[gme-users] various complaints about constraints

[Peter Volgyesi]

I am not sure which default setting is the better (though your sitation
suggests it is the "first violation ends it" default behavior).
 
If you'd like to (pre)set this feature in your model/paradigm file (that you
might distribute to your users), create a registry key in your root folder:
/ConstraintManagerSettings/ViolationCount  and set it to 1.
 
--
peter
 
 


  _____  

From: gme-users-bounces at list.isis.vanderbilt.edu
[mailto:gme-users-bounces at list.isis.vanderbilt.edu] On Behalf Of Matthew J.
Emerson
Sent: Thursday, April 22, 2004 5:54 PM
To: gme-users
Subject: RE: [gme-users] various complaints about constraints



Yes, I have the short-circuit logic operators enabled - this does not make
the "implies" operator a short-circuit operator.  As a work-around, I've
replaced used of the "implies" operator with judicious use of &&, ||, etc.
in my constraint equations.  But I would rather "implies" were a
short-circuit operator itself.

 

As for setting "Evaluation ends after first violation," setting this does
prevent GME from crashing.  But, I can't force the user to set this option.
Since my paradigm won't work properly unless it is set, it would be nice if
I could in some way configure my paradigm to automatically set this option
for the user.  Or, what would be better and easier is if "Evaluation ends
after the first violated priority level" were the default configuration for
GME.  I think that's the setting that makes the most sense as default
anyway.

 

--Matt

 

-----Original Message-----
From: Peter Volgyesi [mailto:peter.volgyesi at vanderbilt.edu] 
Sent: Thursday, April 22, 2004 9:06 AM
To: gme-users
Subject: RE: [gme-users] various complaints about constraints

 

Hi,

 

To the short-circuit evaluation problem: have you tried to enable
short-circuit logical operators? (File/Display Constraints/Settings)

 

Also, make sure that "Evaluation ends after the first violation" is set on
this page. 

 

--

peter

 

 

 


  _____  


From: gme-users-bounces at list.isis.vanderbilt.edu
[mailto:gme-users-bounces at list.isis.vanderbilt.edu] On Behalf Of Matthew J.
Emerson
Sent: Sunday, April 18, 2004 7:16 AM
To: gme-users
Subject: [gme-users] various complaints about constraints

I don't think that the constraint manager really evaluates constraints in
the order of their priority.

 

I have two priority-1 constraints that need to be checked before all others,
because if those two constraints evaluate to false then the other
constraints will crash gme.  Well, no matter what priority I make the other
constraints, gme crashes.  If I delete the lower-priority constraints, gme
does not crash.  If I turn my priority-1 constraints into constraint
functions and then call them as part of the lower-priority constraints as
pre-conditions, then gme does not crash.  But I'm going to have a hard time
finding every case where I need to check these pre-conditions and inserting
the necessary function calls.  I'm going to miss some and gme is going to
soundlessly crash when the constraint manager goes into an infinite loop.

 

I think that if the constraint manager would just check my priority-1
constraints first and then stop checking constraints if they evaluate to
false, my life would be much easier.

 

Also, if the "implies" statement used short-circuit evaluation, it would
make my life easier.  For instance if you say "A implies B", and A is false,
then don't check B.

 

Finally, it would be good if some priority levels beyond level-1 could
result in critical evaluations.  Assuming that the constraint manager
eventually checks constraints in priority order as I described above, it
would be good to be able to specify an evaluation order among critical
constraints.

 

--Matt 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.isis.vanderbilt.edu/pipermail/gme-users/attachments/20040422/20b76d17/attachment.htm