Subject: Re: Java portability (and not about Lisp at all!)
From: Erik Naggum <erik@naggum.no>
Date: 1999/02/18
Newsgroups: comp.lang.lisp
Message-ID: <3128341921012330@naggum.no>

* "Jim White" <jim@pagesmiths.com>
| By virtue of the fact that performance is directly a function of
| resources devoted to development, as time and resources continue to be
| devoted to the JVM, it will eventually outperform practically every
| other language compilation/runtime solution.

  to get this behavior from the business community, you would have to
  encourage investors with a linear growth in features.  once you go
  logarithmic on investors, they find something else with exponential or
  linear growth.  Microsoft's pyramid game of trust in its future ability
  to fix past mistakes will be killed just this way.

  I'm frankly suspicious of the high interest in Java at the moment -- from
  what I have read about the "AI summer", people's hopes and expectations
  were so high that actual achievements were disappointing, and after that
  followed the "AI winter" and reasonable, realistic people turned away
  from AI out of a sense of having been fools to have believed the hype.
  the still frosty reactions to Lisp stem from this period.

  the treatment that AI got was bad enough, but I don't see any evidence
  indicating that something of importance has been learned from it, and
  this is a bit surprising considering the number of Lisp people in the
  core Java community.

  so I'm frankly not sure Java will sustain investor interest long enough
  to overtake much anything since the principle attractor is currently a
  disturbing (to me) growth in features with premature specifications,
  which only _add_ to the distance between Java and mature languages.  add
  to this that fancier features require significantly more resources to
  specify well because fewer people really know how they should work and
  much fewer people will have the mental capacity to relate them to the
  rest of the language if they are not well-designed.

  Java effectively killed the investments in C++ that were also slated to
  outperform every other language according the hype that has now been put
  to shame.  _perhaps_ Microsoft's Java killer "COOL" (geez, how nerdy can
  you _get_?) will just be enough of an investment detractor that COOL will
  be another Microcruft product and Java will struggle with buggy and fancy
  features for a long time to come?

  I hate it when political overtake technological reasons for success, but
  Java may well have put itself in such a position by having its own
  political motivation and in effect playing Microsoft's con game.  those
  who live by politics will die by politics.
  
#:Erik