Retriggerable VC One-Shot?

[WeAreAs1 at aol.com]

Hello folks,

I am in need of a circuit that can do the following:

*  Create a clean one-shot pulse, the duration of which is under voltage 
control.
*  The pulse "High" amplitude must be constant (5 volts or 10 volts would be 
OK).
*  The pulse length must be continuously variable from around 10ms to about 
10 seconds.
*  The control voltage range will be 0 to 10 volts.
*  The response to control voltage should be as close to linear as possible 
(within reason).  That is to say, if 10 volts gives a ten-second pulse, then 
5 volts should give a 5-second pulse, and 1 volt would yeild 1 second, etc.  
(BTW, it would be OK if the control voltage response is the inverse of this, 
i.e. 0 volts = longest pulse, and 10 volts = shortest pulse)
*  The variable-width pulse circuit is to be triggered by a digital 
positive-going edge.  It will ignore the falling edge of the trigger pulse.
*  It would be best if the one-shot pusle could also be re-triggerable, if a 
new positive-edge trigger pulse comes along before the one-shot has completed 
its full duration.   Kind of like oscillator sync, if you will.

How would you guys approach this?  Would you use a 555 timer in one-shot 
(monostable) mode?  If so, how would you put its duration under voltage 
control?  How about those TTL multivibrator chips?  (74161, I think?)  Are 
they good for this sort of thing, and can their pulse length be voltage 
controlled?  Any other ideas?  Discrete transistor circuits?  Flip-flop 
configured as a monostable?  Binary counter (4040 with its outputs OR'd?) 
clocked by a fast VCO?  XR4151/RC4151 monostable?  The simpler, the better, 
of course.  Without sacrificing timing stability, CV linearity, and waveform 
cleanliness, of course.

The variable pulse will be used as a synthesizer Gate signal, so the edges 
need to be clean and sharp, but need not be surgical-scalpel-sharp, if you 
know what I mean.

I will greatly appreciate any replies, even those that wish to ridicule.

Michael Bacich