At 08:48 PM 10/19/1999 -0400, Jeffrey C Honig wrote:
>> Agree that the new protocol SHOULD include compatibility with the old
>> protocol for both pitch and catch (or source and sink, client and server,
>> heckyll and jeckyll, etc.).
>
>I think that this is one of the big problems with existing protocol.
>It is not very flexible and any changes have the potential to break
>other vendor's implementations. We considered some enhancements at
>BSDI, but had to abandon them in favor of backwards compatibility.
After thinking it over, I want to withdraw my comment (quoted above) that
back compatibility is important. Let's ignore all the existing
infrastructure, including port assignments, choice of protocol, and
location of the configuration file (getting to the implementation details)
and avoid the burden Jeff describes of having to make the new syslog work
with existing syslog implementations.
Otherwise, we are likely to have to write a specification for the old
syslog protocol before we can write one for the new.
Bob
