• User guide
  • Our services
  • Tokyo main library
  • Kansai-kan of the ndl
  • International library of children's literature
  • Access
  • Photoduplication service
  • User registration
  • Online services
  • List of online services
  • Legislative information
  • Online catalog
  • Electronic library
  • Search guide
  • Online gallery

National Diet Library Newsletter

No. 131, June 2003
Back
Next

Thinking back over the thirty year history of
the NDL host computer
(part two of two)

by Tokiyuki Sugiyama
Senior Secretary, Administrative Department

This is an abridged translation of the article of  the same title in
the NDL Monthly Bulletin No. 502 (Jan. 2003).

Contents
( part 1, No. 130, last issue)
Introduction
1. Childhood - the early days of Japanese processing
2. Adolescence - the age of catalog and index compilation
3. The prime of life - the age of JAPAN/MARC

( part 2, No. 131, this issue)
4. Mature age- the age of online service
5. Old age- handing over the task to the new system
Conclusion

4. Mature age- the age of online service
From the late 1980s to the early 1990s, we began to feel strongly that all the workflow should be systematized.  We called the idea "the total system" and created a committee in 1987 to consider a systematizing process of books in Japanese and periodicals from the comprehensive perspective.  However, we had to face challenges lying between the departments. 

In case of the system for periodicals, we mapped out a plan which covered Western and Japanese serials and newspapers, but we faltered even at the accession process.  It was the worst-case scenario that we could not launch the development even after seven years of consideration.  We finally decided to divide the functions and begin with Western serials.  Hitachi and the NDL staff developed an online system and a batch system respectively and managed to operate the whole system in the eighth year.  After we got over that hump, we applied the system smoothly to Japanese serials and newspapers.  We also succeeded in developing other systems such as the bibliographic information creation system and the binding management system. So we somehow managed to build a kind of "total" system. 

In case of the system for Japanese books, we planned to link materials with online bibliographic data by putting an ID bar code on each accession in order to have overall control of accession records and location.  However, because of many difficulties in linking, we had to give up the idea.  Instead, a checking system that controlled Japanese books by quantity started in 1990.  As well as the system for periodicals, we spent a long time on preparation and produced an advanced system, though it was hardly "total".  In July 1997, ACAT, an online data retrieval and input system, was started.  It controlled and input bibliographic data which links with authority data.  A number of automatic functions were added after that, giving ACAT over one million steps and greatly contributing to the improvement of the Japanese books data quality. 


5. Old age- handing over the task to the new system
The Internet wave hit the library in the middle of the '90s and the NDL LAN was laid in 1995.  LAN was literally "LAN" (means irruption in Japanese) because it opened the gates to competing hardware in the library, where only HITAC had had a monopoly.

Machine organization had been expanded every few years since the start of HITAC 8400 in 1971, and reached the maximum scale in 1997.  The CPU was MP 5600/210 (2 processors).  Its main memory was 1,024MB, a 16,000-fold increase in capacity over the original computer.  The external disk storage capacity was 186.6 GB, 6,500-fold the original one.  Something like 1,000 PC terminals also appeared. 

After clearing the Y2K problem, the host computer reached its final years.  Its last task is to hand over the accumulated data to the new system and act as back-up until it is running steadily.  The data accumulated over 30 years consists of 19 million items of about 40 kinds.  Although not everything has been transferred, all the staff involved will be filled with deep emotion when they see the data from the old host computer living on in the new system.  The host computer is going to be turned off in December 2003.  That will be the end of its thirty-three year history.  

Conclusion
In retrospect, the systems of the host computer were basically handmade by the NDL people who developed the programs.  I cannot deny that it turned out to be disorganized as a whole because it had a lot of different systems without sufficient linking functions.  That means we did not have a clear vision of the overall picture when planning the long-term developments. But actually, it was impossible to anticipate such sweeping changes in computer technology and the network environment. 

The new infrastructure system is an integrated one that will be built in short term with a stupendous budget and input from many people after time has been taken to map out the full picture.  I believe a stable backbone is most important for the new system.  We can take time for side issues later.  I would like the new system to fulfill the dream of the old host computer. 

The NDL host computer (HITAC) was very reliable.  Though it sometimes brought a minor system down, I think its thirty-year service deserves more than full marks.  I hope it will come to a successful conclusion in its last year.  A great number of staff have been involved with the host computer.  I would like to speak for them and close with " Thanks a lot.  You did a really nice job."

HITAC 8400
HITAC 8400

 
Back
Next