Duane,
Please see my comments below. -John Bottoms
On 2/3/2014 3:27 PM, Duane Nickull wrote:
inline:
On 2014-02-03 12:15 PM, "John F Sowa" <sowa@xxxxxxxxxxx> wrote:
On 2/3/2014 12:55 PM, Duane Nickull wrote:
XMl has nothing to do with semantics.
XML has the ability to make data portable and can be used to transfer
ontological or semantic models, or fragments thereof, between
applications.
I agree with both of those points.
JSON notation (or the very similar LISP notation) could have been
adopted by the W3C for exactly the same purposes, but without the bloat
or excessive coding. LISP, for example, has a built-in parser that
requires only two operators: CAR and CDR. (AKA head and tail.)
The W3C has recently adopted a methodology whereby they will define a
technology (like RDF), then define bindings using various syntaxes like
JSN or XML. This is the proper way to architect interchange formats IMO
as they are developed from a good model.
Duane Nickull
Yes, I do understand that is what W3C is doing. And no,
I'm not
sure is the right thing to do.
Stacks essentially trace their beginnings back to Dijkstra's
"THE multiprogramming system
(Technische Hogeschool Eindhoven)".
"The constraint that higher layers can only
depend on lower
layers was imposed by the designers in order to make
reasoning about the system
(using quasi-formal methods) more
tractable, and also to
facilitate building and testing the system incrementally.
(THE
multiprogramming system wikipedia entry)
[if you are interested his PhD thesis is
at:
http://www.cs.utexas.edu/users/EWD/PhDthesis/PhDthesis.PDF]
The important kernel of the layer system, which
ultimately
became the ISO standard for communications, is that messages
are passed between
layers and layers do not "dip into" other layers structures.
The only
place I have seen exceptions to this is in tunneling, in
communications, where
priority or time take precedence. When binding between layers
takes place you
are effectively gluing the layers together and that removes a
critical open
architecture feature; namely the ability for any other
component to talk to
that interface. IP solve this problem by having a field that
specifies who
(which protocol) the message should be sent to. This way the
forwarder can send
the message to TCP, Telnet or DLSw.
The W3C approach can be followed but if it is a
static
one-to-one binding then we will lose the flexibility that has
been developed
for so many years. Further, the binding requires compatible
updates between the
layers or something will get broken when the layer
specifications are updated.
The IP approach allowed for the addition of new numeric values
without changes
to the protocols or layers when protocols were modified or new
ones added. All the protocols have standard numeric values
that
are kept in tables so the forwarders can operate across many
different
protocols. Notice that this design contains no bindings. All
forwarding is dynamically determined.
<rant>The more I see of W3C
architectures, the more it looks like an
intentionally arthritic skeleton, designed to prevent usefully
flexible
designs.
I am beginning to think it is time for an web system
designed by scientists, not programmers.</rant>
-John Bottoms
FirstStar Systems
Concord, MA USA
_________________________________________________________________
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