Re: language design tradeoffs

alvin@eyepoint.com (Alvin Starr) writes:


>Is a little verbosity a bad thing? I work with a language(Turing) that will
>allow short forms so that
> if a < b then
> x := y
> end if
> can become
> [a<b:x:=y]
> the second is much harder to maintain when an 80,000 line program is
> completly written this way.


Is it really? I disagree completely. If you are thoroughly familiar with a
language, then short forms present no problem. Using >> for loop
termination (or whatever it is - I don't have my Turing report handy)
becomes second nature if you use the language a lot, and so it doesn't
become a problem. Another example is the ? : construction from C.


And yes, I do think that verbosity is a problem. I'm not running so close
to my quota that I am worried about the space taken up by the extra words,
but who would honestly rather write begin...end a few hundred times than
{...} ?


>Turing has no statment seperators/terminators.


I stand corrected. I don't actually use the language much - it seems too
much of a computer science language rather than one for serious
implementation.


While I have the floor, I'd like to ask if anyone out there can explain
why Pascal allows gotos but disallows early termination of loops or inline
returns, which are functionally equivelant? It seems to me that the whole
language wasn't thought out very well.


Andrew Bromage
--

Return to the comp.compilers page.
Search the comp.compilers archives again.