Maybe you are thinking about an ordinary transistor saturation and linear.
A Mosfet is the opposite:
1) It is turned on hard like a switch when it is a linear resistor.
2) It is an amplifier with plenty of drain to source voltage when it is saturated.
In the "linear" region, the Mosfet is a linear voltage-controlled resistor with fixed transconductance.
In the "saturation" region the Mosfet transconductance is not linear.
There are many articles about it in Google, here is one: **broken link removed**
In the "linear" region, the Mosfet is a linear voltage-controlled resistor with fixed transconductance.
In the "saturation" region the Mosfet transconductance is not linear.
There are many articles about it in Google, here is one: **broken link removed**
* For sufficiently small VDS voltages (below the saturation region) the FET can be used as a voltage controlled resistor, however it is not a "linear" resistor. The transconductance is not a relevant parameter.
* Transconductance gm is defined as dId/dVgs. This relation cannot be identified in the Id=f(Vds) set of curves. Instead, you need the slope of the Id=f(Vgs) characteristics.
This characteristic has a quadratic form - and, hence, the slope (and the transconductance gm) is not constant but depends on the bias Vgs.
I have used as Jfet (not a Mosfet) as a voltage-controlled resistor for stabilizing the output level of a low distortion Wien bridge oscillator. It is linear (very low distortion) only when Vds is much less than only 100mV.