Subject: Re: New commercial product written in Lisp looks like a winner.
From: Erik Naggum <>
Date: 1998/08/23
Newsgroups: comp.lang.lisp
Message-ID: <>

* Hannu Koivisto <> [rant abridged]
| I haven't been complaining about the situation before simply because I
| think nothing is going to happen to it anyway.

  sometimes, when I look at unemployment statistics that inexplicably make
  the news and about all the people who somehow "want a chance", I cannot
  for the life of me forget all the stuff that needs doing, yet doesn't get
  done.  this is part of the reasons why I decided long ago that I would
  neither employ people nor be employed in the regular sense.  I just want
  to take care of some of the things that need doing.  somebody will want
  to pay for having annoying problems _solved_, especially when they pay a
  lot for people who keep it there, only somewhat less annoying, and if
  there's one guy or company who wants to pay well enough that I can make a
  comfortable living off of it, hey, that's pretty cool, and it hasn't
  failed me yet in 14 years of running my own business this way.  I don't
  _expect_ anything for free, but I do get a lot of stuff for free because
  people think what I do is useful, yet recognize that it is impossible to
  provide it in an employer-employee relationship or in our wonderful
  "taxation economies".  _most_ of the stuff that needs doing is like that:
  it doesn't make sense to suffer the enormous costs of employing people to
  get it done.  what we've done to the small needs that societies before us
  satisfied (with full employment, btw) is to make them _impossible_ to
  satisfy because it isn't worth the overhead it takes to provide them in
  the goverment-recognized fashion with taxes and employment rights and all
  that stuff.  and worse yet, even if we _could_ employ people without all
  the costly stuff, it costs too much just to live in our modern societies
  that it wouldn't attract anyone to actually do it.

  the same holds for features in programming languages in the open market.
  it doesn't pay to set up a support organization for every small feature
  that a handful of people need, yet don't want to pay any real money for.
  it's not that we're not rich enough to pay for it, it's that doing it for
  others requires so much more than doing it for ourselves, and doing it
  for pay requires _yet_ more, making it economically infeasible to do
  stuff that very few people want.  however, every single feature has at
  least one supporter behind it.  sometimes, that supporter just _needs_
  it, like the unemployed people need a job.  sometimes, the supporter can
  go ahead and implement it all by himself and he doesn't even stop to
  think that others might find it useful, or he does, and then frowns at
  the high costs of advertising it, selling it, supporting it, etc, and
  then somebody else, who just _needs_ it, won't know about it.  this is
  all a _societal_ problem.

  it's all a matter of communication: if you could describe what you've
  done or what you need done in such a way that a computer can help you
  find what somebody else wants or has already done, without the high
  overhead involved in normal transactions and without the high cost of
  expression yourself very clearly, then that computer or network of
  computers could perhaps again make possible what smaller societies that
  _had_ total communication did, only now with computers taking care of the
  total communication, and also the almost cost-free shipping of the works
  involved.  gotta find a way to make money transactions cheap, though.

  until this wonderful world unfolds itself, you need to talk to the people
  you want to get something from yourself, and you need to offer them some
  incentive to give it to you, or they won't.  if you "complain" and think
  it's never going to happen, then, voilà! _nothing_ happens, just like you
  thought it would.  if you don't even help build that total communication
  network of computers, your _best_ bet is that somebody else does, and I
  wouldn't hold my breath for that one, either.
| So, I thought instead to shut up and start writing my own compiler.  If
| all goes well, perhaps I'll write next time after a year or two about how
| I'm proceeding.

  what I don't get is why you don't write the software in Common Lisp (or
  Scheme, if you're as much into reinventing wheels as you seem to be),
  profile it and optimize it to death or rewrite the parts that needs tight
  CPU and memory control in languages suited to _that_ task.  (which, by
  the way, is more likely to be Fortran than C if your goal is fast numeric
  code -- C and especially C++ are actually _bad_ at it.)

  but go ahead, write your own compiler just because you can't get feature
  X without even _talking_ to the vendors about your needs and desires and
  willingness to fund their development.  it's just like reinventing a
  whole new societal structure just because you can't a job in this one
  when you never ask anybody what they would like to pay you to do.

  the future of Lisp, indeed any future, lies not in the people who want
  something solved for them, but in the people who solve problems without
  having to find a huge market _first_.  even academia, where this was once
  the working motto of most researchers, is now succumbing to the masses
  and their "needs".  but the masses only have "needs" as shown on TV --
  they could never invent something new to need if it didn't already exist
  -- _that's_ the task for people who can solve them and then people become
  (or are made) aware of the problem that can go away.  however, this is
  not going to make anybody billionaires.  solving problems never does
  that.  if you want to get rich, you must _never_ solve a problem, only
  make more and more bogus ways to make it progressively less annoying, yet
  keep it in everybody's mind at all times.  I sincerely hope Lisp doesn't
  get that kind of future.

-- is about my spam protection scheme and how
  to guarantee that you reach me.  in brief: if you reply to a news article
  of mine, be sure to include an In-Reply-To or References header with the
  message-ID of that message in it.  otherwise, you need to read that page.