From: Yuhai Ma [Y.Ma@Advantest.com]
Sent: Tuesday, October 21, 2003 10:29 PM
To: 'dave_dowding@agilent.com'; Don Organ; ejwahl@agere.com;
mosley@san-jose.tt.slb.com; jim_oreilly@agilent.com;
jose.santiago@philips.com; larry.moran@teradyne.com; dsprague@us.ibm.com
Cc: tom.micek@motorola.com; tonyt@synopsys.com; g.a.maston@ieee.org
Subject: RE: Test Flow Working Group Updated: FlowNode/SubFlow Diagrams
Attach ed

Hi Dave and All,

Thanks, Dave, for updating the diagrams which have very good information.

I am now traveling, so just a thought from traveling...

When an airplane arrives and stops at a gate, every passenger stands up,
takes his/her luggage and prepares for exiting the airplane (like Exit
Actions (7) in our diagrams), and then everyone exits from the same
exit/gate (ExitPath (8)) and leaves for their own routes (making connection,
baggage claim, rental car, and going home, etc.)...

So, with this in mind, I am think we may not need to itemize different Exit
Actions and Exits in our diagrams since there are so many different
situations to consider/analyze, and two or three Exit Actions and Exits may
not be able to describe them all. So I propose to use only one Exit Action
and one Exit instead of two or three in the diagrams. We don't have to worry
about where a flow goes after exiting from a FlowNode, because we know that
it will go to one of FlowNodes or Terminal Points, and the next FlowNode
(making connection, baggage claim, rental car...) or Terminal Point (going
home) will carry on from there. Like at the entrance point (EntryPath (2))
of FlowNode, we don't define/describe where the flow comes from...

Just my thought for your reference. And since I am in travel, I will not be
able to make the conference call tomorrow, 10/22. Sorry about that.


Have a nice meeting,

Yuhai Ma


-----Original Message-----
From: dave_dowding@agilent.com [mailto:dave_dowding@agilent.com]
Sent: Monday, October 20, 2003 4:23 PM
To: don.organ@inovys.com; ejwahl@agere.com; mosley@san-jose.tt.slb.com;
jim_oreilly@agilent.com; jose.santiago@philips.com;
larry.moran@teradyne.com; dsprague@us.ibm.com; Y.Ma@advantest.com
Cc: tom.micek@motorola.com; tonyt@synopsys.com; g.a.maston@ieee.org
Subject: Test Flow Working Group Updated: FlowNode/SubFlow Diagrams
Attach ed


Greetings,
	Though late, I am hopeful that this will cause some e-mail
discussion. The attached diagrams are to help (not hinder) closure
on what we think of the subflow Module per our discussion in last
week's call. Having read the Meeting Notes (thank you Jim 0.) I 
was not sure what we had agreement on. 
	As to the diagrams:
1.	The first page is intended to show corrections agreed upon.
2.	The second page contains Figure 2 which is graphically what  
I think was agreed to as good models for the Module. 2A shows my
view of a pointer to a TestMethod. 2B shows my view of a Module 
that contains a subflow of predefined FlowNodes... one entry and 
one exit. I think the latter as shown is too limited. The FlowNodes
that make up the SubFlow have their Arbitor/ExitActions negated
and are dependent on always having a later Arbiton. I am NOT saying
that this kind of FlowNode arrangement is not needed! I simply don't
think this has the diversity needed. 
3.	But, I needed to back up and determine what kind of (what I 
am calling) "out-flow" FlowNode configurations are agreed upon as 
"legal". By "out-flow" I refer to the different paths we expect are 
needed following the Post-actions of a FlowNode. 
Specifically, I am interested in pre-defined FlowNode reuse... even
in subflows.
I drafted Figure 3A, 3B and 3C to show three kinds of out-flow
configurations that we have discussed. 3A is where a FlowNode has
either one exit path, or the two exit paths go to the same place.
This infers that a later arbitor will resolve the results of this
and subsequent FlowNodes or the same configuration. 
3B shows a pass/fail out-flow configuration where a "terminal point" 
is needed if a fail result occurs. (There may be other kinds of 
"points", but I am trying to keep the discussion simple here.) 
3C is an out-flow configuration where you want to skip subsequent 
FlowNodes and go to a "named" Post-entity... which could be the 
PostAction. This is a FlowNode that is likely to NOT be intended
to ever "stand alone" but be in a sequence of FlowNodes. 
4.	Now if those are all legal, then is the SubFlow Module of Figure 4
a legal Module SubFlow? If it is not altogether correct, then what parts, 
if any are correct?

Comments??? Please,
Dave Dowding