Experimental Physics and
Industrial Control System

"Liyu, Andrei" <[email protected]> · Wed, 20 Jul 2005 11:42:43 -0400

	I like idea to call CAS directly! 

	I suppose.
	1. "Call CAS directly" is not completely correct. IOC must have
server level. Look - when a record calls Monitor(), server level has
list of clients and checks subscribers for the record. Am I correct?
	2. CA protocol works with memory. In record base there is
wrapper for different type structure. Particle record type fills end of
memory. There is similar splitting in client side. 
	3. CAS is part of server level. Server level has the same are
errlog, archive, alarm, timing, seach engine for IOC records (of any
type - need define only head) and clients.

	Maybe you will be VERY KINDLY and will think about definition of
server level API. And two levels on CA client side. After that Epics
users get freedom
	- to use standard types and records. It will be useful for a lot
of application that haven't any conflict. And you could really define
MIN types for MAX usages!
	- to do what they want for some applications. And raw dataAccess
(by Marty Kramer's definition) is very acceptable answer. Of course,
control system developers will be responsible for correct usage on both
sides in this case.

	You have had code with well structure!!! I am about sure it is
very easy to implement :-)

Have a good day, Andrei.

-----Original Message-----
From: Dalesio, Leo `Bob` [mailto:[email protected]] 
Sent: Wednesday, July 20, 2005 10:47 AM
To: Kasemir, Kay; EPICS Core Talk
Subject: RE: Type descriptor vs. enum

There are use cases on the wiki.

There are two thoughts on this
-- you are told the type

-- you are not told the type

Could we work out the use cases with either approach?
Ben - you handle -- not told the type
Kay - you handle -- told the type
For the first level calls to channel access

Jeff - you handle how these calls tie into the existing library design

We were going to complete the use cases in two weeks - this is week one.

Thanks,
Bob

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System