Re: JESS: Time-varying methods inside a and in a test
(and) does a short-circuited evaluation of its operands, but it is still a single CE. Whether it returns false due to the first or second operand does not matter: it'll have to be reevaluated over and over again. Using two boolean functions in two different CEs may result in less invocations of these functions since a true result of the first CE is not lost. Wolfgang On 18 July 2011 17:00, Skeptic . skeptic2...@hotmail.com wrote: Hi, Is there a good reason why a rule like this : (defrule r1 [...] (test (and (time-varying call1) (time-varying call2)) = [...]) behave differently than one like this : (defrule r1b [...] (test (time-varying call1)) (test (time-varying call2)) = [...]) Thanks !
RE: JESS: Time-varying methods inside a and in a test
r1 is fired, r1b is not (which is the desired behavior). From: ejfr...@sandia.gov To: jess-users@sandia.gov Subject: Re: JESS: Time-varying methods inside a and in a test Date: Mon, 18 Jul 2011 13:06:56 -0400 Although they do compile into different networks, the effect should be basically the same. What's the difference you're observing? On Jul 18, 2011, at 11:00 AM, Skeptic . wrote: Hi, Is there a good reason why a rule like this : (defrule r1 [...] (test (and (time-varying call1) (time-varying call2)) = [...]) behave differently than one like this : (defrule r1b [...] (test (time-varying call1)) (test (time-varying call2)) = [...]) Thanks ! - Ernest Friedman-Hill Informatics Decision Sciences Phone: (925) 294-2154 Sandia National Labs PO Box 969, MS 9012ejfr...@sandia.gov Livermore, CA 94550 http://www.jessrules.com To unsubscribe, send the words 'unsubscribe jess-users y...@address.com' in the BODY of a message to majord...@sandia.gov, NOT to the list (use your own address!) List problems? Notify owner-jess-us...@sandia.gov
Re: JESS: Time-varying methods inside a and in a test
It would be interesting to see these functions. Basically I'm inclined to believe that this is not possible without some weird side effect achieved by one of the functions. -W On 18 July 2011 19:42, Skeptic . skeptic2...@hotmail.com wrote: r1 is fired, r1b is not (which is the desired behavior). From: ejfr...@sandia.gov To: jess-users@sandia.gov Subject: Re: JESS: Time-varying methods inside a and in a test Date: Mon, 18 Jul 2011 13:06:56 -0400 Although they do compile into different networks, the effect should be basically the same. What's the difference you're observing? On Jul 18, 2011, at 11:00 AM, Skeptic . wrote: Hi, Is there a good reason why a rule like this : (defrule r1 [...] (test (and (time-varying call1) (time-varying call2)) = [...]) behave differently than one like this : (defrule r1b [...] (test (time-varying call1)) (test (time-varying call2)) = [...]) Thanks ! - Ernest Friedman-Hill Informatics Decision Sciences Phone: (925) 294-2154 Sandia National Labs PO Box 969, MS 9012ejfr...@sandia.gov Livermore, CA 94550 http://www.jessrules.com To unsubscribe, send the words 'unsubscribe jess-users y...@address.com' in the BODY of a message to majord...@sandia.gov, NOT to the list (use your own address!) List problems? Notify owner-jess-us...@sandia.gov
JESS: Basic questions about modules, context, and focus
I¹ve read through the documentation, JESS in Action book, and mailing list archive, and I can¹t seem to figure out this basic question: How do I assert facts from Java into different modules (or ??contexts??)? To be more specific, the application is a discrete event simulation with agents interacting with each other and objects in the environment. Each agent currently has two separate RETE engines, each with it¹s own CLP file and only the MAIN module, and I use call-on-engine¹ function calls to coordinate between them. While this functions correctly, it¹s cumbersome and inefficient. Plus I need to expand the categories of rules by two and using the current scheme would get too complicated. Most of the discrete event simulation happens in Java, with facts asserted to the agents so they can react to the environment and each other. Currently, this is how I assert facts: engine.assertFact(new Fact(found-something-interesting, engine)); I would like to move to an architecture where each agent has only one RETE engine, one CLP file with multiple modules -- one module for each category of rules. But how do I assert the fact in a specific module? Is it in the string associated with the fact? I found classdoc for ³public Fact assertFact(Fact f, Context c) throws JessException -- Assert a fact, using the given execution context. ...², and then tried to understand of execution context¹ was the same thing as module or focus stack or what. The classdoc for Context was very unhelpful in this regard. Also, what¹s the difference between set-current-module¹ and focus¹ functions? When do you use one or the other? I presume I¹ll be using java methods engine.eval(³(focus BLARG)²);¹ if I need to change the current module, but then what is set-current-module¹ for? I really wish there was a complete worked example showing how to use a rule set with multiple modules that are accessed and manipulated through Java. I need the ³dots² to be more ³connected². Struggling, Russ Thomas
JESS: RE: Basic questions about modules, context, and focus
Hi Russ, Facts are assigned to modules at the template level, not at the individual fact level. When you define a template using deftemplate, either the template specifies its module by using a name like modulename::factname, or if the name is unqualified, the template is placed in the current module. The current module is the one that unqualified constructs will be assigned to; usually it's only relevant during the time your code is being loaded, and not while your program is running. The focus module, on the other hand, is the one from which rules are being fired. It's unrelated to which module things will be defined in. It is relevant at runtime only. If you use assertFact with an unqualified name as you've shown below, Jess will first look in the current module and the default module for a template with that name; if one isn't found, it'll be created in the current module. As the current module won't necessarily be well-defined or predicatable while your program is running, it is always a good idea when using modules to specify the full name of the fact, including the module name, so that you don't inadvertently create templates with the same name in different modules by mistake. Finally, note that when a rule or query is compiled, the unqualified template names it mentions are resolved in the module in which the rule or query is defined, or in the default module. The current module is thus effectively set to the rule's module while a rule is being defined. Once the rule is compiled, the templates are already resolved, and the current module has no effect on how the rule operates. From: owner-jess-us...@sandia.gov [mailto:owner-jess-us...@sandia.gov] On Behalf Of Russell Thomas Sent: Tuesday, July 19, 2011 10:00 AM To: jess-users Subject: JESS: Basic questions about modules, context, and focus I've read through the documentation, JESS in Action book, and mailing list archive, and I can't seem to figure out this basic question: How do I assert facts from Java into different modules (or ??contexts??)? To be more specific, the application is a discrete event simulation with agents interacting with each other and objects in the environment. Each agent currently has two separate RETE engines, each with it's own CLP file and only the MAIN module, and I use 'call-on-engine' function calls to coordinate between them. While this functions correctly, it's cumbersome and inefficient. Plus I need to expand the categories of rules by two and using the current scheme would get too complicated. Most of the discrete event simulation happens in Java, with facts asserted to the agents so they can react to the environment and each other. Currently, this is how I assert facts: engine.assertFact(new Fact(found-something-interesting, engine)); I would like to move to an architecture where each agent has only one RETE engine, one CLP file with multiple modules -- one module for each category of rules. But how do I assert the fact in a specific module? Is it in the string associated with the fact? I found classdoc for public Fact assertFact(Fact f, Context c) throws JessException -- Assert a fact, using the given execution context. ..., and then tried to understand of 'execution context' was the same thing as module or focus stack or what. The classdoc for Context was very unhelpful in this regard. Also, what's the difference between 'set-current-module' and 'focus' functions? When do you use one or the other? I presume I'll be using java methods 'engine.eval((focus BLARG));' if I need to change the current module, but then what is 'set-current-module' for? I really wish there was a complete worked example showing how to use a rule set with multiple modules that are accessed and manipulated through Java. I need the dots to be more connected. Struggling, Russ Thomas
