[clean-list] Clean and scientific programming
Thu, 30 Aug 2001 19:53:44 +0200
My dear, how many times have I seen postings entitled "Pascal and
scientific programming", "Haskell and scientific programming", "Java ..."
I have no answers to such questions as "why people use XYZ which is not
a language, but a disease instead of ZYX which is fabulous". First, because
usually I have no answers to questions which are a bit silly and insinuating.
A popular language is a popular language whether we like it or not.
Second, because there are modes and styles, which are not always rational.
Why people like some TV shows, or some style of music? (and make a lot of
money, or lose a lot of money in these domains)...
But I have some questions instead of answers.
1. How many papers showing some development or some applications in Clean
have been presented during last 4 years on international conferences?
Published and made accessible on Clean pages?
2. BTW, How often the people from the Clean team go to ICFP conference
and talk to others? Did anybody propose a Clean workshop ever?
3. Yes, I am aware of the nice game library. How many games have been
constructed with, and where are they? Mike Wiering pages are dead.
Haskell community asked for years such questions as the original: "why the
scientific community doesn't use it"... But they form a mature
community now, known by others and Hear! Hear! apparently appreciated by
Microsoft. They communicated, there was and there IS a lot of life around.
There is a problem with having a nice set of concrete applications, but
they will come, I believe that a given threshold has been reached.
Clean needs more life and MORE COMMUNICATION. I mean external communication,
not such a relatively closed list.
I must say that I use Clean for some years regularly, mainly for teaching
and around. I wrote a silly paper on modelling of 3D objects using Clean,
showing how to operate with higher-order functions to implement extrusions,
distortions, etc., all parametric stuff, in a very elegant and compact way.
I partly suggested to Thorsten Zoerner the idea of using the functional style
to parameterize implicit surfaces. My students wrote a working ray tracer
in Clean, and I wrote a package for the generation of procedural shaders
for such an engine. I rewrote in Clean my stuff for doing lazy automatic
differentiation, including differential forms. All that remains somewhere
on my disks, and so what?
We might produce a lot of libraries scientific or others, and it might not
help. As a former physicist I would remind you that the biggest libraries
such as the CERNLIB, later NAG, etc. have been constructed incrementally.
People built working, CONCRETE applications, and pieces of code which were
potentially shareable were distilled into library modules. In such a way
* the modules have been tested, and are thus more trusted.
* there was an "automatic" communion between various developers, who have
at least to adapt to a common documentation protocol.
New Clean home pages are nice (and slow to load, and a white text on a light
blue colour needs eyes and screen better than mine...) But when I read:
"Currently most applications of Clean are in rather technical areas
(control systems, measurement tools, simulators, compilers,
etc.) and in GUI-programs (editors, development tools, etc.). This
has two reasons. First, ..."
I want to scream: I DON'T CARE ABOUT REASONS! Tell me where are those
applications! I want to inspire myself from concrete examples, the
existing (IO tutorial, GameLib, MolVac...) are duly acknowledged, but
I want more, especially those simulators and compilers.
That's it. I wonder how many Cleaners will be at ICFP this year, which
begins in 3 days....