[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
UML Point/Counterpoint: Mellor vs. Selic
Stephen J. Mellor
Modeling Complex Behavior Simply
http://www.embedded.com/2000/0003/0003feat1.htm
described what he thinks are the problems with UML.
Bran Selic
"How to Simplify Complexity" is by at
http://www.embedded.com/2000/0003/0003feat2.htm
fights back, and Mellor replies again.
In "Misuse of hierarchy", Mellor argues that UML has "non-
transparent" higher levels in the state machine hierarchy.
Mellor says "In a well-formed hierarchy, the larger unit
is transparent: it simply groups state machines together
with no additional semantics." In his Last Word Mellor
says (after reading Selic's convincing arguments)
that "I began to dislike using flat state machines myself,
until I remembered that my argument involved partitioning
into multiple smaller state machines." Selic argues
that UML state machines are possible to understand, and in
fact needed to handle the inherent complexity of a normal
system.
If I have a PAR in occam, the father process freezes, and
I cannot communicate with it. Neither can it communicate
with anybody else. It only connects the lower level state
machines (processes). It is "transparent". However, I could
have been given a channel "down there" to communicate
three levels up and two down. I haven't missed anything
on that point in occam.
Q1. Does this make occam more like Mellor (Structured) or
Selic (UML).
Q2. How does clean CSP handle this. Like occam?
--
(------------------------------------------------------------------(
) Oyvind Teig ) oyvind.teig@xxxxxxxxxxxx ) Tel: +47 )
( Navia Maritime AS ( oyvind.teig@xxxxxxxxxxxx ( 73581268 (
) div. Autronica ) ) Fax: +47 )
( 7005 Trondheim ( http://www.autronica.no ( 73919320 (
) Norway ) ) )
(------------------------------------------------------------------(
) http://www.autronica.no/pub/tech/rd/index.htm )
(------------------------------------------------------------------(