I do not like giving references as part of the candidate assessment process

This message is a cleaner re-write of an earlier assault on the
irksome convention of some employers requiring references as part of
the interview process.

URL to previous discussion thread:

   http://www.mail-archive.com/[email protected]/msg00433.html

WHY REFERENCES SUCK IN GENERAL

1/ They require the time of the reference. 15 minutes per reference
phone call means 2-3 hours out of each references work week during a
typical job search. Your reference (and his boss) will love you for
taking up his time.

2/ The reference and/or his company can get in legal trouble. Oracle,
for example, has a policy that its employees will not give references
to former employees. The extent of referent information is manager
name and time and location of employment. I worked at Oracle. I can't get
a character or technical reference from them. I love the skeptical looks I
get when I try to explain this to people.

3/ They allow a bad boss one more shot at tyranny. If you worked for a
jerk, then what? As an extreme example, assume that Martin Luther King
worked for 3 closet white supremacists. Would he ever get a good
reference? Probably not. But that doesn't mean that he is a bad
person.

4/ People move, retire, die, or go belly up: I worked during the dot-com
era. That's 3 years worth of references that I will never be able to
resource because all of those people have bitten the dust. I guess
that makes me a bad person.

WHY TECHNICAL REFERENCES SUCK

1/ If you want to know how good I am technically, break out the
appropriate technical test.

WHY CHARACTER REFERENCES SUCK

1/ Let's assume that a character reference is a good way of knowing
how good person X is. If this is true, then you really must have
references for the references of person X before you can trust the
references of person X. And you must have references for the
references of the references of the references of person X before you
can know how good the references for person X are.

Logically:

   X good => (references for X good)
  (references for X good) => (references for (references for X) good)
  (references for (references for X) good) =>  
       (references for (references for (references for X)) good)

So we see that we have a recursive and non-terminating algorithm for
determining if X is good.

The self-referential unsolvable nature of the character reference
proof system shows its fallibility.