It's a matter of what kind of precision you need with regards to what you work with. I became a little frustrated with drift of some less-than-best reference oscillators in some of my test equipment when working at GHz frequencies. I now use a HP Z3816A GPSDO feeding a HP 5087A distribution amplifier whose multiple outputs feed all of the equipment on the bench via their external reference inputs. It's nice when generating or measuring a signal to know that you are basically spot on.
And the reference standard is there if needed to calibrate other oscillators. For that I use either a HP K34-59991A phase comparator, a 3575A gain-phase meter or simply an oscilloscope. The scope and K34 unit is for less demanding applications while the 3575A gives me reference/DUT phase comparisons at 10 MHz down to 1/10 of a degree accuracy.
And, no, I am not a "time nut." My wristwatch runs about 30 seconds fast each month and the old analog clock on the wall keeps me properly on time for everything in my life.
