Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Thu, 19 Dec 1996 16:04:02 -0500 (EST)

Rozelle wrote:
> I finally looked closely at the dfanout which is in the EPICS
> distribution.  All of the values (including of course HOPR and LOPR)
> are defined as long integers.  This explains why you get an error
> when you try to set HOPR to .5 from the opi. 
> 
> Does any one else have a copy of a floating point data fanout already?

Why do the dfanout, sequence and select records have to have a data type 
in the first place?  Their fuction is data handling, not processing, so 
they don't need to understand the data at all.

[I'm excluding the high/low/median function of the select record for the 
purposes of this discussion -- we still need a numeric Maximum, Median, 
Minimum record type which should conceptually also include the index 
selection mechanism as well]

It ought to be possible to create type-less versions of these which could 
also work with strings and arrays, the data type of the record value being 
determined by whatever is input (or by an FTYP field).  You lose out on 
the ability to specify deadbands and alarm limits on the select/dfanout 
value, although these could be implemented by adding a suitable output 
record before the dfanout or after the select.

- Andrew
   ___
 .'   `.	Andrew Johnson, Head of Electronics
/ Royal )	Royal Greenwich Observatory
\ Greenwich	Madingley Road, Cambridge, CB3 0EZ
| Observatory	Tel: +44 1223 374823   Fax: 374700
+----------	WWW: http://www.ast.cam.ac.uk/~anj

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System