An auto-focus rule changes the module focus right at the moment the
rule is activated, which happens during pattern matching -- i.e.,
right at the "assert" call below. Focus would have been shifted to
MAIN by that assert, if MAIN weren't already the focus module, which
it is since (reset) moves the focus to MAIN. But then right after the
assert, you explicitly shift the focus to the MOD module, which
effectively undoes what the auto-focus would have accomplished, anyway,
On Jan 26, 2008, at 11:57 PM, D Moon wrote:
(deftemplate item (slot id))
(deftemplate lost (slot id))
(deftemplate counted (slot id))
(set-current-module MAIN)
(defrule lose-an-item
(declare (auto-focus TRUE)(salience 100))
(item (id ?c))
(not (lost (id ?c)))
=>
(assert (lost (id ?c)))
(return)
)
(defmodule MOD)
(defrule count-item
(declare (salience -100))
(item (id ?id))
(not (counted (id ?id)))
=>
(assert (counted (id ?id)))
)
(watch all)
(reset)
(assert (item (id 1)))
(focus MOD)
(run)
I'm expecting the MAIN rule lose-an-item to fire first, because the
rule is declared auto-focus TRUE, but my expectations are off.
Here's what happens:
Jess> (batch main.clp)
==> Focus MAIN
==> f-0 (MAIN::initial-fact)
==> f-1 (MAIN::item (id 1))
==> Activation: MAIN::lose-an-item : f-1,
==> Activation: MOD::count-item : f-1,
<== Focus MAIN
==> Focus MOD
FIRE 1 MOD::count-item f-1,
==> f-2 (MAIN::counted (id 1))
<== Focus MOD
==> Focus MAIN
FIRE 2 MAIN::lose-an-item f-1,
==> f-3 (MAIN::lost (id 1))
<== Focus MAIN
==> Focus MAIN
2
I see that both rules properly activate upon assertion of the
matching (item) fact, but I want the MAIN rule to fire first, so I
gave it salience 100 (and gave the MOD rule salience -100 to see if
that would make difference). Why doesn't MAIN auto-focus when I
(run), regardless that I have explicitly given MOD module focus?
One other experiment I tried was to put an (agenda) call
immediately before and after (focus MOD) to see what's what:
Jess> (batch main.clp)
==> Focus MAIN
==> f-0 (MAIN::initial-fact)
==> f-1 (MAIN::item (id 1))
==> Activation: MAIN::lose-an-item : f-1,
==> Activation: MOD::count-item : f-1,
[Activation: MOD::count-item f-1, ; time=2 ; salience=-100]
For a total of 1 activations in module MOD.
<== Focus MAIN
==> Focus MOD
[Activation: MOD::count-item f-1, ; time=2 ; salience=-100]
For a total of 1 activations in module MOD.
FIRE 1 MOD::count-item f-1,
==> f-2 (MAIN::counted (id 1))
<== Focus MOD
==> Focus MAIN
FIRE 2 MAIN::lose-an-item f-1,
==> f-3 (MAIN::lost (id 1))
<== Focus MAIN
==> Focus MAIN
2
So the watch activations shows MAIN::lose-an-item activating upon
assertion of the matching (item) fact, yet (agenda) reports only
the MOD rule. Seems like something happened to the activation of
MAIN::lose-an-item between the time I asserted the (item) fact and
the time that I executed the (agenda) function. Is there something
subtle here that I am missing? Please advise even if it's obvious
and not subtle. ;-\
[Aside: I recall a discussion about auto-focus rules and the
(return) function on this list a while back, but removing (return)
from my lose-an-item rule gives no difference other than the final
==> Focus MAIN not occurring. With (return) in, I tried this:
(reset)
(assert (item (id 1)))
(printout t (get-focus-stack) crlf)
(agenda)
(focus MOD)
(printout t (get-focus-stack) crlf)
(agenda)
(run)
(printout t (get-focus-stack) crlf)
and saw:
Jess> (batch main.clp)
==> Focus MAIN
==> f-0 (MAIN::initial-fact)
==> f-1 (MAIN::item (id 1))
==> Activation: MAIN::lose-an-item : f-1,
==> Activation: MOD::count-item : f-1,
(MAIN)
[Activation: MOD::count-item f-1, ; time=2 ; salience=-100]
For a total of 1 activations in module MOD.
<== Focus MAIN
==> Focus MOD
(MOD MAIN)
[Activation: MOD::count-item f-1, ; time=2 ; salience=-100]
For a total of 1 activations in module MOD.
FIRE 1 MOD::count-item f-1,
==> f-2 (MAIN::counted (id 1))
<== Focus MOD
==> Focus MAIN
FIRE 2 MAIN::lose-an-item f-1,
==> f-3 (MAIN::lost (id 1))
<== Focus MAIN
==> Focus MAIN
()
Note how it appears that the MAIN module refocuses after (run), but
then my final call to (get-focus-stack) shows the empty stack. I
don't know why, don't know if it's related to my above-described
problem, but mention this parenthetically just in case.]
V/R,
Moon
---------------------------------------------------------
Ernest Friedman-Hill
Informatics & Decision Sciences Phone: (925) 294-2154
Sandia National Labs FAX: (925) 294-2234
PO Box 969, MS 9012 [EMAIL PROTECTED]
Livermore, CA 94550 http://www.jessrules.com
--------------------------------------------------------------------
To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]'
in the BODY of a message to [EMAIL PROTECTED], NOT to the list
(use your own address!) List problems? Notify [EMAIL PROTECTED]
--------------------------------------------------------------------
