Python, C++, Java, Smalltalk, Eiffel,...

[Donn Cave]

Quoth "Michael Abbott" <michael at rcp.co.uk>:
...
| I also was encouraged by the presence of OCaml.  I'm troubled by the absence
| of strong typing in Python, and it's good to see a strongly typed language
| with comparable expressive strength.
|
| I haven't yet tried writing anything in OCaml, but I'm anticipating being
| let down by the library support in comparison with Python, particularly on
| Win32!

In a completely coincidence, I happened to be researching network
file distribution options yesterday and one of them turned out to
be written in OCaml (Unison), with Windows and UNIX support.

But that strong typing thing, I'm not so sure.  Like others who
have commented on this lately here, it hasn't been a significant
real problem in practice for me.  Python's willingness to allow
any object to appear in any context is a freedom, as usual with
benefits and costs, and when you look at how it fits with the OO
model you can really see the benefit.  It puts the object in the
driver's seat.

I managed to implement a tree in OCaml last night - not balanced
or anything, just insert in alphabetical order and print.  That's
how much I know about it - next to nothing.  But for what it's
worth, I was thinking the value of the way it handles types might
be rather accidental to the ostensible intention.  They talk about
reasoning about program correctness, but that sounds pretty ambitious
to me.  Meanwhile though, they have come up with type dispatching
and other features that end up getting some programming mileage out
of types that is missing in the miserable encumbrance of Pascal, C
et al. types.  Maybe instead of FLP, we should say TOP, Type Oriented
Programming.

Have only skimmed the OO part, but it looks like its contribution
to the language might be debatable.  Just as the typing would be
in Python, if it were grafted in somehow - a thing that considered
on its own seems to be of value, but ends up a burden on the language
that already has enough to work with.

	Donn Cave, donn at u.washington.edu