PRTV(ID:4554/prt001)


for Peterlee Relational Test Vehicle.

Earliest relational algebra language


Related languages
ISBL => PRTV   Commercialisation of

References:
  • Todd, S. J. P. "The Peterlee relational test vehicle - a system overview" view details Extract: VDL and PRTV
    This is the earliest known application of VDL to database specification. The actual work was carriedout in the sunmer of 1974 when A. Hansal was an undergraduate student. The report gives a fairly complete and faithful model of the "Peterlee Relational Test Vehicle" (PRTV) - otherwise known as ISBL. PRTV is a relational algebra based DBMS actually implemented by the IBM UK Scientific Centre in the early 1970s.
          in IBM Systems Journal, 15(4), 1976 view details
  • Verhofstad, J. "The PRTV optimizer: the current state", IBM Scientific Center Report: UKSC-83, May 1976, IBM UK Scientific Centre, Peterlee, England. view details
          in IBM Systems Journal, 15(4), 1976 view details
  • Sibley, E. H. review of Todd (PRTV) view details Abstract: This is a well written paper about an interesting working implementation of a relational database system (PRTV) at IBM's United Kingdom Scientific Centre in Peterlee, County Durham. As the abstract says:

    . . . PRTV is not a fully-fledged data-base system, but rather an evolving prototype which is expected to aid in solving some of the problems that have been encountered in using relational data bases . . .

    Thus the system does not have any backup and recovery; it has a relational algebra interface, and it does not attempt (at present) to provide a relational calculus type input language. However, it is a real, operational, test vehicle -- as the title implies.

    The article first discusses some concepts and facilities, including the separate use of selectors and domain names (i.e., the name used for describing or selecting an element may be different from that used internally for the domain), the method of creation (data definition), operations (selection, projection, union, intersection, difference, and join), and query operators (degree, selector, domain, cardinality) which also return relations so that all operations are homogeneous in form. The concept of join is given two possible results: quadratic join, which involves all the results of a coordinate product, and natural join; it is also possible to rename a selector prior to join in order to produce joins without association on the same "column name", e.g., for use in joining a relation to itself without matching on one or more columns. There is a "user extension" capability which appears very well developed. Other features of interest are in special use of read and write relations, multiple usage, and the possibility of special commands to defer binding (providing a type of virtual data file which changes automatically when the relations are updated). A short example is provided of a library application.

    The Implementation section discusses the two major sections (the top and bottom end) and the CIL (Common Intermediate Language) interface, which uses prefix Polish expressions to pass "bricks" and 'streams" to and from the top and bottom end. [ This reviewer hopes the new terms will soon be as dead as the Information Algebra' s "clump" ] . The description of the logical and storage structure is quite adequate for an article of this size, though the author misuses "optimise" (like most computer scientists talking about programs) when he really means "improvements".

    This is, in fact, an excellent article on an important experiment, and it should be read by all who do not already have a working knowledge of the PRTV.
          in ACM Computing Reviews 18(11) November 1977 view details