Re: how is Haskell not robust?

Jon Harrop <jon@xxxxxxxxxxxxxxxxx> wrote:

I do not believe I could recover the costs because the Haskell
market is just not commercially viable yet. Moreover, this is a
chicken and egg problem: many people will not invest any serious
effort in Haskell until these fundamental problems are addressed
and they will not be addressed by industry because it is not
commercially viable to do so. Furthermore, the academics are not
likely to address such problems because it does not constitute
compelling research.

In other words, Haskell has not reached critical mass. IMHO, it
never will.

Well, in the last few years, Haskell development progressed quickly.

In what sense? They're partway to catching up with OCaml as far as IDE
support is concerned but they're still a decade behind F#. They've
invested a lot of time in their own proprietary package system and
have a thousand packages but still poor support for mainstream package
systems like apt and their libraries don't even include a decent hash
table implementation. The core of GHC is not only buggy (e.g. the hash
table problem) but even has serious design flaws that nobody knows how
to address. I cannot even see any evidence that anyone is even
attempting to address them.

It's a sad truth that Haskell has not made it into the commercial market
yet. If anyone would make money with the language or a compiler
package, then it would progress much faster, I think. It has worked for
Linux as well, which has been a nerdy tech toy in its early ages.

The language itself is great IMO, because it's concise, elegant and
interesting. Currently I'm using F# at work, but I'd love to switch to
Haskell.


There has certainly been a lot of noise surrounding Haskell but
virtually no progress, IMHO.

A few years ago, GHC produced much slower code and there were only a
handful libraries, most of which were incomplete or very limited. This
has changed. Although still most libraries are incomplete, there are
much more useful libraries by now.


If new standards could be created quicker and there were more
libraries,

Lack of documentation and interoperability are much bigger problems,
IMHO.

The documentation is there. It's rather that there is actually too much
documentation available, particularly about monads, and most of it is
either redundant or badly written. Beginners don't know where to start.
Real World Haskell is a first step, but as far as I've seen it needs a
lot of improvement.


then it could well get commercially viable. As a language it
definitely has the potential.

Perhaps it might become viable for specific domains but I cannot see a
language with such poor support for large-scale programming garnering
any serious traction in industry.

I don't think that language-level support is poor. It's actually quite
good. What's really missing is a package manager, which better
interoprerates with system-level package management, FFIs for more
languages and better support for dynamic linking.

I'd also like to see explicit OOP support together with FFIs to common
OOP languages.


Greets,
Ertugrul.


--
nightmare = unsafePerformIO (getWrongWife >>= sex)
http://blog.ertes.de/

.

Relevant Pages

  • Re: Where to begin?
    ... to start learning the language. ... There is no support for simple data structures such as ... SwiftForth doesn't even provide a FIELD ... Use my novice package: http://www.forth.org/novice.html This will run ...
    (comp.lang.forth)
  • Re: Is it just a matter of taste?
    ... found Haskell the easiest for me to understand, ... say Erlang does better at parallel execution, ... I've been reading some papers about functional language's SMP support ... formula as to which language to prefer for a particular case. ...
    (comp.lang.functional)
  • Re: What do you LISPers think of Haskell?
    ... And I don't want to "get around" the language. ... Haskell was designed to be the Common Lazy Typed FPL. ... is much deeper in its support for FP than Common Lisp, ...
    (comp.lang.lisp)
  • Is it just a matter of taste?
    ... to take a closer look at functional programming languages... ... found Haskell the easiest for me to understand, ... I've been reading some papers about functional language's SMP support ... General purpose language. ...
    (comp.lang.functional)
  • Re: Question about OO programming in Ada
    ... package, then you can use that type in the child packages. ... But note that, in Ada, a class is a set of types, whereas in C++ ... There is an alternative unique to the Ada language. ... Lists and Nodes ...
    (comp.lang.ada)