Why atpg fault coverage is always 100%

[Idatarm]

Episoln \approx 0
Episoln *100/episoln \approx 0*100/episoln
Now lim rousing->0+
100 =0
Coverage =0+0=0+100=100

What do you think?

 

[dick_freebird]

I don't recall ever seeing 100% test coverage, although it's
been a very long time since I cared.

How that formula is anything but theoretical, I am not seeing.
Looks to me like you can make epsilon anything you want,
until you start dealing with reality and just how acceptable
infinite iterations and run time are, when paid by the hour.