Seadrive that's the way forward I think with the water bath-thanks guys for all your suggestions!The problem is not the temperature of the developer; the problem occurs when the temperature changes during a printing session, so that test prints developed at 20C don't provide useful information, because your developer has now warmed up to 25C.
If the temp in your darkroom is warm but constant, just shorten your development time. You need to know the time at which the darkest values in your print first appear in the developer. At 20C, that may be 20 seconds. If you're developing for two minutes, then your development factor is 6 (20 secs * 6 = 120 secs = 2 mins).
The temp on a particular afternoon is 25C, and your darkest values now appear at 15 seconds. Using your development factor of 6, you reduce your development time from two minutes to 1 1/2 minutes (15 secs * 6 = 90 secs). As long as the temp remains at 25C, you're good to go.
If the temp starts at 20 and rises during your printing session, well, now you've got a problem. You have to either pay close attention to the "first appearance" times, and do the calculation in your head as you develop the print, or you need a way to control the temperature of your developer.
I use a water bath, myself. I place my 8x10 developer tray in a 12x16 tray, and add about a liter of 18C water to the larger tray. Every 20 minutes or so, I add an ice cube or two to the water bath, and every once in a while, I check the temp of the developer.
The best solution, of course, is to find one of Fred Picker's Zone VI Compensating Development Timers, a gizmo that automagically adjusts the time on your timer as the solution temperature changes. They're hard to find, and they're not cheap!
Anyway, hope some of this helps a bit.