Brovnik, I agree with you. There have been many discussions about the 'code dreaming' both as nodes, in the CB, in my office, and other places where I meet coders. I think the reason that this 'suprises' people, is that usually we don't assume that our dreams are worthy of 'producing' solutions.
I think this dream utility depends upon how 'tight' you are with your current project. I know it works that way for me. I problem solve in my sleep all the time. It could be code, or construction, it doesn't matter. The common factor is: "How much is this problem on my mind?"
I've only worked with Perl for about two months now. Just yesterday I used DBI.pm for the first time. I read the perldoc and 'used DBI', and had it working in about an hour. I was so proud of myself... I didn't have to post a node to figure it out! :) What I soon realized is that the DBI really is about that darn SQL that I haven't touched for about 2 years (much like you're comment about not utilizing perl for quite a while.) Low and behold, I dreamt of nothing but SQL problems last night. (BTW: not a restfull sleep ;| )
I know there are many stories about this, and I think it is worthy of discussing these 'dream solutions' on occasion. It's a different way of debugging (That's for sure.) I know that in my office we talk of our dream insights all the time. The first few times were kind of 'wierd and funny', but now we really don't think anything of it at all. It's like a bonus resource that appears as a gift in the night.