> Modify programs like netstat(1) to use SNMP (or just grab the freely
   > available tools and use them, where possible).

   Ick. I wouldn't mind having netstat use sysctl(), but I think that
   going for SNMP is REALLY BAD here.

So it could use the local interface if a host name isn't specified.
This is a detail.

   > 4) With everything accessible via SNMP, machines can be managed
   > remotely.

   [...]

   I would argue that rather than thinking in SNMP terms, we should take
   the part that is good -- the idea of having sysctl() support returning
   enough information that clients can walk its MIBs extensibly and
   cleanly -- and put lots of other stuff in sysctl(). This has all the
   advantages you cite and none of the SNMP cruft.

No; in fact, it has only one of the advantages I mentioned.  Notably:

1) It would *not* be similar to an existing standard.

3) We *would* have to write wrappers for everything made available
through snmpd (which is the main reason snmpd doesn't exist in any
particularly useful form).

4) Machines cannot be managed remotely, unless we also invent a new
network protocol.

5) Doing anything based on string matching will bloat the kernel code
significantly.

SNMP will clearly do the job here, and is more flexible and simpler
than any of the alternatives I've seen suggested.  Why should we
invent a new, incompatible protocol, when an existing one works just
fine?  That's completely illogical.