Database Theory Links
Thu, 6 Apr 2000 11:38:47 -0700

From: Kyle Lahnakoski []
>"Brian T. Rice" wrote:

>Is DB theory fundamental to a reflective system?  Apparently you do not
>think so; maybe you're right.

Hmm.  I would tend to say that DB theory is at the least very important to
any system.  We need to have better databasing systems in place.

>Implementing my DBOS reflective system I
>believe that any information stored in source code is hard to 
>extract as
>meta information.  You may have noticed my lack of enthusiasm for Joy
>for just this reason.

A misplaced lack of enthusiasm :-).  The cool thing about Joy is that it
makes information stored in source code easier to extract, partially because
the source code is so much closer to data than ever before, partially
because the source code is so much closer to actual running machine code
than ever before, and again partially because the source code is so much
closer to math than most other notations.

Note that when I say "Joy", I'm talking not about the system on the Joy
pages (that's only a proof of concept), but rather concatenative languages
in general, and an ideal functional one in specific.

Joy source is closer to data than the source of any other language: it
consists almost entirely of blank-seperated words, with almost no parsing
needed.  (The exception is 'quotations' and 'literals', both of which
require a small amount of parsing, although only locally.)

Joy is closer to machine code than any other language, because its words can
be compiled or executed in the order written, with no optimization or other
subtlety needed.

Joy is closer to mathematical purity than most other languages, because as a
combinative language it completely removes the need for named variables;
this makes mathematical transformations and manipulation much easier,
including the transformations which are required for effective

>Kyle Lahnakoski                                  Arcavia Software Ltd.