Doubts about functional programming paradigm

[Doug McIlroy]

Though this post was prompted by beginners at haskell.org, it 
seemed pertinent enough for cross-posting to haskell-prime.

> I just wanted to make a point that learning haskell is *much*
> harder than learning most other programming languages

I would not put it that way. One gets so much farther in
capability almost instantly with a functional language.
I know. At the dawn of functional programming, John McCarthy
could teach in one hour everything from cons up to symbol
manipulation that would have been a month's project in
Fortran. (My account of that hour is at
http://www.paulgraham.com/mcilroy.html.)

What's hard about Haskell is that its landscape extends into
terrains not imagined elsewhere. As long as you stay in the
flat Floridian landscape of Fortran et al, you don't need
to explore the Himalayas of Haskell. But of course the very
exhilaration and inspiration of the high peaks draws one
into the "*much* harder" training necessary to surmount them.

> the learning aids that are available are not yet cohesive
> enough to present a clear path ahead for the average programmer.

I agree with this. Alas, there does not exist a definition of
the Haskell one reads about on the haskell-cafe mailing list.
I treasure my hardcover Haskell 98 report--so rare you can't
even find it in the used-book marketplace. Haskell 2010 was
obsolete as soon as it was promulgated. Contemporary Haskell as
practiced by cognoscenti flaps in a gale of language pragmas.
There is no authoritative source about these pragmas. They are
listed and described in the GHC User Guide, but that source
all too often defines solely by example, not even bolstered by
a formal syntax specification.

I earnestly hope the newly revived Haskell-prime committee can
rectify this state of affairs, and that GHC will provide a
compliant implementation.

Doug McIlroy