[rules-dev] xdrl fixes/enhancements

[Mark Proctor]

if you want full 100% xml, you'll first need to define a full expression 
language in XML - as Drools DRL now supports freeform expressions. If 
someone wants to work on that first, we can see how to utilise it's 
namespace inside of patterns.

Jess XML works as it's expr language is simple. Drools 4.0 worked as 
it's expression language is simpl. With drools 5.2 we now allow any free 
form expression, and thus it's not so simple.

Mark
On 11/05/2011 10:33, Pierre wrote:
> Mark, All,
>
> AFAIK, xdrl is used for systems to exchange rules (perhaps for storing too). This, as opposed to human use, for which XML can be quite awkward to say the least. This to say that putting readability as a major factor might make us miss the main point.
>
> For humans<expr>status == state</expr>   is nicer. However, for systems interaction it's pretty terrible. It's some form of half use of XML where the expression itself can only be validated or transformed with considerable efforts (e.g. regex) when xml tags would allow using standard XML tools (e.g. XPATH). I've got examples of this in the rule translation console: http://yieldrif.appspot.com/ (shameless plug!).
>
> Let's have a look at other XML rule descriptions (please note that I don't imply that these are 100% valid expressions in their respective languages):
>
> JessML
>
> <pattern>
> 	<name>Order</name>
> 	<binding>o</binding>
> 	<slot>
> 		<name>total</name>
> 		<test>
> 			<type>eq</type>
> 			<conjunction>and</conjunction>
> 			<value type='VARIABLE'>__synth0</value>
> 		</test>
> 		<test>
> 			<type>eq</type>
> 			<conjunction>and</conjunction>
> 			<funcall>
> 				<name>&gt;</name>
> 				<value type='VARIABLE'>__synth0</value>
> 				<value type='INTEGER'>100</value>
> 			</funcall>
> 		</test>
> 	</slot>
> </pattern>
>
> RIF-PRD
> <formula>
> 	<name>Order</name>
> 	<formula>
> 		<Equal>
> 			<left>
> 				<Var>total</Var>
> 			</left>
> 			<right>
> 				<Var>-var-__synth2</Var>
> 			</right>
> 		</Equal>
> 		<Equal>
> 			<left>
> 				<Var>total</Var>
> 			</left>
> 			<right>
> 				<Expr>
> 					<op>
> 						<Const type="http://www.w3.org/2007/rif#iri">http://www.w3.org/2007/rif-builtin-predicate#numeric-greater-than</Const>
> 					</op>
> 					<args ordered="yes">
> 						<Var>-var-__synth2</Var>
> 						<Var>100</Var>
> 					</args>
> 				</Expr>
> 			</right>
> 		</Equal>
> 	</formula>
> </formula>
>
> I hope this helps,
> Pierre
>
> ------------------------------
>
> Message: 6
> Date: Tue, 10 May 2011 09:11:24 +0100
> From: Mark Proctor<mproctor at codehaus.org>
> Subject: Re: [rules-dev] xdrl fixes/enhancements
> To:rules-dev at lists.jboss.org
> Message-ID:<4DC8F32C.7080205 at codehaus.org>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> On 10/05/2011 09:07, Mark Proctor wrote:
>
>>>   On 10/05/2011 08:10, Veit Guna wrote:
>>>>>   Hi.
>>>>>
>>>>>   Any chance that this could be reviewed and maybe integrated into the
>>>>>   release?
>>>>>
>>>>>   https://jira.jboss.org/browse/JBRULES-2672
>>>   I'm in two minds on this. Edson and I are thinking of changing the xml
>>>   parser to just be:
>>>   <Pattern object-type="xxxx">
>>>   <expr>......</expr>
>>>   <expr>......</expr>
>>>   <expr>......</expr>
>>>   </Pattern>
>>>
>>>   With any valid  expression allwed in expr, i.e. any combinatino of ||,
>>>   &&, method calls, +, - etc. Which follows what we've done with free-form
>>>   drl.
>>>   Our current approach to normalising expressions in XMl isn't working too
>>>   well, and the more we extropolate that the worse it gets:
>>>   <dro:pattern object-type="String">
>>>   <dro:and-constraint-connective>
>>>   <dro:field-constraint field-name="this">
>>>   <dro:literal-restriction evaluator="!=" value="null"/>
>>>   </dro:field-constraint>
>>>   <dro:field-constraint field-name="this.toLowerCase">
>>>   <dro:literal-restriction evaluator="==" value="true"/>
>>>   </dro:field-constraint>
>>>   </dro:and-constraint-connective>
>>>   <dro:from>
>>>   <dro:expression>fctv_17263.getHeaderValues("X-My-Header");
>>>   </dro:expression>
>>>   </dro:from>
>>>   </dro:pattern>
>>>
>>>   But I do recognise that you patch might give a lifeline to existing XML
>>>   users....
> I should add that I ripped out most of the handler validation
> parent/peer stuff already. This was necessary to get the new compiler
> working, and<expr>   actually already works, see DumperTest.
> <lhs>
> <forall>
> <pattern object-type="State">
> <field-binding field-name="state" identifier="state" />
> </pattern>
>
> <pattern object-type="Person">
> <expr>
>                   status == state
> </expr>
> <field-binding field-name="likes" identifier="likes" />
> </pattern>
> </forall>
>
> </lhs>
>
> Mark
>
>>>   Mark
>>>
>>>>>   _______________________________________________
>>>>>   rules-dev mailing list
>>>>>   rules-dev at lists.jboss.org
>>>>>   https://lists.jboss.org/mailman/listinfo/rules-dev
>>>>>
>>>>>
>>>   _______________________________________________
>>>   rules-dev mailing list
>>>   rules-dev at lists.jboss.org
>>>   https://lists.jboss.org/mailman/listinfo/rules-dev
>>>
>>>
> _______________________________________________
> rules-dev mailing list
> rules-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-dev
>
>