I've been on the Haskell list for over a year now as a bystander interested
in this new area of functional programming. My background is on the
marketing side of the software business and have worked for some of the
large ones (but not Microsoft!).
The recent thread of notes to "Could Haskell be taken over by Microsoft?"
bears out what I've been thinking over the past year about Haskell which is
> ... "How on Earth is this Haskell stuff, not withstanding its merits, ever
> going to make it in the real world?". From what I've read and seen I really
> see Haskell as nothing more than another 'interesting' computer language but
> ultimately confined to the academic/research community.
I disagree. I have been following Haskell for a number of years, thanks to
its Internet availability. I must say that I am a Haskell and Hugs
enthusiast! As it happens, my wife is a top market researcher with a
software background at a large market research company. We've long talked
about Haskell and Hugs, especially lately. I once considered
commercialization of Haskell myself.
> The situation would change dramatically if say, a Microsoft, picked it up
> and ran with it. But they wouldn't do it unless they had full control over
> the language which the Haskell community wouldn't allow.
Someone cited Eiffel as a counter example to your assertion. Other
counterexamples are Standard ML, another excellent language designed by
scientists, which has at least two commerical suppliers now. Another
example of collaborative work was the design of the protocol HDLC, which
supplanted IBM's SDLC. I've heard that IBM promised many delegates trips to
Hawaii if they voted for SDLC instead of HDLC (not out of character for
their earlier history, despite their many scientists).
>On the other hand, Haskell could follow the Linux route with a GNU license.
>No problem with this except how many customers are seriously going to take
>the leap of faith. It is a sad fact of life that the majority of the "real"
>market consists of conservative customers who look for a number of
>(non-product) things from a technology component including supplier status,
>pricing, support, and so on. Having a a number of small companies providing
>Haskell support won't cut the mustard in the real 'big' world.
It is doubtful that business managers will take to Haskell, but
so what? Science is a movement that helped lead Europe from the Middle Ages
and has continued to lead globally, even today. Left to their own devices,
business managers would reinstitute the Middle Ages, possibly excepting the
religious order, with the great irony of more gadgetry.
The analogy, "cut the mustard" ie, the military term "making muster,"
itself reveals the problem with its own line of reasoning: the intent is
to marshall numerous people to grind away at other armies of peole until it
succeds by some means or other. In marketing, this has, at best, a game
theoretic flavor.
Better software might be compared to the engineering of the Middle Ages,
that ultimatly ate away at the military power of the aristocracy: the long
bow, handheld armor piercing firearms and powerful cannons that could
pierce castle walls. Or to the English fleet, advanced 2 or 3 nautical
generations over the Spanish Armada, as is now known.
No one says that the managers have to adopt these innovations, but if not,
their firm is more lilkely to perish. I once worked for a firm that
insisted on programming in assembly language, but it couldn't produce.
Guess what happened to it. The engineering issues with use of a language,
though, are a lot about what paradigms it embodies. Use the right tool for
the job.
>It's fair to say that the jury is still out on Java. It may well turn out
>to be a success and then maybe not but just think of the hundreds of
>millions of dollars that have been spent on its development and marketing in
>the past few years.
Java has a few uphill hurdles:
Dynamic (as opposed to static/lexical binding) limits its use in embedded
systems to non-critical uses. No reactors, pacemakers, satellite launches,
autonomous vehicles, etc., please!
Imperative programming (only!), hence a loss of referential transparency.
This has somewhat the same effect as dynamic binding: the relative
inability to demonstrate, in advance, the soundness of a program.
While writing in Java may be more pleasant than or get to market quicker
than or even be more correct than writing in C++, C++ is not Java's only
competitor. There is a mild tulip bulb flavor to the Java industry,
although this was true of C++ too. Even here in Silicon Valley, I can't
find a shortage of Java programmers, only a shortage of Java programmers
that have worked at Sun, etc.
millions of dollars. Why does the word "shills" come to mind?
>If the Haskell community wants Haskell to be a significant product in the
>computer language and software development markets then the only route open
>is to setup a company whose sole purpose is that ... Everything else (for
>the Haskell community) is just wishful thinking.
One thing we are considering is whether Microsoft will demonstrate some
pride. That could be a symbol of dominance, for example, or a character
property of Mr. Gamesman, himself, Bill Gates. Gates has, for example, made
philanthropic contributions in some cases. Perhaps this could be one. At
any rate, I believe there are many ways to resolve these issues. Let a
thousand flowers bloom!
Byron Hale
Effective Information
[EMAIL PROTECTED]
(408)358-8064
