Rainer: 1. I would propose leaving severities as is & with descriptions. I don't think deviating from widely adopted convention gives much of a benefit here. Did somebody ask for a change?
2. I do agree that legacy facilities need to go. Original legacy facilities came out of system-local use (kernel, printer, system daemons, user-level messages, etc). Essentially, it was a set of standard facilities for one specific product - BSD OS. Whereas we are defining a protocol which is much more broadly applicable. 3. I think defining any *standard* facilities (legacy or not) will be very difficult. For example, for one product it may be acceptable to use just one mail facility, while another mail server product may wish to send messages from different subsystems to different facilities. So, facilities can't be required, only suggested anyway. A broader question... What is a facility? And how do we recommend people use it? I did not see anything about this in the -protocol. Is this something that could be more appropriately called a "logging channel"? This removes the vague idea people may have that this is something like a product or a subsystem. It could be any arbitrary grouping of messages. Essentially, it does not have an inherent meaning -- it looks like it is just another tag, only restricted numerics and placed at the front of the message for easy parsing. Thanks, Anton. > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Rainer Gerhards > Sent: Wednesday, January 28, 2004 11:24 AM > To: [EMAIL PROTECTED] > Subject: Facilities and Severity > > > Hi WG, > > again on the edit: I am replacing the few fixed facilities by > a 4-digit number right now. I wonder if it would still make > sense to mention the traditional facilities in -protocol. > Same goes for severity, where we now also allow additional values. > > I would appreciate comments on this. For now, I will remove > any references. This does not imply that I think they should > be removed. I am actually half-half. > > If that is concensus will be that the traditional values > should be mentioned, I'll move them back in either in -02 (if > comments arrive before end of edit) or -03 (if not). > > Thanks, > Rainer > >
