[Haskell-cafe] Why Haskell is beautiful to the novice

[Alberto G. Corona]

In general I think that haskell has a very low productivity because there
are too much distracting and the gap between the real problem and the
libraries is very high since there is a lot of abstraction in them.  In the
current state, even the haskell EDSLs are far from being close to their
respective domain problems, not because Haskell is not enough flexible, but
because it is too much flexible and too much abstract.

At last, a productive programming language must guide the programmer for
creating valid and efficient solutions fast. The haskell type system helps
in the first very well and this is a great advance, but it does not help on
the latter. There are too much alternatives, too much abstraction and with
a lot of impedance between them. All of this in combination with Monad
transformers and the abstruse errors kill any hope of making Haskell a
productive language.

Haskell is not a high level language in the usual sense, but an abstract
metalanguage where the gap between the platform and the concrete solution
is very long in terms of effort.

Still there is a need for an embedded general purpose high level language
in which a novice can start to program fast in Haskell by freely mixing
effects without a master in monad transformers, packages idiosyncrasies and
error intricacies.

 That is why I'm trying to create Transient: a flat monad with very high
level effects included, (upto distributed computing). That eliminates the
impedances and reduce complexity to a minimum level.

2015-09-01 8:59 GMT+02:00 Mike Meyer <mwm at mired.org>:

> On Tue, Sep 1, 2015 at 12:58 AM M Farkas-Dyck <strake888 at gmail.com> wrote:
>
>> I could compare Java and COBOL; it wouldn't make Java worthy.
>>
>
> Of course not. There are application areas for which COBOL is clearly
> superior to - and hence more worthy than - Java. Or Haskell.
>
> You seem to be suffering from the common misconception that there is some
> independent, objective measure of the quality of programming languages.
> That simply is not the case. It will depend on any number of constraints.
> For instance, much as I enjoy writing Haskell and appreciate it's virtues,
> it is less worthy than C/C++ for many of my current projects for the simple
> reason that Haskell code won't run on the processors those projects need to
> run on.
>
> The same goes for teaching an intro programming/CS/SE class. Your measure
> of "worthy" will depend on your goals and audience. Do you want to
> introduce programming as an art that can be criticized and enjoyed? Then
> I'd say Haskell is clearly the choice you want to make. But if the goal is
> to give the students a feel for what programing is like for most practicing
> programmers today, then Haskell falls well behind many other languages. And
> so on through a long list of other possible scenarios with there own
> metrics of worth.
>
> _______________________________________________
> Haskell-Cafe mailing list
> Haskell-Cafe at haskell.org
> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
>
>


-- 
Alberto.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20150901/3d67db52/attachment.html>