🚀 go-pugleaf

There's been lots of comments about to-semi-or-not-to-semi, including:
lavinus@csgrad.cs.vt.edu:

   There's a reason for the semicolon that seems to be missed here.  Yes, as
   several of you have pointed out, it is perfectly feasible to define a
   language which requires *no* end-of-statement separators - no semicolons,
   end-of-line's, or anything.  One can even drop the semicolons from Pascal 
   pretty easily. 

   But.. this assumes that a compiler for this language is always given correct
   programs.  The semicolons are there so that if the compiler hits an error,
   and gets really confused, it may easily find the end of the statement and 
   recover to continue parsing the program.

I implemented a semicolon-free language (it happened to be a "real-time
blackboard system" language). The compiler never seemed to have any problems
finding the end of statements and recovering; neither did the programmer. The
only usability issue we noticed was that where a statement was incomplete (but
not visibly so) and was followed by a long comment, the compiler would object
at the first token following the comment, rather than at the non-existant
semicolon after the incomplete statement.

This seemed to me a reasonable trade; the programmer *never* made trivial
semicolon errors, and semicolon errors seem to occur with monotonous regularity
in Pascal or C.
--

Regards, Kers.      | "You're better off  not dreaming of  the things to come;
Caravan:            | Dreams  are always ending  far too soon."