Setting the bias point of a discrete MMIC

According to this http://217.34.103.131/pages/pdfs/an60010.pdf it has to be driven by aconstant current source.
I am trying to make one of these super wideband discrete MMICs.
../imgqa/eboard/Antenna/rf-mkyb0iwetgu.png
../imgqa/eboard/Antenna/rf-c4u2g4cimhk.jpg

By reading at the pdf file, I believe instead of using the bias resistors, an active constant current source would be better?

Could you suggest a simple, 1 or 2 BJTs variable constant current source schematic, so I can sed the current of the discrete mmic?

How about this one? it seems nice!
http://www.seekic.com/blog/project_s...ight_LEDs.html

This attachment has a detailed description of how to implement an active bias circuit.

Interesting, but it uses opamps.
What if I keep the voltage stable, would the current then be stable? I am confused by the term "device voltage" used on the mini circuits file. Is it the same as the PSU voltage or it is something internal to the MMIC?

You also can make a constant current with LDO. There are many claasic circuits in web.

What is an LDO?

LDO: Low drop output.

http://www.vidisonic.com/2008/07/10/...iting-circuit/

- - - Updated - - -

http://www.vidisonic.com/2008/07/10/current-limiting-circuit/

This is exactly what I need.
I am thinking of using 2n2222 transistors to drive the two cascaded amplifiers of this one ../imgqa/eboard/Antenna/rf-3xqjjojwpcx (one pair for each cascaded amp)
What would you recommend as a starting point for the resistor values for each cascaded amplifier?