Lisper had to learn OCaml to get a job
Jon Harrop
Here is an interesting case of Lisper blub:
http://www.podval.org/~sds/ocaml-sucks.html
"OCaml Language Sucks" by Sam Steingold
There is also a hint of Greenspun:
"Pattern matching is a very powerful tool, should be easily implemented as
a Lisp macro."
Always good to see. :-)
--
Dr Jon D Harrop, Flying Frog Consultancy
The OCaml Journal
http://www.ffconsultancy.com/products/ocaml_journal/?usenet
Rainer Joswig
Jon Harrop <j...@ffconsultancy.com> wrote:
Ignore Jon Harrop. He is a known spammer and troll.
___________________________
/| /| | |
||__|| | Please don't |
/ O O\__ feed |
/ \ the trolls |
/ \ \ |
/ _ \ \ ----------------------
/ |\____\ \ ||
/ | | | |\____/ ||
/ \|_|_|/ | __||
/ / \ |____| ||
/ | | /| | --|
| | |// |____ --|
* _ | |_|_|_| | \-/
*-- _--\ _ \ // |
/ _ \\ _ // | /
* / \_ /- | - | |
* ___ c_c_c_C/ \C_c_c_c____________
Kaz Kylheku
> Here is an interesting case of Lisper blub:
There is no claim anywhere on the page that Sam Steingold had to learn
OCaml in order to get a job. Sam doesn't reveal at all how he uses
OCaml at work, or why, or even whether it was imposed as a
requirement.
Don Geddis
> http://www.podval.org/~sds/ocaml-sucks.html
> "OCaml Language Sucks" by Sam Steingold
Much thanks. I had a vague distaste for OCaml before, but I didn't know
much about it. I appreciate the link with the many reasons why OCaml's
language design is poor.
-- Don
_______________________________________________________________________________
Don Geddis http://don.geddis.org/ d...@geddis.org
Often, when I am reading a good book, I stop and thank my teacher. That is, I
used to, until she got an unlisted number.
-- Imitation Deep Thoughts
Karol Skocik
How can you say that ocaml language design sucks? Doesn't ocaml have
the *holy pattern matching* built in (tm) while CL only has several
external matchers around? ;)
Dan Bensen
>> http://www.podval.org/~sds/ocaml-sucks.html
>> "OCaml Language Sucks" by Sam Steingold
> much about it. I appreciate the link with the many reasons
> why OCaml's language design is poor.
There's a good discussion on Reddit:
http://programming.reddit.com/info/20dh6/comments
OCaml may not be the language for you, but it's a really
good language for big Enterprise projects. It's too bad
the language is so closely associated with such a troll.
--
Dan
www.prairienet.org/~dsb/
Alex Mizrahi
(you :wrote :on '(Thu, 28 Jun 2007 15:06:53 -0500))
(
??>> I had a vague distaste for OCaml before, but I didn't know
??>> much about it. I appreciate the link with the many reasons
??>> why OCaml's language design is poor.
DB> There's a good discussion on Reddit:
DB> http://programming.reddit.com/info/20dh6/comments
DB> OCaml may not be the language for you, but it's a really
DB> good language for big Enterprise projects. It's too bad
DB> the language is so closely associated with such a troll.
Sam was not convincing?
really nice code for Enterprise:
(Int64.to_float (Int64.sub (Int64.mul q (Int64.of_int n)) (Int64.mul s s)))
/. (float n)
type system so advanced, that it cannot do automatic arithmetic
promotions -- but it's so conceptually pure! really, who needs readable
arithmetic at Enterprise? ah, certainly such notation makes Enterprise
projects much bigger, if that was your point.
this programming language is so powerful, so "meta" level, that programmers
need manually keep track of where they need to write +. and where +, or
where that ugly Int64.add, and then they also need to manually do all type
conversions..
well, maybe some people like this language, as it fits their thinking.. some
people even like Perl, and use it for Enterprise programming.. as for me,
Caml looks like some statically-typed Perl :).
sorry, but in my opinion Caml is very idiosyncratic for lispers. even SML,
being essentially same, is less annoying -- at least it handles arithmetics
automatically :)
)
(With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
"scorn")
Jon Harrop
> sorry, but in my opinion Caml is very idiosyncratic for lispers
Blub.
André Thieme
> How can you say that ocaml language design sucks? Doesn't ocaml have
> the *holy pattern matching* built in (tm) while CL only has several
> external matchers around? ;)
OCamls PM is not holy - it is not extensible.
I think it doesn't work for multidimensional arrays and not for
objects. In Lisp you can simply take one of the existing PMs, extend it
or change its syntax as much as you like, just by writing a few macros
on top of that existing PM.
André Thieme
> How can you say that ocaml language design sucks? Doesn't ocaml have
> the *holy pattern matching* built in (tm) while CL only has several
> external matchers around? ;)
OCamls PM is not holy - it is not extensible.
Jon Harrop
> OCamls PM is not holy
It is a good foundation, like the pattern matchers in Standard ML and
Haskell. This is essential if you want good performance.
> - it is not extensible.
You can extend it using macros.
> I think it doesn't work for multidimensional arrays
It works on arrays and multidimensional arrays are just arrays of arrays in
OCaml:
# function [|[|1; 2|]; [|3; 4|]|] -> 0 | _ -> 1;;
- : int array array -> int = <fun>
> and not for objects.
Correct. However, objects are a form of encapsulation, so it does not really
make sense to pattern match over the object itself but, rather, over a
property (which can be done).
# let f o = match o#length with 0. -> 0 | _ -> 1;;
val f : < length : float; .. > -> int = <fun>
> In Lisp you can simply take one of the existing PMs, extend it
> or change its syntax as much as you like, just by writing a few macros
> on top of that existing PM.
Everyone would be better off if someone built this foundation, IMHO.
sds
It would be more accurate to say "a lisper wanted to know whether the
fuss about ML and static typing was justified, so he found a job that
paid him to learn OCaml" :-)