MOTM

It's a fine point, but op amps have output current limiting built in.  
 From the ST Micro TL072 data sheet:

4. The output may be shorted to ground or to either supply. Temperature 
and/or supply voltages must be limited to ensure that the dissipation rating
is not exceeded.

If you stay within the power supply voltage limits and temperature 
ratings, shorting an output does no harm.  I remember this because I 
once fixed a Serge module that had an op amp output shorted to ground by 
a faulty etch.  The signal from that baby was everywhere, seeming to 
come from every jack!  It was driving the ground up to its current 
limit.  Cutting the faulty etch fixed it.  The op amp was not damaged.

1K series resistors on outputs gives a standard output impedance and 
allows such outputs to be mixed together by patching to a multiple, 
though not many people do this.  The Serge used 330 ohm output 
resistors.  I have no idea why.

An op amp is more likely to be damaged by leaving one of its *inputs* 
unconnected.  Because of the extremely high input impedance, a bit of 
static electricity discharged into that pin could exceed the power 
supply voltage by a thousand volts!

-Richard Brewster

>The theoretical output impedance of an op-amp is 0.  Paul designs his
>modules to have an output impedance of 1K if I remember correctly.  The
>reason is circuit protection.  If the output impedance were very low
>(approaching zero) and you accidentally plugged two outputs together,
>WHAMMO!  You'd have little op-amp fires.  The final stage op-amp of the
>module would try to drive the load which would be 0 ohms, so it would try to
>deliver infinite current.
>