I am explicably tired.  My sleeping has varied wildly over the weekend, and I am looking forward to a good night’s sleep tonight.  Looking back over the day, it has been quite successful.  I browse through Papadimitriou before I go to sleep and it is clear now that Henri was right:  my graph assignment problem is probably NP-complete (though I haven’t proven it yet).  It would not, in any case, be a significant result, even if new.  But the point is that I would to prove it myself.  The same is true of the halting problem for affine automata.  I think I can now prove that halting for affine automata with two or more variables is undecidable.

For the last week there is a computational complexity problem everywhere I turn.  One of my students has a SAT solver which we cannot prove is not polynomial.  At least “in a certain sense”.  I think I’ll ponder that question while falling asleep tonight.  But first I have a little proofreading to do…  That will probably finish me off first.