[LCP]what value does the foo() return?

[Mike and/or Penny Novack]

I second David's answer

> The interviewer could have been probing for how much you knew about 
> what compilers actually do.  Undefined is the correct answer; however 
> many compilers I have met return int values from functions in a CPU 
> register, so for those compilers the return value would be whatever is 
> in that particular register when the function returns.
>
> If he seemed unsatisfied did you try asking him what he was looking 
> for?  "Undefined" is definitely the correct answer to the question you 
> have given, and asking for more information in this way wouldn't 
> necessarily be seen as bad.  The problem with the answer "Undefined" 
> is that it's a one word answer, and interviewers usually want to get 
> to know a bit more about you.  Any idiot can read a book and determine 
> the answer to the question, so this answer doesn't in itself 
> differentiate between an experienced programmer and that idiot.

The point is that any ACTUAL compiler implementation will return 
something which is defined. Many programmers are careless and do not 
worry about using results which are formally undefined but "work" (with 
this compiler, version of compiler, this hardwafre, etc.). Very very 
bad. See, someday there will be a new version of the compiler. The 
guarantee is that CORRECT statements will be compiled correctly (and 
thus behave the same way with any compiler) but all bets are off with 
regard to incorrect statements.

A shop which allows the laxity of not removing bugs which "work" (for 
some wrong reason) is in the position of having to retest every last bit 
of software at every major change. I can't tell you how many times I've 
had to expalin to some programmer that no, there is nothing wrong with 
the new version of the compiler, it isn't under any obligation to 
reproduce the error condiiton behavior of previous versions, that is NOT 
a bug with the compiler but in your program, and it doesn't matter that 
it "worked" just fine in production the last 20 years << and yes, I have 
seen bugs of this sort lurk for that amount of time -- causing great 
amusement when they finally hit and the programmer "winning" the "purple 
wiener" (a "desk ornament" for the programmer causing the last 
production hang to be kept till somebody else's goof causes a hang) is 
now a vice president in charge of Information Systems (a several hundred 
person shop) of a major Fortune 500 financial>>

So a BETTER answer to the question would have been something like "The 
result is undefined in C, but of course any particular compiler version 
will be reurning some particular result. However programs must never be 
allowed to use undefined results because then they may unexpectedly fail 
the first time they are compiled by a different compiler version." . I 
don't know what the THAT interviewer was seeking. I know that had I been 
doing the interview, I would have considered the simple answer 
"undefined" correct but unimaginative, possibly just rote knowledge 
lacking understanding of the implications of the problem, while I would 
have considered an applicant giving the fuller answer as exhibiting 
practical knowledge about where the problem lies with using undefined 
results that "work". .

Michael