[Top] [All Lists]

Re: [ontolog-forum] How de facto standards are created

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Mon, 17 Jun 2013 23:52:46 -0400
Message-id: <51BFD98E.50508@xxxxxxxxxxx>
Melvin, John, and Kingsley,    (01)

The point I wanted to make is the importance of de facto standards as
a basis for official standards.  A huge number of official standards
that ignored the de facto standards have been ignored by developers.    (02)

I'd also like to cite a "Law of Standards", which I first enunciated
in a note to the SRKB list (Shared Reusable Knowledge Bases) in 1991:    (03)

 From http://www.jfsowa.com/computer/standard.htm
> Whenever a major organization develops a new system as an official
> standard for X, the primary result is the widespread adoption of some
> simpler system as a de facto standard for X.    (04)

Prediction:  According to the Law of Standards, I predict that the
Semantic Web notations will be replaced by de facto standards based
on much simpler pre-existing languages:    (05)

  1. OWL will be replaced by a de facto standard based on Aristotle's
     syllogisms.  The majority of published OWL ontologies do not use
     any features that go beyond Aristotle.  Examples:  Good Relations,
     BFO, and many others.  Those syllogisms have been expressed in
     controlled natural languages for over two millennia, and they will
     continue to be expressed in CNLs.    (06)

  2. SPARQL and SQL will be replaced by de facto standards based on
     a typed version of Datalog, which can also be mapped to and from
     simple CNL sentences.  The types will be specified by sentences
     of the CNL used in point #1.    (07)

  3. For more expressive power beyond #1 and #2, typed Datalog can be
     extended to a full Horn-clause logic-programming language.  The
     usual notations for LP languages can be used by people who know
     them, but the statements could also be translated to CNLs.    (08)

  4. Various diagrams (UML and others) can be used to supplement the
     controlled NLs for points #1, #2, and #3.  Those diagrams are
     familiar for most programmers, and the learning curve for adding
     CNLs to supplement the diagrams is smooth and simple.    (09)

Some comments on your comments:    (010)

> I think you mean Mosiac, rather than Mozilla.    (011)

Yes, I forgot Mosaic.  Mozilla was designed by the founders of Netscape
(many of whom were also the ones who implemented Mosaic).  But they did
a complete rewrite of the code base for Mozilla.    (012)

> There is an element of luck involved too.  Gopher was ahead of
> the WWW, until the U of Minnesota made a licensing mistake.    (013)

Luck is indeed important.  Some people (such as Steve Jobs) make
their own luck and their own mistakes.  Apple developed Hypercard
in the 1980s, but they kept it proprietary.  Tim Berners-Lee used
Hypercard, and it gave him the inspiration for http.    (014)

Another licensing mistake:  Simula-67 was the first object-oriented
language (in 1967).  It was (and still is) and excellent language,
but the developers wanted to charge $20,000 for it.  Philippe Kahn
sold Turbo Pascal for $99, and he got enough orders to fund Borland
without seeking outside investors.    (015)

> I still consider Silversmith the first web (lowercase "w") browser.
> The term "web" existed before WWW...    (016)

Those are interesting points.  Thanks for the history.    (017)

> how do we identify and develop the next killer app? It is easy to
> identify the killer apps after they have major gross revenues.    (018)

Good question.  One reason why Tim B-L's version succeeded is that
CERN was not trying to sell a product.  Their goal was very modest:
enable physicists to share research papers more rapidly.  Academics
from other fields adopted it very quickly.  MOSAIC was also free
because it was funded by the US gov't as free software.    (019)

In general, I would say that every "killer app" started as a solution
to a problem that somebody needed to solve.  CERN recognized the problem
and they asked Tim to solve it.  Then Tim used ideas from a system
(Hypercard) that many people had found useful for related problems.    (020)

Steve Jobs was a good designer because he understood his users.    (021)

> Linked Data has created a killer application for the Web in its ability
> to enable Web-scale structured data representation, publication,
> and publication....
> Google's Guha and Dan Brickley (no strangers to RDF) have also added
> Schema.org [4] to this powerful killer app. cocktail comprised of
> structured data and shared vocabularies...
> Google is encouraging its developers to take advantage of JSON-LD...    (022)

Yes, but.  This is another confirmation of the Law of Standards.
An official standard (the W3C spec's for SW tools) led to de facto
standards based on simpler, pre-existing technology:  Microdata,
RDFa, and JSON can be used with HTML instead of XML, and Schema.org
uses a very simple hierarchy instead of OWL.    (023)

As I've said, OWL hits a "sour spot" in knowledge representation:
too complex to be easy to learn, too limited to be useful for
implementing an application, and too incompatible to be used
with mainstream IT and databases.    (024)

John    (025)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (026)

<Prev in Thread] Current Thread [Next in Thread>