forterpbeta

Documentation

1. Purpose & scope

1. Purpose & scope

The FORTRAN 66 standard — USA Standard X3.9-1966 — sets out the form of a FORTRAN program and how it is to be interpreted (§1.1). Its goal is interchangeability: a program written to the standard should mean the same thing on any conforming processor. (A processor is the implementation — a compiler, or in our case an interpreter.)

What "conforming" means

A processor conforms to the standard provided it accepts and interprets at least the forms and relationships the standard describes (§1.1). The wording is deliberately one-sided: the standard fixes a floor, not a ceiling. A processor is free to accept additional forms the standard does not mention, as long as it does not change the meaning of a program that is written to the standard.

That single sentence is what makes the DEC FORTRAN-10 extensions legitimate. Apostrophe- delimited character constants, the ENCODE/DECODE statements, extra intrinsics, longer source lines — none of these are in X3.9-1966, but offering them does not stop forterp from correctly running a program that stays within the standard. They are additions, not contradictions.

What the standard does not fix

The standard is careful to leave several things to the processor (§1.2.1). It does not prescribe:

  • the mechanism by which a program is turned into something runnable;
  • how a program or its data is transcribed to or from a medium (cards, tape, a terminal);
  • the manual operations needed to set up and run a program;
  • what happens when the rules of interpretation fail — i.e. the behavior of a program that is not standard-conforming;
  • the behavior of a program that exceeds the capacity of a particular system;
  • the range or precision of numerical quantities.

That last point matters in practice: the standard says a REAL datum is "a processor approximation to the value of a real number" (§4.2.2) and never says how good an approximation. Two conforming processors can give different last digits and both be correct.

forterp's place

forterp is a processor for this language, written in Python. FORTRAN 66 (plus the DEC FORTRAN-10 extensions) is forterp's strict base dialect — the default is now FORTRAN 77, so you select F66 explicitly:

      PROGRAM HELLO
      WRITE(6,100)
100   FORMAT(13H HELLO, WORLD)
      STOP
      END

$ forterp --std f66 hello.for
HELLO, WORLD

The literal text is a Hollerith field, 13H followed by exactly 13 characters — FORTRAN 66 has no apostrophe-delimited strings (those are a FORTRAN-10 extension). The leading blank inside the field is carriage control for the line printer and is consumed, not printed; both Hollerith and carriage control are covered in Chapter 7.

forterp notes. Because the standard fixes only a floor, forterp splits the things it can vary along two independent axes, and exposes both:

  • the dialect (the language the front end accepts) — F77 by default, F66 for strict X3.9-1966, FORTRAN10 for the DEC superset;
  • the target (the value model — the "range and precision" the standard leaves open) — NATIVE by default (the host's 64-bit integers and floats), or PDP10 for the faithful 36-bit DEC-10 representation.

Selecting these is covered in the Python API guide and Command-line tools. The few places where forterp deliberately interprets a program differently from a strict reading of the standard — for faithfulness to how real FORTRAN-10 V5 behaved — are collected in Appendix C; each is also flagged in a forterp notes box where it arises.