I've always hated the idea of email bankruptcy, but today I am declaring it.  I am flushing my entire email backlog as of the start of 2012 (midnight GMT).  If you ended the new year awaiting an answer from me, please ask again.

I probably should have done this about 7 years ago -- some of these messages go back that far.  I have to ask myself: why did I keep them in my inbox for so long?  

If you've ever seen my office you know that poor organizational skills could be a factor, but I think there's more to it than that.  I believe that I suffer from a cognitive confusion between thinking about email as a technology to be perfected, and thinking about email as a never-ending stream of stuff to deal with.  I try to perfect my own email correspondence the way I used to try to perfect the reliability of email delivery.  Learning a lesson about the latter didn't seem to help me with the former.

Perfect correspondence is obviously hopeless, but even perfectly reliable delivery was a dumb idea, as I learned a quarter century ago.  It was In 1985, and I was a newly-minted Ph.D.; after four years working on email technology, I was finally doing it full time and getting paid well for it.  It was a great time for me: I was teaming up with my long-time officemate and good friend JR to make it happen. My youngest daughter was newborn, and we had just moved into our first house.   I joined the Andrew Project, at CMU, with grand visions for things like multimedia email on a global scale.  The sky was the limit.

However, first we had to deal with the starting point.  At that time, there were suddenly hundreds -- hundreds! -- of UNIX workstations being deployed throughout the Carnegie Mellon campus.  Each one came out of the box with an email configuration that was designed for a small company's timesharing system.

Email chaos ensued.  All of those machines thought they ran the campus' email, and each had been patched differently to try to fit into the campus system.  The result was that email for any individual could be delivered on any machine, and the user would see it if and only if they logged into that machine.  People would send email, and then phone and ask if their email arrived.

So, if we ever hoped that anyone would take email seriously enough to justify our broader ambitions, our first task was to make it reliable in this new environment.  We worked on that one goal for months, becoming the email delivery police, investigating every report or even hint of lost or misdelivered email.

Then, one day, we were in the office of a faculty member who had reported lost email, asking some questions and looking at her machine.  She finally took pity on us, closed the door to her office, and said, "Look, I got the mail, OK?  It's just that not getting the mail was the best excuse I could come up with for not having responded."

When that happened twice in a row, we declared that our email delivery was reliable enough.  Perfect reliability, we had learned, is not the right goal for a system that is intended to support all the complexities of human communication.  Sometimes, the possibility of an error permits the kind of minor falsehoods without which society would grind to a halt in a paralysis of truthfulness.  Perfection isn't just elusive, it's undesirable.

It's quite a rationalization, I know, but today I'm happy to report that as an email correspondent, I'm not perfect either.  I've gotten so far behind that I've  had to declare bankruptcy and flush all my unanswered mail.  I'm telling myself that there's some sense in which that might be a good thing as well.   Let me know if you can think of what it might be!  (And write back if you're awaiting an answer from me!)