Date: Tue, 14 Nov 89 11:24:47 -0500
From: rap@OSF.ORG
Message-Id: <8911141624.AA07309@gozer.osf.org.osf.org>
To: argus2@ATHENA.MIT.EDU
Subject: messgae types...

I do not agree that the messge type of a rule should not be in the rule.
That is I still contend that the message type of a rule (i.e. what kind of
message this rule was created to match) should be part of a rule.  Typing,
of course, does not type a rule (as Brian says) but a rule is designed to
match a msg (that is the predicate of a rule) and so should incorpoeate all
aspects of a message: including message type.  

Also, I don't think that it would be appropriate to have some message types
insert the "Message-Type:" field into the predicate of a rule and others not
to.  As for the "MSG" message type, of course it should come up with a field
called "Message-Type" if a user has explicitly added it from within the
ruleeditor.  Remember that no message types have an actual field
"Message-Type:" in their declaration.  As per excluding the insertion of
that field when the user asks for a particular message type (i.e. MSG) I
respond with an example: lets say someone starts off with a phone message
type, deletes all the fields and then adds the MSG message type (this
example also demonstrates why we need the MSG message type in the
ruleeditor's option of message types).  The "Message-Type:" field should
come up.  Also, remember that the "Message-Type:" field is an artifact of
creating a rule, or sending a message (i.e. with alcomp), and does not exist
apriori.  The point is that no particular message type should be excluded
from inserting its identity into a rule because as far as all message types
are concerned they are all equal: MSG is no better than any other message
type.

Rambling yours,

	__Rich.