Am Mittwoch 23 September 2009 02:51:59 schrieb Jimmy Hartzell:
On Sep 22, 2009, at 8:01 PM, Jimmy Hartzell wrote:
I am in love with this proposal: http://www.haskell.org/haskellwiki/Accessible_layout_proposal (Richard O'Keefe:)
I hadn't read it before. Now that I have, I really do not like it. "Syntactic sugar causes cancer of the semicolon" as Alan Perlis once said, and to my taste this proposal definitely counts as cancer of the semicolon. In effect, its purpose is to overload vertical white space.
Any time that you have something where you think you need this, it's likely that a better solution is to break what you are doing into smaller pieces.
I don't like it either. I have not nearly a s strong feelings as Mr. O'Keefe, but to me it doesn't look right.
Well, look at code like this:
wrapParens str = concat [ "(", str, ")" ]
(And yes, I realize you can do something like this with 'printf "(%s)" str'.)
Or, what I do: concat [ "(" , str , ")" ] (of course, here I would just write '(' : str ++ ")"). I admit it looked odd for the first couple of hours, but now I find it nice, clean and systematic.
First off, there is a serious issue with the commas. You should at least be allowed to have a comma after the last element, a la Python. Otherwise, the last one is randomly special in this list, and in a format like this, I regularly edit code accidentally leaving off commas, yielding stuff like:
concat [ "(", str, ")" -- (oops, no comma!) lineEnd -- forgot I needed this ]
And that is avoided, because a missing comma leaps to the eye.
which (of course) results in a very confusing type error. Meanwhile, you have to format your code very awkwardly, as the closing bracket can't be in the left-most column,
Which is a good thing in my eyes.
In summary, I have to spend a good portion of my time coding Haskell dealing with the fact that I have a lot of {'s, ['s, and ,'s to keep track of, and they rarely fit on one line (records, ADTs, lists). I have to spend a significant amount of my coding time finagling the layout to look sensible, and I don't think anyone would claim that I just shouldn't use records or ADTs.
I see your point but remain not liking the proposal.