http:^^www.cs.cornell.edu^info^misc^usr-public^ufq^4th-edition.html

来自「This data set contains WWW-pages collect」· HTML 代码 · 共 1,413 行 · 第 1/4 页

Even my chalk refuses to write this.
<p><em>79 Nov 14:</em>
Now the amusing claim (I'll make several)...
<p><em>79 Nov 16:</em>
Be brave, take the complement of the language.
<p><em>80 Feb 21:</em>
Perverse means nice.
<p><em>80 Feb 28:</em>
I could show you my scars . . . psychological scars.
<p><em>80 Feb 28:</em>
(On the nonconstructive gaps in the gap theorem:)
Ergo, Constable can't see them.
<p><em>80 Feb 28:</em>
That's not the only thing he can't see.
<p><em>81 Mar 13:</em>
This department is run on the model of the Austrian empire: benevolent
despotism moderated by incompetence.
<p><em>81 Oct 15:</em>
Gary Levin is the guy David Gries tries to imitate.
<p><em>82 Mar 01:</em>
While you're being interrupted, let me just comment ... .
<p><em>82 Mar 01 :</em>
The recursion theorem is just like tennis.
Unless you're exposed to it at age five, you'll never become world class.
<p><em>82 Jun 28:</em>
The books I should be writing are written by John Hopcroft.
<p><em>82 Jun 28:</em>
This `A' is skinny, as you see, so it is a sparse set.
<p><em>82 Jun 28:</em>
If the polynomial hierarchy collapses, there'll be panic at some
institutions in Washington.
<p><em>82 Jun 28:</em>
Somebody once asked John Hopcroft about the problem of P and NP. He answered:
``On Tuesdays,
I try to prove that they are equal, on the rest of the week - that
they are different.''
I believe that he has reduced the time to try to show that
they are equal to Sunday afternoons.
<p><em>82 Jun 28:</em>
I strongly suspect that this set is complete.  I have not written yet a proof
because none occurred to me.  No, the correct statement is: ``I have not written
yet a proof because none occurred to my students.''
<h2>John Hopcroft</h2>
<p><em>74 Sep 09:</em>
(About the first 60 pages of Aho, Hopcroft and Ullman:)
It's kind of leisure reading.
<p><em>74 Oct 11:</em>
Everything in this class is a power of two.
<p><em>74 Oct 11:</em>
I'd have to think about it; I'm not in the mood to think.
<p><em>74 Oct 11:</em>
Not everything I say is correct.
It's correct modulo the little details you're going to have to worry about.
<p><em>75 Sep 10:</em>
Somebody should warn me when we get to the end of class by walking out,
or I'll tend to keep on talking.
<p><em>75 Nov 10:</em>
I have an n floating around, which I guess I haven't mentioned today.
<p><em>75 Nov 10:</em>
Without loss of generality, for a reason that temporarily eludes me, ...
<p><em>77 Sep 07:</em>
One thing I like to ignore is details.
<p><em>77 Sep 07:</em>
Two sets of half size -- \(14 being approximately \(12.
<p><em>77 Nov 14:</em>
Student: That means zero to the zero is zero.

Hopcroft: Observations like that lead me to
believe that you understand what I'm doing.
<p><em>77 Nov 21:</em>
I suppose none of these algorithms are really useful.
<p><em>79 Jan 24:</em>
Whenever it doesn't make sense, stop me,
because there's a good chance it's wrong.
<p><em>79 Mar 14:</em>
(After messing up a proof by forgetting how to use induction:)
I should never do things carefully.
<p><em>79 Mar 16:</em>
I'd hate to leave you for a vacation without showing you
how to prove that a problem is NP-complete.
<p><em>79 Mar 28:</em>
I'm going to call that a picture proof.
<p><em>79 Apr 18:</em>
This is obvious to an educated person.
<p><em>79 Apr 20:</em>
I'll have you know I spent 13 hours preparing this lecture
-- so nobody had better sleep.
<p><em>79 Apr 25:</em>
Let @ L sub 0 @ be your favorite NP-complete problem ...
<p><em>79 Sep 24:</em>
Does everyone see the intuitive idea before we get into the details?
Does anyone <em>want</em> to see the details?
<p><em>79 Oct 19:</em>
How about a proof by example?
<p><em>79 Oct 19:</em>
Now we can prove problems right and left to be NP-complete.
<p><em>79 Oct 26:</em>
Let me show you how easy it is to show things P-space complete.
<p><em>79 Oct 26:</em>
If me can't move, you win.
<p><em>79 Nov 30:</em>
You can cube it if you don't like squaring it.
<p><em>80 Oct 17:</em>
Do you mind if my prime is a power of two?
<p><em>81 Oct 27:</em>
We have to figure out how to let interesting things happen.
<p><em>82 Apr 23:</em>
(On the Hopcroft & Ullman book)
I can't make sense of what's in the book.
<h2>Debbie Joseph</h2>
<p><em>82 Feb 17:</em>
Joseph: If you go to MIT, at some point you should try to lose the audience.

Hartmanis: If you don't, they feel insulted.
<h2>Daniel Leivant</h2>
<p><em>80 Feb 13:</em>
You can grasp infinity easily, but how can you grasp @ 2 sup 45 @?
<p><em>80 Mar 26:</em>
If you have a cycle, you can spread truth anywhere.
<h2>Frank Luk</h2>
<p><em>78 Sep 04:</em>
Of course we do not wait until <em>i</em> goes to infinity.
<p><em>78 Oct 02:</em>
The accuracy should not be doubted, because numerical analysis
sends people to the moon.
<p><em>78 Oct 06:</em>
I won't give you a sketch of the proof because the sketch wouldn't be sketchy.
<p><em>78 Oct 06:</em>
The proof of this is so simple that I'm going to give it.
<p><em>78 Oct 09:</em>
It seems we have transformed a difficult problem into an impossible one.
<p><em>78 Oct 11:</em>
So zero is strictly less than one.
<p><em>78 Nov 08:</em>
If we do not know @ f(x) @, its @ (n+1) sup st @ derivative is
even harder to come by.
<p><em>79 Jan 24:</em>
Usually, of course, you stop after a finite number of steps.
<p><em>79 Feb 16:</em>
This is a very vague argument, but it's valid in practice.
<p><em>79 Nov 16:</em>
You have seen more norms than you want to.
<h2>Gerry Salton</h2>
<p><em>75 Feb 13:</em>
It's obviously obvious to you.
<p><em>75 Mar 13:</em>
Incidently, I was Aiken's last Ph.D. student, if you can believe that.
<p><em>76 Sep 23:</em>
Now what we were taught to do at Harvard ...
<p><em>77 Mar 17:</em>
From where I sit, and it's total ignorance I'm sure,...
<p><em>78 Sep 07:</em>
Business people, by definition, are not so smart.
<p><em>78 Sep 28:</em>
(After a loud yawn from a student:)
I feel the same way; I do this as a duty to the undergraduates in the class.
<p><em>78 Sep 28:</em>
(After 68 minutes of a 75 minute class:)
I haven't told you yet anything of interest.
<p><em>78 Oct 17:</em>
Student: Would you like to help grade the 211 exams?

Salton: No. I probably am not qualified.
<p><em>78 Oct 24:</em>
In principle, you need to know nothing.
In practice, you need to know something, but not much.
<p><em>78 Oct 26:</em>
I could say that this is a non-subject, but then you would wonder
why anybody was teaching it.
<p><em>80 Sep 25:</em>
So I shift by either 2 or 2, so of course I shift by 2, and I
have matched bananas.
<p><em>80 Sep 25:</em>
Here is an algorithm that is really useful for a change.
<p><em>80 Sep 25:</em>
We have come full circle -- we are back in the 1950's, which
makes me very happy.
<h2>Fred Schneider</h2>
<p><em>79 Sep 05:</em>
Programs don't execute, at least not in 613.
<p><em>79 Sep 14:</em>
If you never have to use a machine but you need one to
give pathological examples, then the 360 is my favorite machine.
<p><em>79 Oct 10:</em>
(While discussing the Dining Philosophers:)
This is known as the Law of Conservation of Forks.
<p><em>79 Nov 02:</em>
We can talk about reality later.
<p><em>79 Nov 12:</em>
I think Modula's about as close as you're going to get to
reality in this course.
<p><em>79 Dec 07:</em>
Let's warm up with deadlock.
<p><em>80 Sep 06:</em>
Proof by Toyota -- you asked for it, you got it.
<p><em>80 Sep 15:</em>
(On aliasing:) We've got enough problems without worrying about tricky naming.
<p><em>80 Sep 15:</em>
(In a reprimanding tone:) You're thinking again!
<p><em>80 Sep 25:</em>
I'm going to be a hacker when I grow up.
<p><em>80 Oct 09:</em>
Prove your programs by means of Heisenberg's Uncertainty Principle.
<p><em>80 Oct 24:</em>
If you can build semaphores, you can conquer the world.
<p><em>80 Nov 03:</em>
This is my dirty joke of the day.
<p><em>80 Nov 05:</em>
A rendez-vous monitor, a.k.a. a bordello...
<p><em>80 Nov 07:</em>
God knows you forget half the things I tell you anyway. <em>Forget that</em>.
<p><em>80 Nov 19:</em>
Truth . . . is impervious to hisses.
<p><em>80 Nov 21:</em>
(On distributed processing:) It's a very sexy topic these days.
<p><em>80 Dec 03:</em>
<em>Nice</em> in the sense it's pathological.
<p><em>81 Feb 09:</em>
At ten o'clock you can laugh at the jokes.
<p><em>81 Feb 23:</em>
(On chess:)
Look, there's a lot of klugey rules.
<p><em>81 Mar 13:</em>
Forks and monitors are really just the same thing.
<p><em>81 Mar 13:</em>
You could always kill a philosopher.
<p><em>81 Mar 23:</em>
Philosophers are degenerate forms of octopodes.
<p><em>81 Apr 29:</em>
(On the next instruction after a disk head seek to track 1,000,000)
You have to do a CALL(service repair person).
<p><em>81 May 11:</em>
There are those who don't want to be impersonated by Dave Wright.
<p><em>81 May 11:</em>
Some systems ask you personal questions.
Not <em>that</em> kind!
<p><em>81 May 11:</em>
There are just some things in the world you can't do.
That's what 481 and 482 are about.
<p><em>82 Feb 03:</em>
Euclid was a different disaster.
<p><em>82 Apr 05:</em>
I've led a sheltered life.
<h2>Tim Teitelbaum</h2>
<p><em>74 Apr 17:</em>
Everything I say is plus or minus one.
<p><em>76 Jan 05:</em>
The last ten percent of your grade will be determined by other
considerations, mainly sexual.
<p><em>77 Mar 28:</em>
Why am I so goddamned lost?
<p><em>78 Jun 23:</em>
I have a story to tell you which, besides from the fact
that it has nothing to do with communication, is completely irrelevant.
<p><em>78 Dec 05:</em>
Student: Tim, there was a blatant case of cheating on the program.
What should we do?

Tim: What did she look like?
<p><em>81 Feb 26:</em>
I'm running out of time, so I'm going to use very short variable names.
<h2>Charlie Van Loan</h2>
<p><em>76 Sep 06:</em>
It's a trick.  That's what separates NA from ordinary mathematics.
<p><em>76 Sep 15:</em>
3/5 is a calm number.
<p><em>76 Oct 04:</em>
@ A sub ij @ is not @ A sub ij @.
Don't quote me on that -- it sounds like a madman!
<p><em>76 Nov 15:</em>
I have never presented this without getting it backwards.
My solution this year is to leave it as a homework exercise.
<p><em>79 Oct 11:</em>
In n-space you have a unit baseball. Under |A| you have a unit football.
<p><em>80 Sep 19:</em>
This is the algorithm we are going to adulterate.
<p><em>80 Sep 19:</em>
\&...a delinquent set of matrices.
<p><em>80 Sep 22:</em>
Positive definite matrices are beautiful.
<p><em>80 Sep 24:</em>
Band solvers are dynamite.
<p><em>80 Oct 06:</em>
I hate least squares.
<p><em>80 Oct 10:</em>
Householder did his work in the fifties - he was one of the founders of
Numerical Algebra.  He is now retired in Malibu.  This will happen one
day to me too:  the ``Van-Loan Matrix'' has two columns at
the top and three at the bottom.  For that they'll let me retire in
Elmira.
<p><em>80 Oct 17:</em>
If you can find a quadratic equation with three roots, come find me at my
office hours, or even on my estate.
<p><em>80 Oct 22:</em>
It's hard to differentiate FORTRAN subroutines.
<p><em>80 Oct 22:</em>
This is crude, but, um, I'm that sort of guy.
<p><em>80 Oct 31:</em>
Here's the theorem we are going to prove:  it is fairly painless and the result
is quite dramatic.
<p><em>80 Oct 31:</em>
If you're a FORTRAN dude like me, . . . .
<p><em>80 Oct 31:</em>
Write your matrix on a
@ roman { m o dotdot bius } @ strip.
<p><em>80 Nov 03:</em>
So now whenever one mentions how smooth a spline is at a cocktail, you can
prove it on the back of a napkin.
<p><em>80 Nov 12:</em>
I think too much about matrices.
<p><em>80 Nov 14:</em>
If you start thinking about reals, you get nervous.
<p><em>80 Nov 17:</em>
Back row people are very hard to satisfy.
<p><em>80 Nov 19:</em>
I am anxious to give you an example where Newton's method <em>really</em>
works - a little result I proved a couple of years ago in my thesis.
<p><em>80 Nov 19:</em>
How the hell did I get in these things?
<p><em>80 Nov 24:</em>
Perhaps these are electron energies, and you know if it's 10 miles away it's > 0
and if it's on the palm of your hand it's < 0.
<p><em>80 Nov 26:</em>
Let's have a post-algorithm chit-chat.
<p><em>80 Nov 26:</em>
There exist three Newton brothers: Isaac, Quasi and Fig.  The latter had the
most impact on my life.
<h2>John Williams</h2>
<p><em>75 Nov 24:</em>
I might be dumb, but I'm not that dumb.
<p><em>76 Oct 07:</em>
There's apparently a lot of ad hoc-ery going on here.
<p><em>77 Sep 19:</em>
That is clearly -- no, opaquely -- a call to a machine language subroutine.
<p><em>77 Sep 29:</em>
Guarded commands are like flying buttresses.
<p><em>77 Sep 29:</em>
There's something that should bother you in there.
There's something that bothers me -- that nothing bothers you.
<p><em>79 Nov 15:</em>
Functional programming is not NewSpeak. It is possible to utter ugly things.
<p><em>81 Oct 23:</em>
And last year when I talked about it,
Gries was opening his mail.
<h1>Mathematics</h1>
<h2>Kenneth Brown</h2>
<p><em>76 Oct 08:</em>
The easiest way to assert that something is true is to
state that it is clearly true.
<h2>Anil Nerode</h2>
<p><em>79 Oct 25:</em>
I'm always one off, because I can never remember where I am.
<p><em>79 Nov 15:</em>
I'm going mad.