Things I've Learned as an ECE Grad Student
Weaver's Laws regarding ECE Graduate Work
- Never attempt to use a project developed by grad students in a way
not exactly like their published results. It won't work.
- Corrolary 1: even if you do use it in exactly the same way
as their published results, it still won't work.
- Corrolary 2: It probably didn't work for them either.
- Corrolary 3: Computers don't work at all, we're all delusional.
- It is impossible to write an x86 processor simulator without abusing
the C-preprocessor beyond all recognition.
- If you submit a paper more than 5 minutes before the submission
deadline you probably aren't doing interesting work.
- If you've done all of the homework problems except one, the students
who come to your office hours will ask you about that one question.
- Computer Scientists are incapable of expressing any idea without using
at least one backwards or upside-down letter. And once they do they'll
all claim that what they see is somehow "intuitive".
- All modern computer research is focused on getting a 5% increase in speed
out of the 26 SPEC2000 benchmarks and nothing else.
- Corollary: actually no one bothers with the Fortran or C++
Papers I'd Like to Write Someday:
- Weaver, Vince. "A Novel Cache Coherence Protocol Using Killer Robots"
Abstract: Modern multi-processor computer systems use complex
protocols to keep cachelines coherent. I propose the use of miniature
robots to arbitrate the cache lines. For each potential conflict, the
robots will fight to the death and the winner gets to keep the cache
line in the "Owned" (or PWNED) state. With the further advances of Moore's Law
I project that killer robots will be able to be made exponentially smaller
and thus would support having larger cache-lines with more robots in
the near future. As an optimization the robots can fight until one is
severely injured, rather than destroyed, otherwise you have to replace
your CPU once you run out of robots.
- S Tater, D Bindel, P Murray, V Weaver. "A Particle theory of Procrastination."
Proceedings of the Talbot Hall Cafeteria Conversations, 1998.
Abstract: We propose a particle called the "Procrastinon" that is the
elementary particle responsible for procrastination. When large accumulations
of lazines occur (such as those in the dorm rooms of English majors) a
high procrastinon-flux occurs, and field lines radiate outwards. Unlike
electro-magnetic phenomena, the procrastinon responds in an inverse-linear
manner not inverse-square, causing its effects to have far greater range.
The anti-particle is known as a "Virtuon" and is unfortunately much weaker
Back to musings page
©VMW Web Design ( vince _at_ deater . net) : 10 March 2006