Talk:IBM SQL/DS

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Additions to the article[edit]

I am Emil T. Cipolla; who was the lead planner at IBM Endicott NY in 1978 thru 1985, and prepared the marketing material for the first several releases of SQL / DS. I can provide much of the history of the development of the first several releases of SQL/DS:, and have most of the original IBM publicly available Announcement letters. The first version operated on the DOS/VS operating system; the second on the VM/CMS operating system. Subsequent releases were made to provide additional functions and improved performance. When IBM released the DB/2 relational DBMS for the MVS/VS operating system three years later, SQL/DS was later rebranded as DB/2 for DOS/VSE and CMS.


Brief history of the events preceding the first release of SQL/DS: Ed (Ted) Codd at IBM San Jose Calif developed the relational model - System R, (as a prototype). One of the original test sites was Pratt & Whitney Aircraft. The language was Structured Query Language (aka SEQUEL) ; key developers included Chris Date and Don Chamberlain (and others -) who can provide more of the background of System R and SEQUEL.

At the time, the dominate data base was IBM's DL/1 for the DOS environment,and IMS DB, for the MVS environment. IMS was developed by IBM Santa Theresa; Both DL/1 and IMS/DB were hierarchical data models, and the program had to know the detailed structure and relationship among the various records and sub records and fields. One 'navigated' thru the data base structure by providing specific detailed search parameters, and testing when sub -record segments were crossed. The significant 'attractiveness ' of the relational model is that one simply could retrieve the desired data fields among various records (two dimensional tables) simply by asking for them, and specifying the relationship among them. The Relational Database program knew where these segments were stored, and provided them to the application. Programming under the relational model was an order of magnitude easier than the hierarchical model.

This article will not attempt to describe the Relational model (see other articles for Codd's rules for first, second,and third degree of normalization and the relational model).

We developers of SQL/DS summarized Codd's rules as : 'The key, the whole key, and nothing but the key, so help me Codd '

IBM Santa Theresa had little incentive to introduce a competitive data model to the high end main frame IMS DB/DC MVS market segment. IBM Endicott, under the Lab Development managers Jack Cook, Jerry Becker, Larry Whitney and the outstanding programming development team, was anxious to provide a simpler and easier to use data model for the mid-range customer set. They 'ownership' of the System R prototype code was transferred to Endicott in late 1978, which modified it to execute on the IBM mid-range operating systems and hardware.

The initial vision of SQL / DS was to be an on-line ad hoc data management query tool for simple small data sets - performance was not the primary design goal. However, the customer reaction was so positive, that capabilities were added to make SQL / DS a robust data base model.

My previous assignment in IBM was as a system engineer whose customer set was mid range operating systems. My role was to use my customer experience to develop a business case, forecast, price point, and the marketing program material for target customer set (the mid range DOS/VS and VM/CMS operating systems.

Shall I continue with this ?? Is the history of interest ?

--68.14.144.145 (talk) 01:54, 18 October 2019 (UTC)[reply]

— Preceding unsigned comment added by 68.14.144.145 (talkcontribs)