Am Mittwoch, den 05.03.2008, 10:10 -0500 schrieb Walter Banks:
> Software innovation has dramatically slowed down. Much of the software
> development in educational institutions depends on free software platforms
> often based on 20 year old technology. We are still for the most part declaring
> variables that depend on the implementation environment and not based on
> application requirements. (Recent thread on variable information typing
> covered a lot of the issues in this regard)
True, but historically, software innovation in programming languages has
been partly commercial, partly research (i.e. noncommercial), so I
wouldn't ascribe that to open-source effects.
Here's a shortlist, sorted in roughly historical order:
* Assembly (abstract away from bit encoding)
Commercial, but software was just thrown into the hardware and not sold
separately, so it's a bit unclear.
* Lisp (be simple)
Research origins. Lisps wasn't commercially for several years.
* Fortran (efficient formula translation to assembly)
Commercial, but IIRC software still wasn't sold separately.
* Cobol, PL/I (try a higher abstraction level)
Invented by standards bodies, with heavy IBM influence. Compilers for
these languages may have been the first that were actually sold, though
I'm not sure (selling compilers as a practice may have evolved later).
* Algol 60, C, Pascal (nestable struct/record and array types)
Algol 60 was a research effort.
C was "just for fun" (to port a game).
Pascal was just for teaching programming.
Decidedly noncommercial.
* Smalltalk (first attempt at systematic OO)
Research project.
* Shell, Tcl, Perl (command-line scripting languages)
None of them ever sold as far as I'm aware (there may be exceptions, but
obviously they never got into the mainstream).
* C++, Eiffel (statically-typed OO)
C++ was a research project, compiler vendors came later.
Eiffel was a research project initially, later commercialized by its
inventor.
Again, research roots.
* PHP, Python, Ruby (web scripting languages)
All were Open Source from the beginning. PHP is the basis for other
services for its main supporter (Zend).
* Java (web-enabled, so-so OO)
Developed for commercial purposes. Not sold by inventor, given away to
help increase sales of networking appliances.
Each of the items is an innovation step (though some are of debatable
usefulness). Still, the vast majority of these steps started with
research or OSS projects.
IOW language innovation seems to be independent of commercial
activities.
> The final comment has to do with what is learned from customer feedback.
> Free software generally speculates on applications requirements and does not in
> the organized way that most commercial packages do to seriously address
> feedback from customers real needs.
That's probably the reason why FOSS software generally has a less
polished UI.
> The counter is of course everyone contributes to free software. When was the
> last time an application developer added 24 bit data types to GCC because
> their application needed it?
Right.
If you need something, you can either program it yourself or pay
somebody to do it. I.e. if a customer needs commercial-type support,
it's still available.
Regards,
Jo