Hiroki Konaka TRC Massively Parallel Software Laboratory

Profile

Affiliation:	Massively Parallel Software Laboratory, Massively Parallel Department, Tsukuba Research Center, RWCP
Date of birth:	July 16, 1964
Birthplace:	Hyogo
Background:
1987	Graduated from the Electronic Engineering Department, Faculty of Engineering, University of Tokyo.
1989	Completed the Master Course of Electric Engineering, Postgraduate School of Engineering, University of Tokyo.
1989	Joined Mitsubishi Electric Corporation.
1992	On loan to the Real World Computing Partnership
Hobbies:	Golf, playing musical instruments.

It is almost four years since I came here to join the Tsukuba Research Center. At first, I didn't know much about this city, except the section surrounded by the main streets that run north to south and west to east. But I've recently found many interesting spots and things in the community and its neighborhood such as very cheap, good quality agricultural products, a "three-star" Soba noodle restaurant, and an unbelievably cheap golf course. I have also found a lot of relaxing places within a few hours drive, such as Kasama and Mashiko, both famous for pottery, hot spa resorts and waterfalls.

It may also be relevant to my research activities here at the TRC. It was in the spring of 1993 when a research subject started to take shape in my mind on the theme of a programming language within a parallel object-oriented framework as the infrastructure of research on massively parallel programming models. When I was a university student, I developed a music information processing system and the type inference systems for an object-oriented logic programming language and a concurrent logic programming language. After joining Mitsubishi Electric Corporation, I programmed neural networks and other things; I was always a user of programming languages. When I started my new research here at the TRC, I had an image of the language I'd like to develop, but I didn't have even a basic knowledge of language processing systems. I had to start learning from scratch, so I could not afford to look around new language. The new language was soon named "OCore" for some reasons. OCore is pronounced as "O KOWA," and its meaning is "fearsome" in a dialect of the Western area of Japan. So, the development of OCore was launched with a sense of uncertainty while whispering "OCore".

Eventually, in December 1993, a small language processing system finally began running on a Paragon machine although it contained a number of bugs. At the same time, some parallel applications started to run. Around that time, I began to see an environment around the OCore language from machines to applications. Since then, many improvements and modifications have been made to the language with much support from many colleagues, including Mr. Maeda and Mr. Tomokiyo of the Massively Parallel Software Laboratory, and some application programs finally ran on several machines. The speech recognition program of the Information Integration Laboratory is a good example; it gave me an unexpected pleasure of achievement when such applications that I hadn't even conceived of at the start finally ran on actual machines.

I will continue to improve and modify OCore so that parallel applications that other conventional parallel programming languages cannot handle can be run on computers. In addition, I will work on description of parallel application programs that are closer to human processing such as speech recognition.