Re: Current work in compiler/language design.

Bob Hathaway writes:
> Yes, my opinion is biased towards the OO paradigm, but so is the
> programming language community, in my opinion. ...


I'd disagree, Object Oriented is good, but it's a methodology, not a
programming language specific thing, I can write OO code in Pascal,
or C, or Scheme, or ML, some languages support it better. I don't
really think that the programming language community is biased toward
OO, I suspect that if it were actually looked at what most code was
being written in, that it would be some god awful language like
fortran or cobol.


> Functional programming solves mathematical problems in a clean way. Is
> this general purpose? No, ...


No, SML, one of the most commonly used functional languages bootstraps
itself in SML, as does LML, both of the compilers are written in ML, and
they both compile to machine code. I don't understand what you mean about
only allowing single (constant) assignment, as I can return anything I
please from a function in ML. I've written some semi-complex programs in
ML, and seen a number of complex ones written.


I've written some programs in ML that I would challenge anyone to implement
cleanly in another language, the simplest of which is the infinite stream of
prime numbers.


[lisp stuff]


I don't really like Scheme, I think the syntax sucks. I know people that
disagree, and program in Lisp/Scheme all the time, I suspect it is to some
degree what you are familiar with.


[programming style comments]


Bad code can be written in any language. My dad talks about how he
wrote structured Assembly code. It can be done, it's just more painful.
---


It seems to me that in general the problem is that what you think is the
direction of programming stuff depends on what you have used recently, and
what you are using in general. I have my own opinions about what I would
want a programming language to have in it, but to a large degree it amounts
to I want the compiler to figure out what I mean and implement it, which
just isn't very feasible now.
--

Return to the comp.compilers page.
Search the comp.compilers archives again.