WISHLIST (Last updated 05-Oct-2004) -- aka TODO

	Version, Features and Dates are all projected.

0.93: (Apr, 2004)
	* multiple textfiles
		. by groups?
		. by state (all, 'down' status..)
	* pageinterval configurable per host
	* pings per check configurable.  also configurable per-host.
		loss+numbers to send (min_pings, send_pings) (half done)
	* textfile displays objects that are down when the parent is also down
	* config flaptime <minutes>;
		display host in html/text file for next <minutes>
		amount of time as green after host came back up
	* native dns checks (send packet, do we want recurs, aa checks? need to
		parse response) - need addl feats, docs
	* http://example.com:1345/ textfile+auth+object history
	* per-object pmesg (DONE? - needs docs+testing 20030708)
	* snmp custom "down" message
	* snmp custom "up" message
	* queueinterval configurable per-host 
		(DONE? - needs docs+testing 20030708)
	* ability to compare two snmp oids w/ each other and alert if
		equal, not equal, etc..
	* ability to configure descriptions for snmp errors on a per-object
		basis
	* show FILENAME in errors (standardize format?)
		parser_file_error(char *filename, int lineno, "msg");
		or replace msg with integer pointing to canned message
	* documentation
	* need to aggregate snmp queries to a single host/ip in one
		larger snmp message with centralized snmp dispatcher
	* what to do w/ snmp traps? (by 0.94)
	* config savestate /path/to/file (save state for sw upgrades)
	* snpp notification
	* make paused data visible in curses client
	* packetloss check 
		keep history for 24 hours (configurable glob and per-host)
		tolerance config for pings lost in a period
		if total loss, just report as down
		example: far-away-site pktloss 15 1 2 "far-site" mail@site
		will send 15 pings every 1 minute, if more than 2 lost, report
	* rtt check? (use external rtt agent?)
	* SAA?
	* ircd check? (connect, send quit, no banner)
	* documentation
	* config authfile (file-type sysmon|passwd|htpasswd|etc)
	* curses client fix
	* make daemon log port listening on via syslog()
	* revoke root!
	* make WAKEUP_CHECKS go out and actually do that, instead of just 
		be killer...  also go out and make us pass better states 
		around such that wakeup_checks actually only calls checks 
		every few times, instead of every iteration
	* config minnumfailures (but before it appears in txt file, ie: 
		becomes yellow)
	* object suppression.  if one of the dep's is down, don't monitor
		or alert.  useful in circuit based environment if nei rtr
		is down.
	* rework queuer to sort by age and queue longest since check
		instead of current pseudo-linear+random system
	* ability to cap queue size based on check-type
	* All bugs gone :)
