PL/I

From EDM2
Jump to: navigation, search
George Radin
1931-01-22 to 2013-05-21 - One of the principal developers of PL/I at IBM

A procedural programming language introduced in 1964 by IBM, development was started in the early 60's under the auspices of George Radin with the aim of making an updated version of FORTRAN that could better handle generic programming tasks, but in the end it was decided to implement a new language based on ALGOL that would also incorporate features from COBOL and FORTRAN that were considered useful.

PL/I is a rather large language that had some design goals that might seem quaint by now such as "you shouldn't have to learn all of PL/I to use it", but this actually a feature: The low level and systems language features of PL/I are separate parts of the language specification and you can use it as a generic application domain language without touching any of the low level constructs, while having the options to use them if need be.

Another quirk is that it does not have any reserved words/commands, this is something it inherited from ALGOL and may be a reaction to COBOL but at the time PL/I was being developed that language was dominant and the list of reserved word you had to memorise was quite staggering, over 300 and growing with each new implementation, modern versions of COBOL have reduced that somewhat, this is similar to modern Basic variants such as GFA Basic that can have over 1000 reserved words.

The language is sometimes incorrectly known as PL/1, while the language is indeed named Programming Language One, the one was always meant to be expressed as a Roman numerical rather than an Arabian one, as is the tradition with computer languages even today. The Arabic 1 is only used on vintage computer and communications systems like USENET where the lower case only system or display limitation means that there is no Roman "I" available.

There are some variants of the language that are by now completely forgotten such as the safe variant SPL/I, that was designed at the behest of the USA Department of Defence in the early 1970's and was intended to be the language used for all non-business type programming done for the DoD but it end they gave up due to the variability of implementations and the difficulty of writing compilers and the DoD went with Ada instead.

Another variant was PL/C developed at Cornell University, it was designed for beginning programming students and the basic idea was that the compiler front end was smart enough to try and work out what the student meant rather than give just an error when he tried to compile. Since the background correction was silent, unless the compiler had to give up, you could easily end up with working programs that did something completely different than you intended to write.

OS/2 implementations of PL/I

GUI and application generators with PL/I output

  • CA-Telon - CASE tool/application generator - Commercial - Discontinued
  • IBM PL/I Toolkit (Visual PL/I)

Translators that generate PL/I output

  • C to PL/I (c2pli)
  • FORTRAN to PL/I Translator (F4-PLI.PLI) - for FORTRAN IV

Workframes and or IDE's

  • IBM Workframe/2

OS/2 programmer's editors with PL/I support

  • IBM Enhanced Editor (PL/x keywords for REXX and PL/I)
  • IBM LPEX (Live Parsing Editor eXtended)

PL/I source code snippets, archives and collections

Small programs or routines that you can integrate into your own programs or study to learn from, but are not delivered in library form.

PC DOS implementations of PL/I

  • DR PL/1-86 - The old Digital Research PL/I, available for PC DOS and CP/M - Discontinued - Commercial - now freeware
  • Liant LPI-PL/I - commercial
  • RUN - interpreter for limited PL/1 subset

PC DOS Libraries and class libraries

  • DR PL/1-86 Libraries - The old Digital Research libraries for the above compiler, disassembled - Discontinued - Commercial - Now freeware.

Generic or cross platform source code

Publications

Books

  • Mary L. Douglas and Frank Bates: Programming Language One, Third Edition: With Structured Programming – Prentice-Hall 1975, ISBN 0-13-730473-0
  • Barnes: PL/I for Programmers - North-Holland 1979, ISBN 0-444-00284-7
  • Hughes: PL/I Structured Programming - 3rd Ed., Wiley 1986, ISBN 0-471-83746-6
Earlier versions of this book were for many years the standard introductory PL/I text used in USA colleges and universities, they were OK and the text is reasonably clear and examples reasonably varied but there were better introductory texts out there, the third edition with its structured programming additions is however a considerable improvement over the earlier versions.

External articles

Links

Tutorials and other learning material
Tips'n'Tricks
Mailing lists & forums
  • PL/I Developers - On Yahoo groups - very low volume list intended for people that are developing PL/I tools, not for general PL/I programming discussion.
USENET

Standards

A standard for PL/I was not introduced until 1976, until then companies such as Burroughs and AT&T that wanted to implement the language used the published IBM specifications.

  • PL/I Language Specifications. IBM C28-6571 - 1965
  • ECMA TC10 Working group preliminary report - The first preliminary standard - sponsored by the IBM Vienna Lab (IBM-WT).
  • PL/I Language Specifications. IBM GY33-6003 - Originally published in 1967 but received a number of updates in the 70's.
  • ECMA-50 First Edition - First international standard, later adopted by various national and international organisations, best known being ANSI X3.53-1976. Now withdrawn by ECMA but still valid in some jurisdictions.
  • ANSI INCITS 74-1987 (R1998) - A PL/I Subset for generic programming (Subset G)

PL/I history

A procedural programming language introduced in mid 1960's by IBM, originally development was started in 1964 under the auspices of George Radin with the aim of making an updated version of FORTRAN that could better handle generic programming tasks, but although very popular and hugely influential, FORTRAN really only shone at numerical computing and I/O. When the team hit difficulties in shaping FORTRAN to their needs they decided to start afresh and implemented an ALGOL 60 variant under the name "New Programming Language" or NPL.

IBM later brougt on people from other organistaions that used IBM hardware to serve on the PL/I design committee including M. Douglas McIlroy from Cornell University.