Re: [rules-users] Facts, context, rules and queries
For the simple example you have provided, both forward and backward reasoning work equally well. Distinction sets in if there are more reaoning steps invilved. I do not quite understand why you need to have two fact databases. There is no reason I know of to prevent you from implementing an expert system for anything, entirely in either Drools or Jess or any other production rule system; and using Prolog would also be an option. The choice might be influenced by the data model you need. If your data is best represented as (Java) Objects, then Drools is in favour. OTOH, some ontology representations might be handled better with logic programming. -W On Wed, Apr 14, 2010 at 10:59 PM, Patrick Sannes patr...@i-dev.net wrote: Hi Im working on a project for which I need a toolset. From my study I worked with prolog and with jess. So I have a basic understanding of those tools. I want to create a factbase about a scenario. Those facts are learned. I need to query those facts. In the meantime a contextual factbase is created. The facts in this contextual factbase can influence the other factbase based on rules that are also learned. So for example, I like meat and vegetables, but on sunday I like only vegetables. So when i ask to the system, what do I like, it should tell me on all days meat and vegetables, but on sunday it should say only vegetables. The thing is that I have the feeling that the type of queries i like to make are brilliant to do in prolog, but the rules part is a drools type of problem. Is there anyway to do this in one environment? Or maybe Im not thinking in the right way. Patrick ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Facts or rules ?
Hello,
Thank you for your answer and experience feedback.
Your explanations are very clear.
Regards,
Benoît
Date: Fri, 9 Apr 2010 09:59:28 -0700
From: greg_bar...@yahoo.com
To: rules-users@lists.jboss.org
Subject: Re: [rules-users] Facts or rules ?
It matters on the nature of the processing. Rules are useful when the
algorithm involves iterative change, i.e. a rule fires, changing the data,
which cause another rule to fire, changing the dataetc until a resting
state has been reached.
They're also useful if you want to handle odd cases, or global constraints,
especially if these can directly affect the data driven processing. This is
what I was talking about when I said you can do both below. Let's say your
data driven approach involves generating the pricing rules from the data.
(Instead of having generalized rules that act on inserted objects.) Using this
approach you (or the business folk) could then define other rules that execute
alongside the generated ones for both conditions that don't follow the pattern
of the generated rules, and global conditions
that affect all price calculation. This is like mixing decision tables and
hand written rules together, a rather common practice.
That make sense?
Including a BRMS is warranted when 1) there's going to be change, and 2) funky
processing is anticipated. In my experience both cases are frequent. The
marketing folks always come by and say, can your system handle this promotion
we want to do? :)
--- On Fri, 4/9/10, Bertrand Grottier b_grott...@hotmail.com wrote:
From: Bertrand Grottier b_grott...@hotmail.com
Subject: Re: [rules-users] Facts or rules ?
To: rules-users@lists.jboss.org
Date: Friday, April 9, 2010, 10:29 AM
Hello
I am coming back to this post with new a piece of information regarding the
number of rules: this latter may attain 5 * 10^5... This figure confirms the a
data driven approach.
However, if I represent business rules using facts, is there really an added
value in using a brms ? (compared to a hard-coded component).
Thank you in advance for yours answers.
Regards,
Benoît
Date: Wed, 7 Apr 2010 07:44:09 -0700
From: greg_bar...@yahoo.com
To: rules-users@lists.jboss.org
Subject: Re: [rules-users] Facts or rules ?
The answer to whether this is a good approach is, It depends. :)
Does the 10^4 and 10^5 figure mean there could be that many rules? If so,
definitely do the data driven approach.
If not, will business folk (i.e. non technical) will be
writing pricing rules? If so, do the rules based approach as data driven
rules are generally less understandable by business types. (They're a bit more
abstract and influenced by the structure of your data model. Business people
should not have to deal with either abstraction or the implementation details
of a data model.)
There's no reason why you can't do both, though.
--- On Wed, 4/7/10, Bertrand Grottier b_grott...@hotmail.com wrote:
From:
Bertrand Grottier b_grott...@hotmail.com
Subject: [rules-users] Facts or rules ?
To: rules-users@lists.jboss.org
Date: Wednesday, April 7, 2010, 9:23 AM
Hello,
I am a newbie with business rules systems.
I would like to implement a component that compute the price of products.
There can be a great number of potential prices (maybe 10^4 or 10^5).
My first idea was to create template rules, feeded by a database. In other
words, the information about prices would be contained in the rules.
Eventually, I decided to create Prices objects, that are also inserted in the
Knowledge session (Prices become facts). The rule engine just matches the
appropriate price object with each product (there are other rules then to
compute the final price).
Price
{
_productId ;
_beginDate ;
_endDate ;
_price ;
}
Product
{
_id ;
_date ;
...
_
}
Are my explanations clear enough ?
Is this a good approach or is it stupid ? Why ? What are the good practices ?
Thank you in advance for your help.
Benoît
Envie de naviguer sur Internet sans laisser de trace? La solution avec Internet
Explorer 8
-Inline Attachment Follows-
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
Hotmail débarque sur votre téléphone ! Paramétrez Hotmail sur votre téléphone!
Gratuit !
-Inline Attachment Follows-
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
_
Hotmail arrive sur votre téléphone ! Compatible Iphone, Windows Phone,
Blackberry, …
http://www.messengersurvotremobile.com/?d=Hotmail___
rules-users mailing list
rules-users@lists.jboss.org
Re: [rules-users] Facts or rules ?
It matters on the nature of the processing. Rules are useful when the
algorithm involves iterative change, i.e. a rule fires, changing the data,
which cause another rule to fire, changing the dataetc until a resting
state has been reached.
They're also useful if you want to handle odd cases, or global constraints,
especially if these can directly affect the data driven processing. This is
what I was talking about when I said you can do both below. Let's say your
data driven approach involves generating the pricing rules from the data.
(Instead of having generalized rules that act on inserted objects.) Using this
approach you (or the business folk) could then define other rules that execute
alongside the generated ones for both conditions that don't follow the pattern
of the generated rules, and global conditions that affect all price
calculation. This is like mixing decision tables and hand written rules
together, a rather common practice.
That make sense?
Including a BRMS is warranted when 1) there's going to be change, and 2) funky
processing is anticipated. In my experience both cases are frequent. The
marketing folks always come by and say, can your system handle this promotion
we want to do? :)
--- On Fri, 4/9/10, Bertrand Grottier b_grott...@hotmail.com wrote:
From: Bertrand Grottier b_grott...@hotmail.com
Subject: Re: [rules-users] Facts or rules ?
To: rules-users@lists.jboss.org
Date: Friday, April 9, 2010, 10:29 AM
Hello
I am coming back to this post with new a piece of information regarding the
number of rules: this latter may attain 5 * 10^5... This figure confirms the a
data driven approach.
However, if I represent business rules using facts, is there really an added
value in using a brms ? (compared to a hard-coded component).
Thank you in advance for yours answers.
Regards,
Benoît
Date: Wed, 7 Apr 2010 07:44:09 -0700
From: greg_bar...@yahoo.com
To: rules-users@lists.jboss.org
Subject: Re: [rules-users] Facts or rules ?
The answer to whether this is a good approach is, It depends. :)
Does the 10^4 and 10^5 figure mean there could be that many rules? If so,
definitely do the data driven approach.
If not, will business folk (i.e. non technical) will be writing pricing rules?
If so, do the rules based approach as data driven rules are generally less
understandable by business types. (They're a bit more abstract and influenced
by the structure of your data model. Business people should not have to deal
with either abstraction or the implementation details of a data model.)
There's no reason why you can't do both, though.
--- On Wed, 4/7/10, Bertrand Grottier b_grott...@hotmail.com wrote:
From:
Bertrand Grottier b_grott...@hotmail.com
Subject: [rules-users] Facts or rules ?
To: rules-users@lists.jboss.org
Date: Wednesday, April 7, 2010, 9:23 AM
#yiv69327948 .ExternalClass #ecxyiv1286396095 .ecxhmmessage P
{padding:0px;}
#yiv69327948 .ExternalClass #ecxyiv1286396095 .ecxhmmessage
{font-size:10pt;font-family:Verdana;}
Hello,
I am a newbie with business rules systems.
I would like to implement a component that compute the price of products.
There can be a great number of potential prices (maybe 10^4 or 10^5).
My first idea was to create template rules, feeded by a database. In other
words, the information about prices would be contained in the rules.
Eventually, I decided to create Prices objects, that are also inserted in the
Knowledge session (Prices become facts). The rule engine just matches the
appropriate price object with each product (there are other rules then to
compute the final price).
Price
{
_productId ;
_beginDate ;
_endDate ;
_price ;
}
Product
{
_id ;
_date ;
...
_
}
Are my explanations clear enough ?
Is this a good approach or is it stupid ? Why ? What are the good practices ?
Thank you in advance for your help.
Benoît
Envie de naviguer sur Internet sans laisser de trace? La solution avec Internet
Explorer 8
-Inline Attachment Follows-
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
Hotmail débarque sur votre téléphone ! Paramétrez Hotmail sur votre téléphone!
Gratuit !
-Inline Attachment Follows-
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Facts or rules ?
Hello,
I am a newbie with business rules systems.
I would like to implement a component that compute the price of products.
There can be a great number of potential prices (maybe 10^4 or 10^5).
My first idea was to create template rules, feeded by a database. In other
words, the information about prices would be contained in the rules.
Eventually, I decided to create Prices objects, that are also inserted in the
Knowledge session (Prices become facts). The rule engine just matches the
appropriate price object with each product (there are other rules then to
compute the final price).
Price
{
_productId ;
_beginDate ;
_endDate ;
_price ;
}
Product
{
_id ;
_date ;
...
_
}
Are my explanations clear enough ?
Is this a good approach or is it stupid ? Why ? What are the good practices ?
Thank you in advance for your help.
Benoît
_
Découvrez comment SURFER DISCRETEMENT sur un site de rencontres !
http://clk.atdmt.com/FRM/go/206608211/direct/01/___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
Re: [rules-users] Facts or rules ?
The answer to whether this is a good approach is, It depends. :)
Does the 10^4 and 10^5 figure mean there could be that many rules? If so,
definitely do the data driven approach.
If not, will business folk (i.e. non technical) will be writing pricing rules?
If so, do the rules based approach as data driven rules are generally less
understandable by business types. (They're a bit more abstract and influenced
by the structure of your data model. Business people should not have to deal
with either abstraction or the implementation details of a data model.)
There's no reason why you can't do both, though.
--- On Wed, 4/7/10, Bertrand Grottier b_grott...@hotmail.com wrote:
From: Bertrand Grottier b_grott...@hotmail.com
Subject: [rules-users] Facts or rules ?
To: rules-users@lists.jboss.org
Date: Wednesday, April 7, 2010, 9:23 AM
Hello,
I am a newbie with business rules systems.
I would like to implement a component that compute the price of products.
There can be a great number of potential prices (maybe 10^4 or 10^5).
My first idea was to create template rules, feeded by a database. In other
words, the information about prices would be contained in the rules.
Eventually, I decided to create Prices objects, that are also inserted in the
Knowledge session (Prices become facts). The rule engine just matches the
appropriate price object with each product (there are other rules then to
compute the final price).
Price
{
_productId ;
_beginDate ;
_endDate ;
_price ;
}
Product
{
_id ;
_date ;
...
_
}
Are my explanations clear enough ?
Is this a good approach or is it stupid ? Why ? What are the good practices ?
Thank you in advance for your help.
Benoît
Envie de naviguer sur Internet sans laisser de trace? La solution avec Internet
Explorer 8
-Inline Attachment Follows-
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
___
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users
[rules-users] Facts or rules ?
Date: Wed, 7 Apr 2010 07:44:09 -0700 From: greg_bar...@yahoo.com To: rules-users@lists.jboss.org Subject: Re: [rules-users] Facts or rules ? The answer to whether this is a good approach is, It depends. :) As usual :-) Does the 10^4 and 10^5 figure mean there could be that many rules? If so, definitely do the data driven approach. These figures will be refined soon but yes, there could be that many rules. If not, will business folk (i.e. non technical) will be writing pricing rules? If so, do the rules based approach as data driven rules are generally less understandable by business types. (They're a bit more abstract and influenced by the structure of your data model. Business people should not have to deal with either abstraction or the implementation details of a data model.) There's no reason why you can't do both, though. Ok. _ Consultez gratuitement vos emails Orange, Gmail, Free, ... directement dans HOTMAIL ! http://www.windowslive.fr/hotmail/agregation/___ rules-users mailing list rules-users@lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users
