5

I was recently trying to find the reasons that led to the use of curly braces for blocks (compound-statements) in the C language and other languages of the same family.

Thinks like square brackets for arrays or parentheses to change operator precedence seems to come from the mathematical notation used for these things so it makes sense to me. The semicolon makes sense too, it's meant to separate related sentences with independent clauses, and it does something like that in C.

But I couldn't find something similar for curly braces: they are used for sets in mathematics, and something somehow similar is their use in musical notation (Ensemble staves) where they are used to indicate that the music on all the staves is to be played simultaneously (not quite the same use but I thought it was closer than their use in mathematics for sets).

Also, I was wondering if C is actually the first language to introduce this block syntax or did it just popularize it?

Paul Hankin
  • 54,811
  • 11
  • 92
  • 118
evuez
  • 3,257
  • 4
  • 29
  • 44
  • 1
    Could be as simple as `( )` and `[ ]` were already taken. Why do we use `< >` for templates / generics ? – H H Jun 04 '16 at 07:57

1 Answers1

4

The first curly brace programming language was BCPL in 1966, although it allowed $( and $) to delimit blocks in addition to { and }.

BCPL was based on CPL which used the unusual (to modern programmers' eyes) § to open a block and the same symbol with a line through it to close the block. Martin Richards, the inventor of BCPL in "How BCPL evolved from CPL": says:

CPL used a section symbol (§) to be equivalent to Algol’s BEGIN and a section symbol overprinted with a slash to represent END. BCPL adopted $( and $) for these tokens and as with CPL such section brackets could be tagged, allowing a close section bracket to close multiple sections. Unfortunately this convention lead to rather obscure programming errors and so when curly brackets ({ and }) became available, they were only used as untagged section brackets and the use of $( and $) was discouraged.

Curly braces came to C (1972) via B (1969).

Paul Hankin
  • 54,811
  • 11
  • 92
  • 118
  • 1
    I think you are guessing right. The availability of special characters in character sets and on keyboards was (and still is sometimes) a problem for portable programming. This is why `{` and `}` still have alternative encodings as trigraph (`??<` and `??>`) and digraph (`<%` and `%>`). – Jens Gustedt Jun 04 '16 at 08:10
  • My guess was wrong, so I edited it away -- I found a reference from Martin Richards and replaced my guess with that. – Paul Hankin Jun 04 '16 at 08:15
  • Thanks for the link to "[How BCPL evolved from CPL](http://www.cl.cam.ac.uk/~mr10/cpl2bcpl.pdf)", it's pretty interesting to have an explanation on the evolution from a syntax / language to another! – evuez Jun 04 '16 at 08:58