Subject: Re: becoming a better programmer
From: Erik Naggum <erik@naggum.no>
Date: 16 Sep 2002 10:59:07 +0000
Newsgroups: comp.lang.lisp
Message-ID: <3241162747773498@naggum.no>

* synthespian
| I don't understand your point of view, in the light of the fact that a
| couple of weeks ago you said that SGML and XML were "braindead."

  Well, let me put it this way: The XML crowd believes that if you only add
  enough markup, everything humankind has ever dreamt of will suddenly emerge
  from the chaos.  I believe that this is a very serious misunderstanding of
  both chaos and emergent properties and that more markup will, in fact,
  prevent them from emerging.  XML-style markup effectively prohibits the
  multiple perspectives on the same information that makes it usable for
  multiple purposes.  In this sense, the more you employ XML to achieve your
  goals, the more irrelevant the result will become.

| That really kept me wondering for days...I kept thinking "well, how would
| you view, then, a thing like the Semantic Web from a Lisp perspective?"

  Note that the Semantic Web so far is being realized with tons and tons of
  markup that fail miserably in being the flexible declarative language to
  describe semantics in web pages that we need.  I consider the efforts by the
  people who work on this without any understanding of how people have dealt
  with semantic classification of information prior to the Internet to be a
  colossal waste of effort.  They will, in all likelihood, reinvent everything
  badly, being more interested in concocting useless syntactic monsters than
  the information infrastructure that is necessary to realize their ideas.

| Would you care to expand on your views regarding the Semantic Web, XML being
| "hopeless", SGML being "braindead" and all that in what regards a Lisp
| approach to those important questions?  That is, how would *you* go about
| it, if I may ask?

  This is a very broad question.  I think the Semantic Web will be realized
  when the analytical capacity of software has progressed to the point where
  people are actively encouraged to help the computer by communicating their
  intent and our human-computer-human communication changes out language and
  our communication skills.  For instance, when it becomes rewarding for the
  user to make the steps of his argument clear to the computer so the computer
  can help him communicate with other computers that might more effectively
  argue his point with the originating human who will listen more to his
  computer than to other people, the user will benefit from communicating in a
  way that removes much of the ambiguity of our current language.  This will
  not happen if the amount of work necessary to communicate with the computer
  is as immensely complicated as the current XML-based Semantic Web.

| All these things matter a lot to me, being from the medical community, and
| quite aware of the importance that being able to integrate the data that
| continues to grow at exponential rate in the field (genome, for instance,
| the barrier that tradional statistics methods are up against in a large
| medical database, etc...)

  I envision a future where people are sufficiently encouraged by computers to
  learn the skills of rhetoric, argumentation, and logic to actually achieve
  what education cannot achieve when there is no clear benefit to learning
  either of these skills.  I also believe that something like Dewey's or the
  Universal decimal classification system needs to be taught as the optimal
  means of searching for information with computers.  (The actual numbers is
  not the point -- the hierarchical structuring of human knowledge that is
  embodied in the coding into numbers is an unparallelled achievement.  The
  silly re-invention of "ontologies" in the Semantic Web context is pathetic
  in comparison.)  All of this actually means that philosophy, epistemology,
  the nature of concepts, etc, will have to become fundamental in people's
  approach to information.  Today, we are hampered by many false starts in
  these areas and much muddled thinking and expression thereof.  I believe
  that over time, probably many decades, perhaps even centuries, we will get
  rid of many of the historical accidents in our language and communication
  and will consider the new accident of using computers a more fundamental
  property than it is today.  When you cannot get anything /done/ being sloppy
  and incoherent, people will just adapt.  The school that says that computers
  should adapt to people is wrong for many reasons, the most important of
  which is the misguided notion that what people do today is better than whet
  they will do tomorrow under new technological influences, that yesterday's
  (or yestercentury's) technology that shaped our language and thinking, is
  better than our current or future technology, is keeping us all back.

  I am, obviously, not advocating that we go into some "logical" language or
  that we "speak mathematics" (frequent hysterical rejections of improvements
  in human communication usually take this form by mathofobes), but that we
  consider the benefits that befall us when adapting our natural language to
  the requirements that follow from wanting to get the most out of exchanges
  with our communication peers.  When our peers become computers, as the
  Semantic Web clearly foresee, people should, indeed must, change to make
  themselves better understood and the exchange more fruitful for all parties.

  If the Sapir-Whorf hypothesis holds, we must change our language in order to
  become more "rational" vis-a-vis the computer.  This could become a "class
  distinction" in the future, where the educated communicate mostly with their
  computers and the illiterate mostly with people.  It is at times like this
  that my desire to see immortality in our lifetime just gets more intense.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.