Saturday, January 28, 2006
Generative Methodology Glass Bead Games
possibility of complexity arising in grid computing
[368] ß [comment on four issues
(Richard Ballard)
[367] ß [Four Issues facing
Ontological Modeling (Paul Werbos)
[371] ß [Discussion about founding
the Second School (Paul Prueitt)
[151] ß [Comment by Paul Werbos on invariance of natural laws
Communication
to part of the SOA Blueprint Technical Committee (at OASIS) à [144],
Hi
Judith (Rosen),
Yes
indeed, the relationist approaches are slowly (and against much resistance...)
trickling into many scientific fields, after long detours since Heraclites...
It is a fascination process and a wonderful experience being part of it,
especially if one has found (as we think we have) an operational method to
describe the relationist paradigm which unfortunately has been missing. Since
we only understand what we can make (Vico/Nietzsche), this missing
link was behind the lack of broad acceptance of the relationist approach until
now...
About
your questions:
When you talk
about "storing relations instead of data"... how do you assess those
relations? From the information on your site, it appears that "Pile"
is based on an assumption that all relations are considered equal to one
another. Have I misunderstood? Do you store relational patterns (patterns of
relations which are peculiar to certain system types and patterns of relations
which repeat across system types) as well? How would these patterns be
assessed?
The
relation space of Pile is a run-time space with its (uniform!)
objects dynamically generated. What is recorded on disk/memory are
only handles to (some) Pile objects. The relation space where operations
are done is a purely logical space, the translation between the physical memory
space and the Pile space being performed by the Pile engine.
The
Pile space holds only uniform objects in the form of complex addresses (these
addresses are like code, we call them 'combinatory pointers', which is the
actual IP (patent pending) of Pile. They are self-connecting to other such
objects and self-organize into the Pile structure (which by itself can
vary depending on what Pile agent you use or design).
What is
new here is that
a) all Pile objects are relations between other
Pile objects
b) all these relations are referable objects as
well
(So you overcome the distinction between object and
relation or node and link etc.)
c) all interpretations are OUTSIDE Pile space, so you
have no 'semantic contamination' (if you don't mind me calling it such) of
the 'pure' relation space.
Unlike
in a data space, where you always have syntax and semantics mingled up, here
you have a strict separation which allows you to truly formalize data (by
computing pure relations and generating data as result). Just think computer
game, and if you don't have access to those, ask your kids or the neighbor's
kids...:-).
These
relations in Pile form patterns (any data string is generated by a pattern of
relations as a path in Pile). The interesting thing is, of course, that these
patterns can be interpreted by different
(arbitrary) atomic elements (also representing different data types), so you
can see relational pattern similarities across different contexts and data
types. This should be very useful in computational biology (I do not really
care about some of the other applications that could benefit from it, but
possibly without benefit for the rest of us...)
Judith Rosen
said:
I look forward
to hearing more about this. I noticed that there is a link to an open source
site, at www.pilesys.org where it appears the software is downloadable for free
so that people can explore its potential and applicability, is that correct? If
so, I think that's a marvelous approach to avoiding the problems of isolation
and tunnel-vision which can be consequences of overspecialization in any particular
area or field. Impressive.
We have
an 'open lab' under www.pileworks.org
where the inventor, Erez Elul and his partner Miriam Bedoni publish their more
technical papers. You might find these texts a little hard to penetrate, we try
to provide 'translations' at Pilesys.com. The lab version of Pile is
downloadable (as open source) and so is a first demonstrator. At this point I
can recommend this only to absolute geeks, however...
There
is also a very good blog by Ralf Westphal, an experienced developer and trainer
from the mainstram IT world at
http://weblogs.asp.net/ralfw/category/10111.aspx ,
giving
a very good introduction.
We
are also forming an association and institute to foster interdisciplinary
research and exchange, and the initial response is amazing: from
sociologists, computer scientists, historians, physicists to media people and
linguists. I think the time is ripe indeed and the Internet is a perfect communication
structure to spread such ideas even when they are not
(yet) accepted in the ivory towers and corner offices...
We also
have a Sourceforge project and site for programmers, with interesting
discussions about understanding Pile in depth from mathematical complexity,
information science and programming angles. Our current handicap is that
the inventor is one of these rare outsiders, self trained and without any
academic background or history in IT.
Sometimes,
it takes such a fresh outsider look to advance science, but it is
often hard to understand and even harder to accept. But we also have a
very good group of professionals working hard on translation of the theory and
re-implementation of the code (for industrial and scientific use), so these
handicaps should be overcome in a few months.
We even
find growing investor interest now, including from some major players. This was
that biggest problem over the last years, when nobody dared to touch such an
outsider project... But I guess all radical innovators have to go through this
phase, even in our time and age where everybody talks about the need for
innovation but hardly anyone has the vision to see it or
the courage to support it...
Cheers,
Peter
Peter
Krieg
Pile
Systems Inc