Re: CA protocol documentation.

["Brian McAllister" <brian@hotrats.mit.edu>]

Jeff, Ralph, et. al:

Is it just me, or do you sound just a bit paranoid ?

You seem to suggest that the only use for a protocol spec would be to write
"alternate implementations of the core system libraries".

Suppose we want:

- to know exactly what is going across our networks.

- to be able to put a network analyzer on the wire and configure it to
  display EPICS traffic in some meaningful fashion.

- to know what the current capabilities and limitations of the protocol
  (rather than the API) really are.

- to understand how the protocol works so we can have meaningful opinions
  on how it might be further developed.


Just because I want to have the TCP/IP spec for reference doesn't mean I'm
interested in writing my own protocol stack.

The argument that fully documenting CA requires significant effort that may
not be available given current resources and priorities seems quite
reasonable.
The idea that it is somehow "dangerous" to provide the details (which can
of course be obtained by sufficient study of the code) does not.

----
Brian McAllister                    Controls Programmer/Beam Physicist
mcallister@mit.edu                        MIT-Bates Linear Accelerator
(617) 253-9537                                           Middleton, MA