Matching in RFIC and discrete systems

Hi all,
what's the difference of matching between RFIC system and discrete system?
some one said:
"It depends on the operation frequency and dimension of your circuits.
If you are doing IC design for RF, typically the dimension of the metal
trace delivering your RF or LO signal to mixer is much smaller than the
wave length so you may consider it as a voltage signal and directly deliver
to RF/ LO ports. However, if you are doing discrete design or your
operation frequency is high enough, you must treat your circuits in
'microwave manner' and add necessary matching network.

The signal traveling on the interconnection in the form of Power and your
circuits dimension determines the way to treat them."

is this statement true?
is any book I can refred to drill this question.

Best wishes
fxxjssc

If the connection is smaller than the wavelength there are no reflections, but the input needs to be matched to the source impedance that will drive it.

hello vfone,
Thanks. But for integrated PA design, like a two-stage one,
we always do inter-stage matching. why?
Thanks.

For the reason that I mentioned above: the input of the following stage needs to be matched to the source impedance that will drive it.