace, and terminal interrupt. The evaluator supports a programming language based on Scheme, extended with functionality for interfacing with attached devices such as motors and sensors. It also supports mechanisms such as event waiting and asynchronous event interrupts that are necessary for controlling robots.

View full abstract

We propose a web-based software named “picsense” which promotes exchange of digital art works such as illustrations or animations. We propose a unique model in which blogs on the Internet are viewed as galleries or “select shops”. The model uses the blogs as mediation points between art works and audience.
This model uses some “Blog”s as mediation point between art works and audience. This model solves those problems: 1) The needs of ordinal people to show art works are not under pressure. 2) There is the specific difficulty of art works to search.
The development of the software had the following characteristics: a) The requirement specification was analyzed in cooperation with artists. b) Neither difficulty nor novelty was requested in the policy for design choices to solve problems, but novelty and usefulness of the entire composition were pursued. c) Which made it possible for us to develop the software in a short term.
The software has expanded the casual opportunities for general users to see digital art works, and made it possible form them to discover digital art works according to their individual preferences. The software also gave new solutions for requirements specific to searching art works, including needs for fitting with environments.

View full abstract

This paper discusses an availability of statistical property of network topology for the server development and server selection algorithm in distributed systems. As a first step, we introduce new server development and server selection algorithms based on a degree of a server—the number of links for a server—and a scope—coarse grained topological information around a server. Our simulation reveals that the use of both information as a static QoS parameter improves the fairness of server load in a real Internet AS-level topology.

View full abstract

We present a Lisp system that can be downloaded to and used on mobile phones. This system was developed as a side-product of research experiment on real-time garbage collection. Many mobile terminals are now equipped with a tiny Java VM based on KVM so that the users can download application programs for their terminals. In order to realize high performance real-time garbage collection on such mobile terminals, we have been implementing various implementation techniques experimentally. We chose Lisp for evaluation, since there are many benchmark programs available in Lisp for memory management. We built a Lisp interpreter on KVM for mobile terminals by modifying a compact Lisp system JAKLD written in J2SE, and ran Lisp benchmarks for evaluation. JAKLD is designed to be used primarily as an embedded system in Java applications. JAKLD itself is a compact system, but we had to further reduce the size of the system in order to meet the limitation of functionality in KVM. The reduced system became available on mobile terminals by making use of APIs specially designed for mobile terminals.

View full abstract

TARZAN is an open software of Fast Handovers for Mobile IPv6(FMIPv6) which uses a link layer abstraction technique. A delay and packet loss which occur when a movement node changes a connected network have bad effect on communication. FMIPv6 is the most famous solution technique of this problem, and it becomes a foundation of many fast handover studies. However, the effectiveness of FMIPv6 and a study based on FMIPv6 in real environment are not confirmed, because there is not available implementation to the public of FMIPv6 free. Therefore we develop TARZAN, and we published it free. It is possible to run FMIPv6 fast by utilizing a change of link layer information. However, it is unclear how link layer information is used. In addition, it is very complicated to develop an extension on every device driver to obtain link layer information. Therefore we use a link layer abstraction technique and define how abstracted link layer information is used in FMIPv6 operation.

View full abstract

In recent years, the share of a web browser which has many functions — tab browsing, mouse gesture, etc. — to improve web surfine efficiency is getting increasing. By considerng about the reason, we can find that a essence of a web browser is changing from expression to function. In this thesis, we would like to introduce the web browser — Lunascape, which have been developed for five years, as a kind of next generation web browser, and discuss about the future of a web browser.

View full abstract

We aim to increase quality and efficiency of the testing for page generator programs, which are Web application view programs following MVC architecture. This paper proposes three kinds of test coverage, a generation method of test sequences to satisfy the coverage for each viewpoint and an automatic test results verification method. We define Page Model which represents observable behavior of pages and page transitions for Web application. We describe each pages in Page Model as page specification. We develop a testing support tool based on Page Model. Our tool generates test sequences which satisfy the coverage, test verification scripts and input data items list. After filling in the generated input data items list with test data, we can execute test sequences automatically with Canoo WebTest. The preliminary results for a medium scale Web application show that our method is helpful for test of page generator.

View full abstract

Characterizing modal logic in first-order predicate logic is a hot research topic in mathematical logic. Van Benthem provided an elegant characterization in which the standard translation of modal formulas coincides with the class of first-order predicate formulas invariant for bisimulations. Whereas he characterized modal logic in first-order predicate logic at the level of formulas, we characterize modal logic in first-order predicate logic at the level of proofs. Specifically, we provide a complete translation from a term calculus based on intuitionistic modal logic into Barendregt's λP. This characterization, identified as the equality of proofs, is considered significant because a term calculus based on intuitionistic modal logic is expected to realize staged computation.

View full abstract