ABSTRACT
How tall are men in Edinburgh? The interesting thing about that query is when one considers how one applies it, or how it is applied to a data base. In this particular example let us suppose that the data base represents the population of Scotland. I think the query as expressed is fairly accurate. I am obviously interested in the height of men in Edinburgh. The population of Edinburgh is somewhere round about 200,000. Is the implication of making such a query to a data base system that I want to access 200,000 records in order to calculate the average? I want only to look at some statistically significant number of those records. So when one talks about inventing in an extended kind of DML, some high-level query update language, statements such as average, calculate the average, you are not really expressing the problem.
The second query that I want to present is, “How many Swahili-speaking nubile blondes are employed as system programmers in Uttoxeter?” There are a lot of qualifications that I am obviously referring to in this particular case: personnel records which represent female persons of a particular age expressed by nubile, with a particular color hair or skin or something, by blondes with two other fairly particular attributes, one that they work as system programmers, and that they speak Swahili, which as you know is an African language which very few people in England know. The question is, what do you expect the DBMS to do in order to not only understand the particular query, but implement it. I feel that some of the assumptions that one makes when talking about pure relations or sets, that the language processor that obeys that query needs to know something about the population of Swahili speakers, of blondes, of nubile blondes, of systems programmers, and of people that live in Uttoxeter. There is an awful lot of difference, from the point of view of performance, which of those particular adjectives you are going to look at first, and it makes an awful lot of difference which particular relations you actually declare to a system as being maintained by the system. And that seems to be one of the fundamental assumptions of the relational approach.
Index Terms
- Comments on advantages of the data structure set model
Recommendations
The data structure set model
SIGFIDET '74: Proceedings of the 1974 ACM SIGFIDET (now SIGMOD) workshop on Data description, access and control: Data models: Data-structure-set versus relationalI think we have a discussion today; not a debate. Its real purpose is to increase our understanding of the basic concepts underlying data base structures and their models; to note the similarities and differences between them, and to determine their ...
Comments on advantages of the relational view
SIGFIDET '74: Proceedings of the 1974 ACM SIGFIDET (now SIGMOD) workshop on Data description, access and control: Data models: Data-structure-set versus relationalI will be drawing a lot of my material from the application programming paper by Chris Date and Ted Codd. It's an excellent paper and I suggest you read it. The major focus of my talk and most of the arguments I'll be presenting have to do with a ...
Comments on the computer science/computer center interface
SIGUCCS '72: Proceedings of the annual ACM SIGUCCS symposium on The administration and management of small-college computing centersI'd like to find out how many people in this group come from institutions of 4000 or more students. How many of you come from institutions of 2500 or less? How many come from institutions of 1000 or less students? How many of you have your own computer? ...
Comments