Indentation of If-Then-Else

[Cale Gibbard]

These suggestions would, firstly, break lots of code for little
benefit, and secondly, make code harder to read.

The 'then' and 'else' visually separate the parts of the
if-expression, and serve to guide one's eyes when reading code
silently, and one's words when speaking it aloud. Certainly, one can
define a conditional function if' and decide to use that everywhere to
avoid typing 'then' and 'else', but how many people do this? My guess
is that almost nobody does this because it only serves to make code
harder to read. Also, one could simply use case-expressions, but
people still use if-expressions not just for compactness (the
difference there is rather slight), but because they don't read the
same way.

Haskell is not some minimal programming language core intended to be
optimised for the simplest possible machine parsing and fewest
desugaring steps. It's optimised for human readability and
convenience. We have things like do-notation, which while they are not
strictly necessary, serve to filter repetitious noise, while
delimiting things clearly for the reader, and present an expression in
a form which is (often) closer to one's mental model of the meaning.

If-expressions match very closely with the way in which one reads
their meaning in natural language, are not so difficult for machines
to handle, (the current implementations seem to be having no trouble
with them) and don't really complicate the language definition all
that much. Please don't attack non-problems.

On the other hand, I agree that this is usually the "correct" way to
layout an if-expression which does not neatly fit on to one line, or
where the results are at all lengthy. However, given the fact that
there's no chance of ambiguity anyway, I don't consider this to be
something for the language itself to enforce. *That* would be
needlessly overcomplicating the language definition. It's something
for editors to help the user with, perhaps, and it's something for
teachers of the language to point out. Otherwise, it's just not all
that important, let people use whatever style they like best in
whatever circumstance. Neil pointed out lots of reasonable
possibilities, here's another:

if long && complicated condition
   then 0 else 42

I don't use that form all that often myself, but it looks reasonable,
and I certainly wouldn't want the compiler to reject it.

 - Cale

On 22/10/06, Brian Smith <brianlsmith at gmail.com> wrote:
> I agree with Henning. Reserved syntax for "if" is totally unnecessary in
> Haskell. It requires three keywords. Its costs (syntax wise) seem much
> greater than its benefits (which are practically zero). I don't think that
> it makes sense to make further complicate this unnecessary syntax. Instead,
> why not work on eliminating the special syntax altogether? For example, make
> the "then" and "else" keywords optional and schedule them for removal in the
> next revision.
>
> Hennings suggestion that implementations suggest the correct indention is a
> good one.  The existing Haskell implementations do not do a good job
> diagnosing layout problems. As far as I know, they don't even try to recover
> from syntax errors at all. I would like to change both of these problems.
> The proposed layout changes (NondecreasingIndention and DoAndifThenElse and
> the one for case) should be postponed until more work on error diangosis and
> recovery has been done.
>
> Regards,
> Brian
>
> _______________________________________________
> Haskell-prime mailing list
> Haskell-prime at haskell.org
> http://www.haskell.org/mailman/listinfo/haskell-prime
>
>
>