Re: Network databases

Much as we SQL-bigots hate to admit, there is still more data in IMS than in SQL databases today. If you have a system devoted to just one problem, say checking retail accounts, and that problem has a static hierarchical structure, then IMS is at least an order of magnitude faster than SQL. But If you have to share the same data over several systems, then the flexibility and integerity constraints in SQL outweigh the speed.

This is a law of ecology -- an animal that adapts to one environment strongly does so at the expense of not being able to adapt to new environments.

We had a network database language (NDL) standard in ANSI X3H2. It did not match to any commercial product and was only used as a guide for features. All the network products were highly proprietary and depended on particular hardware and index systems. They had no single underlying model or syntax.

Then IBM put SQL into public domain by bringing in the base document for SQL-86 from the Santa Teresa labs to ANSI. In those days, this was very brave since IP law was still brand new.

The SQL-86 was minimal and could be implemented with a little effort. Then FIPS-127 conformance testing started and you had to have SQL to do business with the US government. It helps to be the largest user of data processing on Earth. Received on Thu Jan 13 2005 - 17:51:16 CET