[Top] [All Lists]

[ontolog-forum] Asynchronous processes -- was NASA-Ontolog OKDMS

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Fri, 02 Nov 2007 08:49:08 -0500
Message-id: <472B2AD4.40801@xxxxxxxxxxx>
Paola,    (01)

The notion of asynchronous process is actually more primitive
than the notion of time, and there is no need to assume any
kind of metric time in order to have asynchronous systems.
Therefore, the notion is independent of any 3D vs. 4D issues.    (02)

Synchronization does depend on time, and it is much harder to
implement.  If you have systems that are separated by a distance
that is large compared to the systems themselves, they will
naturally run asynchronously.    (03)

And large is a relative term.  Inside a computer, different
chips will naturally be asynchronous, and some global clock
is needed if synchronization is important (and for most
purposes, synchronization is less important than causal
dependency).    (04)

Petri nets are one of my favorite representations for
asynchronous processes, and they do not require any notion
of time.  Instead, the nodes of a Petri net interact with
one another by passing messages.    (05)

Since it's impossible for any node to respond to a message
before it receives it, the message-passing mechanism
automatically enforces a kind of cause-effect dependency.
In fact, Petri nets are an excellent representation for
cause and effect.  (The activity diagrams of UML are
simplified Petri nets, and the Pi calculus is a generalization
that allows new nodes and links to sprout dynamically.)    (06)

This is one more reason why I believe it is a serious mistake
to put large numbers of axioms into the upper level ontology.
Metric time is a very important, but derivative notion, which
is only relevant if you have clocks.  Primitive societies had
different notions of time, and even European societies didn't
think in terms of metric time until clocks became popular
during the Renaissance.    (07)

That is why I believe that axioms for time should be put into
separable modules so that they can be introduced as needed
for one application or another.  For reasoning about processes,
it is much more fundamental to avoid time unless absolutely
necessary.    (08)

For more about these topics, see the following web page:    (09)

    Processes and Causality    (010)

Figures 5 and 6 show examples of Petri nets, and Figure 7
shows how to represent a clock in terms of a Petri net.    (011)

Message passing is the usual way of representing causal effects
in computer systems, and the simplest way to think about clocks
is to imagine them as nodes that broadcast messages (AKA time
signals) to other processes.    (012)

3D and 4D ontologies don't become relevant until you introduce
clocks.  And for some kinds of problems, metric time is more
of a distraction than an aid.    (013)

John    (014)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: 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 Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (015)

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