That was me, however I was not trying to solve the "uninitialized value used in concatenation" problem. I used that as an example of something that was able to tell if it had been set before or not. And indeed, defined did the same thing, I had assumed (incorrectly) that if I had done my $a = undef, it would have thought the value was 'initialized' and not generate that error. I chock this up to the use of uninitialized instead of undefiend in the warning message.

I was trying to create a behavior that depended on if someone had tried to assing a value, even undef, to a variable. So 'my $a' and 'my $a = undef' would result in different behavior.

This is of course not possible, but I did not know that, and I was very confused. Also this was in the IRC channel, and it was very hard to define my question with enough information to prevent people from assuming I was trying to do something that could be solved with the 'defined' check.

How could I have handled this better? Was it just a communication breakdown cause I was trying to do something impossible that closely resembled something that was possible? Or did I just choose a really bad example?

Well, you had specifically ruled out the correct answer in your question. I think the XY problem still applies. But I don't think you can handle it better than you did. There were some grumpy people around and that's part of what the IRC is about. It took less than a screen to get everything sorted out and in my mind it went pretty smoothly for being on the IRCs. I really didn't think you did anything wrong, aside from confusing us initially.

-Paul

Hmm - so that error message suggests that there is a difference between uninitialized and undefined. This sure is misleading.