Experimental Physics and
Industrial Control System

Benjamin Franksen <[email protected]> · Mon, 19 Mar 2001 21:17:40 +0100

Bernd Schoeneburg wrote:
> 
> If readValue () return with ERROR, udf should not be set to FALSE. If the device
> support read routine (called from readValue) returned '2' (don't convert) it should
> have reset udf. So I think moving of this resetting would break (or 'improve')
> existing device support code and is not nessecary. In mbbi-Record it's the same.
> 
> Carl Lionberger wrote:
> 
> > I agree this sounds OK.  I did notice there is a setting udf to false at the end
> > of convert() (after the smoothing).  I think it needs to either be deleted (its
> > set false various places in device support) or also moved to process().  Surely a
> > converted value is not considered more defined than a smoothed but not converted
> > one.

Is there some generally agreed upon rule which of record support or
device support should reset UDF? I remember I came onto this myself. My
impression has been that sometimes people put it in the device support
because it's missing in the record support (where IMHO it belongs).
Didn't we fix this for some of the records a while ago?

Ben

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System