Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: good references re. ontologies for S*m*ntic W*b



Avril,

I'm a great fan of open-source software, but
there are many different ways of achieving
openness.

AS> It seems that the best thing for CYC would be
 > to just give it away for free in order to get more
 > people using it and not forgetting it, or maybe
 > it is already too late for that.

Actually, that's what they're doing.  They have
already made a large subset called OpenCyc
available for free:

    OpenCyc is the open source version of the Cyc
    technology, the world's largest and most complete
    general knowledge base and commonsense reasoning
    engine.  Cycorp, the builders of Cyc, have set up
    an independent organization, OpenCyc.org, to
    disseminate and administer OpenCyc, and have committed
    to a pipeline through which all current and future
    Cyc technology will flow into ResearchCyc (available
    for R&D in academia and industry) and then OpenCyc.

    http://www.opencyc.org/

It's even available on SourceForge:

    http://sourceforge.net/projects/opencyc/
    SourceForge.net: Project Info - OpenCyc

AS> ... but it is hard to think of new better means
 > of how the evolution process of SW could be guided
 > than those that are already in use.  Any ideas?

I would suggest the development models for the three most
successful open-source projects:  Gnu, the WWW, and Linux.

  1. Richard Stallman started Gnu by contributing a large
     amount of his own software to the world and by starting
     the Open Software Foundation to establish the ground
     rules for Gnu projects and to coordinate the licensing
     and distribution of the results.  Anyone is welcome to
     start a new project, but to get a Gnu stamp of approval,
     they have to agree to OSF conditions for licensing,
     participation, and distribution.  The most successful
     projects, which formed the core of the original Gnu
     and are still the most widely used components, were
     clones or variations of Unix utilities.  The core
     datatypes and APIs of Gnu conformed to the Unix
     standards, and they contributed to the definition
     of the POSIX standards for Unix-like systems.

  2. Tim Berners-Lee and a few colleagues started the WWW
     by designing a subset of SGML called HTML and defining
     the http protocol for using the already established
     Internet.  They gave their software to the world, and
     other groups built many new kinds of tools, including
     Mosaic, which eventually grew into the Netscape Corp.,
     which for the formative years established the de facto
     standards for the WWW.  Every browser in the world today
     is either a direct descendant of Mosaic or an imitator
     of one of its descendants.  There have been many additions
     and modification, but the current "look and feel" of the
     WWW was established by Mosaic and Netscape.

  3. Linux Torvalds began by adding extensions to a small
     Unix-like OS called Minix so that he could run a larger
     subset of the Gnu software on his home computer.  As he
     continued the development, the original Minix was replaced
     piece-by-piece with his own code, which he then made
     available on the WWW.  Other programmers contributed to
     his software, which eventually became Linux.  Over the
     years, Torvalds has served as a benevolent dictator for
     what is accepted or rejected as part of the Linux kernel.
     Although he has the ultimate power to determine what is
     in Linux, his power is checked by the Gnu license:  anybody,
     anywhere in the world who wants to do something different
     from what Torvalds has decreed is free to do so, and many
     have.  That option keeps Torvalds honest:  his power would
     vanish immediately if he started to make decisions that
     failed to win the approval of the overwhelming majority
     of Linux users and developers.

My major complaint about the SemWeb is that they violated
the principles that made Gnu, WWW, and Linux successful:
they did not start with small clones of popular software,
such as Gnu or Linux, nor did they start with tiny projects
that could be quickly written, revised, and rewritten in
competition with other small projects, as in the original WWW.

Instead of starting with a design competition among many
small projects, they edicted untested technology for everything.
GML, SGML, and HTML were proven technologies for adding markup
to documents.  Javscript, ECMAscript, and other scripting
languages were proven technologies for adding function to
web pages and other documents.   But the idea of using XML
syntax for every possible language was untried and untested.
PHP is a very successful new language that demonstrates the
value of the Javascript approach rather than the XML approach.

I would feel more comfortable about the SemWeb if the W3C
would admit that perhaps the Javascript and PHP approach
might be a reasonable alternative for some purposes.

John Sowa

PS to Cathy:  In an earlier note, I suggested Avril's MS
thesis as good survey of ontologies, and I recommend it
for your bibliography:

    http://www.cs.helsinki.fi/u/astyrman/gradu.pdf