Experimental Physics and
Industrial Control System

Kay-Uwe Kasemir <[email protected]> · Thu, 20 Oct 2005 14:06:16 -0400

On Oct 20, 2005, at 11:52 , Ralph Lange wrote:

The current plan (afaik) is to allow even more user-defined

"triggers". Or "filters". Or "events". (Everyone of us has

different interpretations of those terms...)

Your last mentioned subscription won't be enough, as e.g. the Alarm

Hander is also interested in the value, but only wants updates when

the severity changes.

The trigger/events and channels/fields/views aspects of a

connection are orthogonal.

Hi Ralph:

Thanks, the example ("use case") of an alarm tool
wanting to subscribe to value & status, but only
needing updates on status change answers my question.

Yes, we wanted more triggers and filters.
And we're starting over at the level of use case collection
to determine which ones are actually needed.

From the 'office' side, support for the V3 and TANGO events
is a requirement, because the control system API used
by the office applications needs to support those as well.
As far as I can tell, that means:

* update on change in value
 with relative and absolute deadbands,
 each as a positive and negative limit.
 E.g. update when value increases by >= 1% or descreases by <=0.5%.
 EPICS V3 only has a single absolute deadband.

* update on change of 'quality'
 Is that like the V3 'alarm'?

* periodic updates with specified period.
 EPICS V3 doesn't have that on a per-user basis.

* update on archive related changes
 which are designated relative and absolute deadbands,
 again pos. & neg.,
 and in addition a minimum update rate.
 EPICS V3 only ADEL, no additional periodic update.

* 'event' would be another trigger, which Tango doesn't have,
 but V3 already supports as single, numeric events from
 post_event as well as device-supported interrupts.

Filters need to be added to all of these.
A machine-mode masks,
where e.g. 1b'0000001 means 'long pulse',
           1b'0000010 means 'to target A',
allows simple and/or masks to filter only on
long pulses to target A.
The use cases will decide what more we need.

Thanks,
-Kay

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System