[rules-users] More temporal operator questions

[Wolfgang Laun]

Hi,

I haven't been able to figure out what Q#1 is asking.

As for #2: Yes, you can define your own operators, even with (literal)
parameters. See my home page, http://members.inode.at/w.laun/,
Drools, Guides on Selected Topics.

Concerning #3: There is no need to avoid a logical disjunction of
two constraints within a single pattern. It's only a disjunction of
patterns that results in a rule split, which may cause surprising
results.

-W


On 14/12/2011, Mike Melton <mike.melton at gmail.com> wrote:
> If you'll indulge me, I have a couple more questions on temporal
> operators. I have specific requirements where the built-in operators
> don't *quite* fit. I'm finding workarounds but I don't love them and
> basically I'm looking for validation of my approach or better ideas.
>
> 1. I need an operator that means "includes with endpoints". In other
> words, "A includeswithendpoints B" means A.start <= B.start < B.end <=
> A.end. Another way to look at it is: A "includes or coincides or
> starts or finishedby" B.
>
> I've been messing around with the parameters to includes. Ideally, I'd
> represent this as "includes[0, duration]", where "duration" is A's
> duration. This would mean that the minimum distance between the start
> and end timestamps is 0 and the maximum distance between the start and
> end points is Event A's duration. However, I haven't been able to find
> a way to pass a dynamic parameter to the operator.
>
> My workaround in this case is "includes[0, 1000000d]", where 1000000d
> is just an arbitrarily large number. My tests so far have borne this
> out as working, but I don't like it because it's not very
> self-explanatory; everywhere I use it, I will need to include a
> comment. The beauty of the temporal operators is that they are
> intuitive and I feel like this workaround is compromising that.
>
> 2. Given an interval-based event A and a point-in-time event B, I need
> an operator which means A "finishedby or includes" B; i.e., A.start <
> B.timestamp <= A.end.
>
> Similarly, in this case, the workaround is "includes[1ms, 1000000d, 0,
> 1000000d]". This takes the ugliness and need for explanatory comments
> to a new level.
>
> A few questions:
> 1. Is there any technical reason why my workarounds won't work
> (besides the obvious one where there is an event that lasts 1000001
> days)?
> 2. Are there better solutions? (e.g., can I define my own operators?)
> 3. As a general rule, I avoid using "or" logic in my rules. However,
> is this a case where I should consider it? To me, "EventA( this
> includes $eventB || this finishedby $eventB )" is much cleaner than
> "EventA( this includes[1ms, 1000000d, 0, 1000000d] $eventB )", but
> that doesn't mean I like it.
>
> Thanks in advance for any feedback.
> Mike
> _______________________________________________
> rules-users mailing list
> rules-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/rules-users
>