Re: [Caml-list] stl?

[Jon Harrop <jon@ffconsultancy.com>]

On Wednesday 04 March 2009 20:49:14 Yoann Padioleau wrote:
> Jon Harrop <jon@ffconsultancy.com> writes:
> > On Wednesday 04 March 2009 16:48:18 Yoann Padioleau wrote:
> >> I don't think so. I've read the last "history of C++" by Stroustrup
> >> in HOPL-III, who discusses quite a lot about the STL and Stepanov,
> >> and from what I remember unboxing was a big issue
> >> and having "generic" (which is slightly different from polymorphic)
> >> algorithms without introducing performance
> >> penalty that object-solution has with dynamic dispatch was also
> >> a big issue. Those people are not stupid.
> >
> > If they were not stupid, why were their innovations were dropped, e.g.
> > Java, C# and even VB.NET
>
> First, who cares about Java or C# or VB.NET?

About 60% of all professional developers.

> Last time I checked, 
> all the software I use on my machine (linux, mozilla, apache, mysql, emacs,
> X-windows, gtk, gnome, git, transmission, ...) are written in
> C and sometimes C++ (well except the beautiful ocamlopt and ocamldebug).

Of course, most of those programs predate .NET.

> >From what I know, most Microsoft software are still written in C++
>
> (Office, Visual Studio, the kernel, etc), and most Apple software
> too (with sometimes some objective C stuff). Do we have an example
> of a Java killer-app ?

Does Java need a killer app?

> > provide parametric polymorphism from ML (aka generics)
> > instead of C++ templates?
>
> But haven't they added generics in Java because Java programmers
> wanted some of the capabilities of C++ templates ? They even
> use its syntax, and recent Java has added some ugly extensions
> with some star-stuff around it that I don't understand. So I think
> Java generics are closer to C++ templates than ML parametric polymorphism
> and its inference.

Java's generics were by Wadler (Haskell) and Odersky (Scala) and .NET's 
generics were by Syme (F#). They all followed ML and not C++.

> And Java has decided to not follow C++ on many things, they also
> don't have overloading...

Java has overloading but not overloaded operators.

> >> They know about ML.
> >
> > IIRC Alex Stepanov claimed to know ML but then made a series of factually
> > incorrect statements about it, e.g. "phantom types do not exist".
>
> I know ML and I don't know those phantom types thing. And if they are
> "phantom", isn't he right indeed that they do not actually exist?

No. :-)

> Do ML designers really know C++ and all its lastest features too
> as in C++0x ?

The designers of ML obviously didn't know any C++ because they designed ML 
before C++ had been invented. I'm not sure there is much for anyone to learn 
from C++98/C++03/C++0x. They are just flogging a dead horse.

> > His published work includes nothing on anything related to ML:
> >
> >   http://www.stepanovpapers.com/
> >
> > I do not believe they knew about ML.
>
> I think at least Stroustrup mentions in its "the design and
> evolution of C++" book some comparisons with ML.
> But probably at that time ML didn't
> have yet the functor stuff, just the parametric polymorphism.

Stroustrup may well have not known about functors when he was designing C++ 
but not being familiar with parametric polymorphism was silly.

> > I've learned about 20
> > programming languages now and every one taught me something new. C++
> > taught me what happens when you let idiots loose on programming language
> > design and get industry to hype the result regardless of its objective
> > value.
>
> C++ taught me that if you want to be really successful, you need a
> migration path (or you need to wait for old programmers to die or that a
> very different kind of platform arrives so that legacy code
> does not matter any more, e.g. the web).

Indeed. Which raises the question of how I should put an OCaml front end onto 
HLVM...

> > I'm very happy to see C++ dying.
>
> Is it ?

Yes.

> >> > Scripting languages were not so hot at the time, short of Perl, but
> >> > Ruby would easily fit well into the STL idea, just like Lisp also did.
> >>
> >> No, because of the performance penalty of dispatch. Again, those C++
> >> designer guys have strong requirments on performance.
> >
> > Their performance requirements were: destroy the performance of anything
> > we are not familiar with in order to preserve the performance of familiar
> > features
> > regardless of the relative merits of the different approaches. That
> > is really stupid and very counter productive, of course.
>
> Did Stroustrup did that? I never saw Stroustrup criticizing
> other languages (but I've seend many Java people trashing C++).

Stepanov, IIRC, gave a list of languages that he claimed were incapable and 
was wrong. I would not trust his opinion on such matters and consider C++ to 
be more of a get-rich-quick scam than a valuable constribution to 
programming. C++ is one of the few languages I would advise people to not 
bother learning.

-- 
Dr Jon Harrop, Flying Frog Consultancy Ltd.
http://www.ffconsultancy.com/?e