Sparx Systems Forum

Enterprise Architect => Uml Process => Topic started by: AntonS on July 01, 2010, 06:41:19 pm

Title: Modeling scheduled tasks
Post by: AntonS on July 01, 2010, 06:41:19 pm

Hi

Does someone have a good way of modeling scheduled tasks?
I was thinking of using swim lanes but that only gives me columns and not a grid like Microsoft Outlook Calendar with its grid view of tasks.
How would you model tasks scheduled that run:
•      every 30 min
•      only on weekdays
•      only on the 1st of the month
•      only in business hours
•      starts at 09h15
•      etc.

Any ideas would be appreciated.
Anton

Title: Re: Modeling scheduled tasks
Post by: son-of-sargasso on July 01, 2010, 06:55:40 pm

At the risk of ...

Are you trying to model the "tasks" or the "schedules".  But then again, perhaps you are trying to model that "something happens when something else happens".  Which gets us into a whole realm of "causality" upon which, unfortunately (IMHO) OMG has the following to say:
.


Anyway, good luck.
b

(mod: speellink corrunktions onluy)

Title: Re: Modeling scheduled tasks
Post by: Geert Bellekens on July 01, 2010, 07:27:50 pm

Anton,

As bruce said, you'll have to give more info on which level you are trying to model things.
On a use case level I've used a "Time" actor before to indicate that a certain use case is triggerd by a time event.

Geert

Title: Re: Modeling scheduled tasks
Post by: AntonS on July 01, 2010, 07:59:38 pm

Hi

I am trying to model "something happens when something else happens"

Any suggestions
Anton

Title: Re: Modeling scheduled tasks
Post by: son-of-sargasso on July 01, 2010, 09:23:58 pm

Cool bananas!  So, let's talk about these "something else"s.  I presume that your system knows nowt else but that they "happen".

In other words, the "clock-thing" may go "bong" or "ping" or "tweedley-deee" on (as far as your system is concerned) an entirely unpredictable basis, somewhere between every nanosecond and every millenium (or so).

Hmm, sounds like these "clock things" are "external" perturbations of the status quo, as far as the system is concerned....


Are you with me yet?
b

Title: Re: Modeling scheduled tasks
Post by: AntonS on July 01, 2010, 09:33:27 pm

Yes, the timing is external to the system, it is more like batch jobs that could influence each other and the overall system performance.

Anton

Title: Re: Modeling scheduled tasks
Post by: son-of-sargasso on July 01, 2010, 09:36:52 pm

"batch jobs that could influence each other"
how?
b

starting to smell herrings here...

(where's that canadian state machine expert now that we need him)

Title: Re: Modeling scheduled tasks
Post by: AntonS on July 01, 2010, 09:42:39 pm

One batch job is responsible for inserting data by processing files while the other one wants to finalize the cutoff for a billrun.

Anton

Title: Re: Modeling scheduled tasks
Post by: son-of-sargasso on July 01, 2010, 10:09:38 pm

So this has got nothing to do with whether the clock is saying "tick" or "tock" whatsoever???

Anyway, dad says that I'm getting too excited again and that it's time for bed, so I'll catch you later.

b