Kumoi is a virtualization-based cloud platform we have been developing for four years. Kumoi was designed for agility by treating various types of resources, such as virtual machines and networks, as distributed objects and combining them together by scripting. This paper especially focuses on security which was achieved in Kumoi where the range of security features including authentication, authorization, accounting, auditing, and encryption were supported. All these features were enabled by natural extension to the original design concepts and described well in this paper from the software composition view.

View full abstract

Hybrid systems are dynamical systems with continuous changes of states and discrete changes of states and governing equations. Hybrid systems are appliable to diverse fields including physics. HydLa is a declarative language for describing hybrid systems using constraints, and we have developed a symbolic simulator of HydLa named Hyrose. Hyrose is our first publicly available implementation of HydLa. Features of Hyrose include the simulation of systems with parameters, search based on automatic case analysis and bounded model checking. In this paper, we first describe the architecture and features of Hyrose, as well as semantical issues that had to be clarified to implement Hyrose. Then, in order to demonstrate the power of symbolic execution, we show several examples of the analysis of systems whose behavior exhibits quantitative and qualitative changes depending on parameters.

View full abstract

A web browser application for iOS, Libing, proposes a solution to resolve the problems of complicated button-based user interface (UI). The authors implement a gesture-based UI in Libing to realize simple and plentiful operations. Libing also keeps available display area maximum. Users can memorize correspondence between gestures and operations easily because users can assign arbitrary gestures to any operations and/or URLs flexibly in Libing. Touch-input-devices as typified by smartphones and tablet PCs have become popular recently. The market of these application programs has grown continuously. The most application programs are operated by the button-based UI. We can understand it visually. And the button-based UI is popular because we do not have incorrect operations when buttons are sufficiently large. However if we want to assign a lot of functions to buttons, we have to prepare many small buttons and/or a large area on which we place many buttons. We might have to divide the area to several pages to place all buttons. At least the user experience will be compromised because users have a lot of incorrect operations, available display area is compromised and operations become to be complicated. This study gives a solution to these problems which compromise the user experience.

View full abstract

Recently there have been many learning support systems which enable learners to acquire difficult skills such as how to play tennis, golf, and musical instruments and how to sing. Visual examples of how to perform the skill, and the detection of learners' mistakes by the system, allow learners to enhance learning effectiveness. In particular, the presentation of examples is important to decrease the difficulty of learning and keep learners motivated. However, learners need to be able to use their skill without the system support, since there are many circumstances in which they will not be able to use it. Therefore, the goal of our study is to construct a learning system considering withdrawal from a system support. In this paper we analyze conventional learning support systems, and propose methods of withdrawal from the system support. Additionally, we apply our proposed learning method to piano performance, and evaluate the effectiveness of the proposed system.

View full abstract

In this paper, we propose some methods to transform fine-grained editing operation history into coarse-grained one. Fine-grained history is suitable for investigating the detail of editing operations or source code changes, however, not for grasping the outline of editing history. We describe filters of editing operation history, fusion of editing operations, and criteria to transform into coarse-grained history. These are expected as useful devices to reduce the number of frames in replaying editing operation history.

View full abstract

We have designed and implemented a generator for an XQuery processor using compiler technique. It generates the processor for a valid XML document in lex, yacc source programs from a DTD file and a query. Because it knows which nodes of a document tree are actually accessed before processing an XML document, it can avoid making nodes which are never accessed. We classify join operations into two types, comparison of equality and numerical order. We use a hash method for the former, and a binary search method for the latter. We demonstrated effectiveness of a compiler-type XQuery processor; it runs faster with less memory than an interpreter-type processor.

View full abstract

Currently, most of RDF stores adopt a materialization approach to the search with RDFS inference. While on-demand approach based on backward reasoning has several desirable properties such as small data size, short load time, and flexibility, it is not popular since its performance is rather low in particular on large RDF graphs. To address this problem, we propose an optimization method based on query transformation. As major RDFS inference rules are recursive, we could not apply naive unfolding technique. The main idea is to make a contracted graph, which is equivalent with respect to RDFS inference but compact, and exploit the result of search with inference on it. Preliminary experimental results on standard LUBM benchmark are encouraging. For some patterns, we confirmed two order of magnitude speedup.

View full abstract

Using existing source code as third-party code to build new software systems becomes very popular in these days. However, many existing code is keeping on updating during their life circle. Different versions of code, even out-dated, is reused by other software and spreading all over the world. This paper presents an empirical study on the reuse of out-dated third-party source code of several famous open source libraries. Given target source code, using repository mining techniques and file clone detection techniques, we identified the different versions of code in other user projects, and discovered the vulnerability information of the out-dated versions. We also investigated how user projects manage their code. The result shows that a large proportion of open source projects are reusing out-dated third-party code, and many of them are not well managed.

View full abstract