On Syntax and Religion

Ola Bini published a piece showing how the introduction of a special variable, the "it" variable, into the Ioke language makes regular-expression matching less verbose. In short, the it-variable captures the result of the expression examined by the enclosing "if" or "unless" decision. Thus, you can write (pseudo-code):

if(/[a-z]*/.matches("ab")) {
 println(it.matchLength()); }

instead of

var x = /[a-z]*/.matches("ab");
if(x != null) { println(x.matchLength()); }

Personally, I tend to agree with Ola regarding the virtue of this mechanism.

The comments made by readers show how subjective - not to say: biased - the whole issue of "syntactic elegance" is. As of Tuesday evening (GMT), here is a short summary of the various opinions.

Marcelo prefers a Smalltalk-like style:

/[a-z]*/.ifMatches("ab", res,
 { println(res.matchLength()); },
 { println("no match"); })

Stefano wants the variable declaration to be explicit:

if(x = /[a-z]*/.matches("ab")) {
 println(x.matchLength()); }

Paul advocates the introduction of "=>" as a pattern matching operator:

/[a-z]*/ => x

Antares Trader argues for a closure with a parameter:

if(/[a-z]*/.matches("ab")) { |x|
 println(x.matchLength()); }

Johannes says that Option types neatly express the idiom in question.

I think that local optimizations of the syntax (in contrast to global aspects such as: typing discipline, programming paradigm, object model, standard library) are largely a waste of time. Will programmers not use the language because they cannot change the name of the "it" variable? Even fifty (local) improvements to the syntax, will not make a difference. The exact notation chosen for this construct will have nearly no impact on Ioke's future success.

Therefore, my rule of thumb is this: It is worth while to debate over a local syntactic issue only if you have knock-out argument. If you are only going to win by points, don't bother.