For some reason, people have an enduring like for new schemes foremail where the sender stores the message until the recipient wants toread it (the most well known is D. J. Bernstein's Internet Mail 2000 ). Such schemes tend to handwave the socialproblems involved in a transition, but let's set that aside (alongwith why they won't stop spam ) and talk aboutthe practical technical problems, because from my perspective they arepretty bad.

If the sender stores

(This idea is not from me, it's from R Francis Smith . It is just sufficiently nifty and wrong thatI'm going to write it up for posterity.)

Suppose that you have a generic hostname, a hostname that either ismultiple machines (with multiple IP addresses) or a virtual host thatgets pointed to different physical machines from time to time. Furthersuppose that inside your environment, your users ssh to that machine,or at least want to. The traditional problem with this is that

As I've written about, I have a problem machine . I have recently had the opportunity to stopusing it as my office workstation and thus do some systematic testing onit, which has turned up some interesting yet mysterious results.

(It's remarkable but perhaps not surprising how much my mood has improved fromgetting a stable primary workstations and setting up these tests.)

The short summary is that the machine reliably crashes if I dosignificant disk activity and I am not running something thatburns up CPU.