> - the wrong anwers must be realy wrong, not just "a bit less correct than=
 the correct one".

I'm not sure about this.  It's the sort of requirement that leads to people=
 thinking that multiple choice tests are too easy.  I've taken "sneaky" mul=
tiple choice tests, and while they can be annoying, I don't think I'd consi=
der them to be invalid.  (there was a puzzle that used to (?) be published =
in the NY Times; about 20 sentences where you were supposed to pick the "co=
rrect" word from TWO choices.  There was a big prize if anyone got them all=
 right.  Which hardly ever happened.  The explanations were entertaining, t=
hough.)

I guess it partly it depends on whether the grading process allows for argu=
ment.  If you're going to computerize everything so that you can offer a cl=
ass and test to 10k online students, you'd better have unambiguous answers.=
 If you have an in-person class of 20, you might be willing to offer partia=
l or full credit for a convincing argument supporting what you thought was =
a wrong answer.

Multiple choice questions should not be answered from the point-of-view of =
a argumentative jerk.  If you want to answer the capacitor question with "r=
ipple current is the most important capacitor spec because I once designed =
a SMPS where that was so", then you should get the question wrong.  IMO.

My questions were aimed at the target audience mentioned in the original po=
st.


>> An interrupt request on a microprocessor is usually responded to
>> a) within several instruction times
>> b) within several microseconds
>> c) within several milliseconds
>> d) when the microprocessor is no longer busy

Hmm.  I wanted (a), although I admit that I'd be happier if "microseconds" =
were longer compared to instruction cycles.  This was spurred by a discussi=
on on another forum along the lines of "are interrupts so slow that they're=
 making my clock inaccurate?"  I want people to realize that interrupts occ=
ur between instructions, and that instructions are now frequently significa=
ntly shorter than a microsecond long.  (d) is meant to be VERY wrong - the =
point of an interrupt is to go off and do something even though the process=
or IS doing something else.  If you want to claim "I asumed by 'busy' you m=
eant 'in a critical section with interrupts disabled'"=85  See my comment o=
n argumentative.)  The big problem is that a lot of not-really-obsolete mic=
rocontrollers are pretty close to that "microsecond per instruction" timing=
 (4MHz PICs and classic 8051s, to mention two.)

I'm actually pretty happy that people are arguing over subtleties in my pro=
posed answers, rather than thinking that the questions are poor :-)

BillW


--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.