[PRL] annoying types ARGH!

[Matthias Felleisen]

On May 12, 2005, at 7:53 PM, Carl Eastlund wrote:

> On 9/1/05, Matthias Felleisen <matthias at ccs.neu.edu> wrote:
>> I just wrote this little program for Karl:
>>
> [snip]
>>> let distance((x1,y1),(x2,y2)) =
>>>     let dx = x1 -. x2 in
>>>     let dy = y1 -. y2 in
>>>       sqrt(dx *. dx +. dy *. dy)
> [snip]
>> Before I finished, I had_10_ type errors, one was real. Can you guess
>> what went wrong?
>
> Well, I would guess that leaving out the "." after the +,-,* would
> cause most of those.

Correct. They bubble thru and trigger others.

The only proper type error was in the inShape function, where I had 
used the wrong "default" because I was too lazy to finish the function 
complete. The logical error was in <= . I compared it to the wrong 
value of proper type. (I used to keep an error log when I programmed. 
For some 10,000 lines of code a year, this was by far my most common 
kind of mistake.)

> Also, if it were me writing the code I'd forget the "rec" after the
> "let" and probably screw up the commas somewhere (OCaml's tuple syntax
> is dumb if you do anything unexpected with a comma).  I don't know
> which of these errors you'd call "real" though.

I forgot the rec for a moment but it was obvious. Otherwise I coped 
with OCAML syntax after three years of abstinence w/o a hitch.

>> Yes, I am on record saying that I'd use OCAML if I worked in industry,
>> but boy this makes me reconsider this one.
>
> Well, this is my second summer of OCaml hacking and for
> interpreter-writing it still seems like the right thing.  One might
> claim, "but not every job is interpreter-writing", but mine all seem
> to be so this conclusion works for me.  Still, I could write a very
> long list of things I'd change about OCaml.  Starting with decoupling
> the module system from the file system.  ARGH.

Yeap, it's a crime. And mzscheme is only an epsilon better in this 
regard.