Neat, but kinda hamfisted. With your approach, you need to write it into your script. So if you launched something without the notifications you're stuck. You need a way to monitor, rather than report.

Monitoring approaches would use a terminal window (or geektool?) doing something like tailing a log file ("tail --pid $PID -f $FILE") -- you are taking logs, aren't you? -- or a "watch pgrep -lf rsync". More effort would be needed to make use of growl.

The fun fact about monitoring, as distinct from reporting, is that it's applicable to other OSs (yup, even the one from Redmond), and gives you an inkling of what server folk do all day.

[ Reply to This | # ]

true enough. I was mostly interested in the persistent message thing. If you want to focus more on monitoring there's hint from a few years ago here that talks about how to set up something more generalized. combining these two might be a fun project; I just didn't bother. :-)

[ Reply to This | # ]