This directory contains sources for cupl, a UNIX implementation of CUPL, the Cornell University Programming Language. It also interprets CORC, the Cornell Computing Language, the immediate ancestor of CUPL.

CUPL was an early (1966) teaching language implemented as a batch compiler on the IBM/360 at Cornell University. It was descended from an earlier (1962) experimental language called CORC (CORnell Compiler), which was in turn derived loosely from Algol-58 (not, as previously asserted here, PL/1).

CUPL has only one scalar type, a long floating-point real corresponding to C double. It supports vector and matrix aggregates, and has operations specialized for matrix calculations.

CORC is a near-subset of CUPL. There are some minor differences in keywords. In CORC, the statement GO TO <block> is equivalent to CUPL’s GO TO <block> END.

The cupl implementation is a Retrocomputing Museum project. The Retrocomputing Museum http://www.catb.org/retro/ revives long-dead languages for the benefit of students and historians of programming-language design. CUPL was chosen as a representative of a significant class of mid-1960s teaching languages.

Files

Documentation files

README.adoc		-- this file
cupl.doc		-- transcription of the original CUPL documentation
corc.doc		-- transcription of the original CORC documentation
cupl.adoc		-- description and discussion of the language
cupl.spec		-- project history is in this RPM spec file

Source files

Makefile		-- project makefile
cupl.y			-- the YACC grammar
cupl.l			-- the LEX lexical-analyzer description
cupl.h			-- fundamental types and defines
tokdump.c		-- token-dumper code (used for debugging)
interpret.c		-- parse tree interpretation
execute.c		-- actual execution
monitor.c		-- runtime support
main.c			-- cupl's main sequence

CUPL samples

(These are all the non-pathological program examples from the CUPL manual)

test/cubic.cupl		-- iterative solution to X**3 * 3*X + 1 = 0
test/fancyquad.cupl	-- third, fanciest version of quadratic solver
test/poly11.cupl	-- sum for k = 1 to 11 of k!x**k, plus 1
test/power.cupl		-- compute the sum as k goes from 1 to n of k!/x**k
test/prime.cupl		-- test a number for primality
test/quadratic.cupl	-- solve multiple quadratic equations
test/random.cupl	-- print 10 pseudo-random numbers
test/rise.cupl		-- compare iterated +, doubling, squaring, and *
test/simplequad.cupl	-- simplest version of quadratic equation solver
test/squares.cupl	-- print a table of squares
test/sum.cupl		-- compute sum of presented data

CORC samples

(These are all the complete program examples from the CORC manual)

test/factorial.corc	-- compute the factorial function
test/powercorc.corc	-- compute the sum of a power series
test/quadcorc.corc	-- solve multiple quadratic equations
test/sumsquares.corc	-- compute sums of first n integers and squares
test/title.corc		-- illustate use of NOTE, TITLE, and WRITE

For each .cupl or .corc file, there is a corresponding test file used for regression-testing of the interpreter. The interpreter should pass all these tests.

Known problems and bugs:

  • The code chrestomathy in the CUPL manual didn’t include any matrix algebra examples. Because of this, the matrix algebra facilities are incomplete. ALLOCATE, INV and DET are not implemented. Neither are subscript/slice references or assignments. Nor are matrix WRITEs or READs. Nor are the implemented matrix facilities at all well-tested.

  • For true verismilitude, a program listing should be output before results. This is trickier to do in C than it sounds; the lexer doesn’t know about line boundaries.

  • GOTO blabel END always goes to the end of the current PERFORM block, even if the label is that of an enclosing block.

  • Because the cupl interpreter also interprets CORC, the CUPL keywords BEGIN, DEC, DECREASE, EQL, GEQ, GTR, INC, INCREASE, INT, LSS, NEQ, NOTE, REPEAT, TITLE, UNTIL, and $ are also reserved words in CORC.

Send comments, bug reports, and fixes to esr@thyrsus.com. Also see my home page at http://www.catb.org/~esr for updates and other related resources.