but also tricks, which were found by trial and error, which worked in CERTAIN cases. But this does these 'tricks' not necessarily make good design rules...

Ever so often we see something that tries to 'prove' that two wrongs make a right. One example: if the RS-232 port on a PC does not adhere to the standard (for cost saving reasons) then a Mickey Mouse circuit can make a uC communicate with it, but when connected to another PC it will fail miserably. Another example would be the case Kai discuss above where the mistake of a case connected to a dirty ground is 'cured' by RF isolating the case.

Whatever anyone says two wrongs do not make a right

Those that use such approaches may do so at their own peril, I could not care less, but to publisize it as "design techniques" makes my hair stand on end. I guess Prahad found the ferrite to case in some web page, but you can also find a web page that state which motel Elvis is currently staying at.

One recommendatrion to all: do not believe webpages where postulates are not discussed.

Erik