Experimental Physics and
Industrial Control System

[email protected] · Fri, 13 Dec 96 10:53:54 -0500

RE: HIDEOS and EPICS

Jim,

One difficulty with using Hideos is needing a 2nd CPU.  I'm not sure how
well running Hideos under VxWorks is supported -- care to elaborate?

One experience we have is that we are running (Alicia, feel free to chime
in here) a different VxWorks kernel and/or EPICS kernel for the 1 system
we are running with Hideos (successfully, in production use I might add).
Perhaps that is no longer true at APS, but we are not well connected with
these developments, perhaps, which just points out the difficulty of
us using Hideos or DESY's generic device interface.

On a more philosophical note, is it appropriate to insert Hideos (message
passing) to control a simple register based IP ADC?  Can we not come up
with a better abstract driver interface with extremely low overhead for
such devices?  DESY also has a generic interface, but it, too, seems heavy 
weight for simple devices.

I would be estatic if we could merge all of these developments, but it would
require all the relevant developers to stop work on other projects, and focus
exclusively on this problem for a month, plus effort to change some amount
of existing code.  I think other systems have better standards at this
level (LabVIews, National Instruments), and EPICS could stand some real
improvements here.  Takers, anyone?

Chip

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System