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

MIME-Version: 1.0
Server: CERN/3.0
Date: Sunday, 01-Dec-96 18:46:51 GMT
Content-Type: text/html
Content-Length: 52312
Last-Modified: Friday, 26-Aug-94 20:06:15 GMT

<title>Upson's Familiar Quotations (Third Edition) 1974 - 1985</title>

<h1>Al Gaulle, Editor</h1>

<em>September 1982</em>

<!WA0><!WA0><!WA0><!WA0><a href="http://www.cs.cornell.edu/">Cornell Computer Science Department</a>

This report is a compilation of several hundred examples of
context-free language and very irregular expressions.
Contributions were submitted over the past several years
by numerous computer science graduate students who collected
these now immortal words in classes and seminars.  We wish
to express our gratitude to the faculty, guest lecturers,
and students who provided the bulk of this work.

<p>
This work entirely nonsupported by National Science Foundation
grant UFQ-82-256-4934.

<h1>Computer Science</h1>
<h2>Greg Andrews</h2>
<p><em>75 Sep 03:</em>
(On the banker's algorithm:)
There are parallel paths of confusion through it.
<p><em>75 Oct 03:</em>
In six months people figure out how to use their machines until
they're overused.  That's called ``enhancement.''
<p><em>75 Oct 08:</em>
(On the 613 project:)  We've tried to make it as painless as
possible, but it will be painful.
<p><em>75 Oct 20:</em>
(On the midterm:)  All I know is it's a 50 point exam, and McGraw
got a 35 and beat me by 5 points.
<p><em>75 Oct 31:</em>
This is a string of stack snack shops -- uhmm -- snap shots.
<p><em>75 Nov 03:</em>
If the last page of the priority list is three, we take the threeth page.
<p><em>75 Nov 03:</em>
Let's just throw up a few distinct partitions of memory.
<p><em>75 Nov 07:</em>
(On queuing models:)  I'm not going to go through all that crud
-- I don't understand it anyway.
<p><em>75 Dec 05:</em>
Everything has to be written down and not in my head, because
I don't know my head from a hole in the wall.
<p><em>76 Sep 20:</em>
There are two ways to share data.  The first is that you don't.
You either share it or you don't, if you know what I mean.  No, you don't.
<p><em>76 Nov 17:</em>
There's nothing profound here, and that may be true of the whole lecture.
<p><em>76 Nov 19:</em>
This is important, but the stuff I understand is completely irrelevant.
<p><em>77 Sep 05:</em>
\&... then there are idiots like myself who try to hang on.
<p><em>77 Sep 23:</em>
The Dining Philosophers is an interesting problem posed by
Dijkstra.  Many interesting problems with no practical application
were posed by Dijkstra.
<p><em>77 Sep 30:</em>
I'll show you some syntax and you'll go ``yuck'' -- and
that's the state of the art.
<p><em>78 Sep 28:</em>
It would sure be nice to not need another mechanism for every problem.
<p><em>78 Sep 28:</em>
This room is a monitor.  One person is active and everybody else is asleep.
<p><em>78 Oct 20:</em>
I had a nice clean solution in my paper until the referee
told me it wasn't legal.
<p><em>78 Oct 25:</em>
Gries, you're a sweet rat.
<p><em>78 Nov 01:</em>
Student: Excuse me, I've been a little confused lately.

Andrews: Like about 23 years?
<p><em>78 Dec 01:</em>
There's no way I can hope to get done today, so I suspect
I'll dribble over to Monday.
<p><em>78 Dec 06:</em>
It's amazing how much you can do lecturing, when you realize how
little there is.
<p><em>78 Dec 08:</em>
There's a mistake in the notes from the last lecture -- all but
the following line is incorrect.
<h2>Becky Bennett</h2>
<p><em>81 Sep 09:</em>
The computer is getting constipated.
<h2>Corky Cartwright</h2>
<p><em>76 Sep 22:</em>
This is ambiguous, but I'll try to resolve the ambiguity with more prose.
<p><em>77 May 04:</em>
I sort of have the minimal knowledge of hardware you can
have and come out of Stanford.
<p><em>77 Sep 13:</em>
I shouldn't even mention keypunches in connection with LISP.  That's heresy.
<p><em>77 Sep 15:</em>
Can you put my obfuscation into words?
<p><em>77 Sep 20:</em>
I claim that every other possibility is impossible.
<p><em>77 Oct 11:</em>
Is that a question?  I keep hearing these grunts.
<p><em>77 Oct 13:</em>
Let's get on quickly to the handwaving we can do.
<p><em>78 Sep 13:</em>
Computer scientists are so accustomed to reality that,
when they are given the opportunity to create an illusion, they won't.
<p><em>78 Sep 18:</em>
After we solve the reliable software problem, we must solve
the reliable lecture problem.
<p><em>78 Sep 18:</em>
Programming languages should not be subject to whimsical
variations in semantics that depend on the mood of the person
who wrote the compiler.
<p><em>78 Sep 21:</em>
Gries: Why don't you just show us the whole slide?

Corky: Can't quarrel with God.
<p><em>78 Oct 25:</em>
We should come up with programming languages that don't require
a telephone book to describe their little pathologies.
<p><em>78 Nov 06:</em>
Outside this expression language, APL sucks.
<p><em>78 Nov 22:</em>
A program is a comment, a means for getting from a precondition
to what you want.
<p><em>79 Oct 24:</em>
(On SNOBOL:) I'm just someone that looks on from the outside and grimaces.
<p><em>79 Oct 29:</em>
It's hard to say what an elegant SNOBOL program is.
<p><em>79 Oct 29:</em>
This is a real <em>whiz bang</em> program.
<p><em>79 Oct 29:</em>
I would consider that obscene programming, even in SNOBOL.
<p><em>79 Nov 12:</em>
Sound like NA?  Let's get away from it and go on to another topic.
<h2>Bob Constable</h2>
<p><em>75 Oct 15:</em>
If you read the proof in Hopcroft and Ullman, and you'll need to
read it after this ...
<p><em>77 Jan 26:</em>
The only way false can be true is if ...
<p><em>77 Mar 28:</em>
Let me change the theorem to one I can actually prove.
<p><em>78 Feb 17:</em>
Now we only have ten minutes left for hard science.
But that's probably enough ...
<p><em>78 Feb 22:</em>
\&... a really kinky semantics that would support the slickest proof rules ...
<p><em>78 Mar 27:</em>
Student: Number 6 doesn't make sense to me.

Constable: Well, let's write it out bigger.
<p><em>78 Mar 31:</em>
What's going to be amazing is that next time I'll be able to
prove -- that itself will be amazing -- ...
<p><em>78 Apr 23:</em>
Let me back up to where I believed it.
<p><em>78 Apr 25:</em>
This syntactic sugar is looking terrible -- syntactic pepper.
<p><em>78 May 11:</em>
Stanford gobbled up the really good women.
<p><em>79 Feb 09:</em>
We are experimenting with constructive set theory, and so may go wrong.
<p><em>79 Feb 23:</em>
One reason that concept is peculiar is that in mathematics one always
deals with extensional objects, and thus never discusses them,
while in computer science one rarely deals with extensional
objects, and thus never discusses them.
<p><em>80 Jan 28:</em>
This wealth of knowledge will get us confused.
<p><em>80 Jan 30:</em>
We want to lift ourselves out of the primitive recursive muck.
<p><em>80 Feb 08:</em>
(On the Axiom of Choice in constructive mathematics:)
It has the same status as in classical math. You can worry about it.
<p><em>80 Feb 08:</em>
Truth and computation are identical.
<p><em>80 Feb 13:</em>
I don't think I've ever written a RAM program.
<p><em>80 Feb 18:</em>
Sometimes the empty set causes me pause.
<p><em>80 Feb 27:</em>
``To be or not to be'' is not an algorithm.
<p><em>80 Apr 07:</em>
Nothing makes sense in the last five minutes.
<p><em>80 Apr 16:</em>
Let's do a pictorial proof again.
<p><em>80 Oct 13:</em>
We are able to remove logic completely.
<p><em>80 Nov 04:</em>
Student: I've never seen a computer that implements the integers.

Constable: Unfortunately, they're not building them the way they ought to.
<p><em>81 May 15:</em>
Well, it looks like today's lecture
is going to be more confused than normal.
<h2>Dick Conway</h2>
<p><em>75 Feb 10:</em>
PL/19 might have this form, Dijkstra be damned,...
<p><em>75 Sep 02:</em>
This is unreasonable, but I've
been doing it that way for a long time.
<p><em>75 Sep 23:</em>
(On relational data bases:)  I don't know anything about
the mathematical basis of it, but I've read it so many times
I think I can reproduce it verbatim.
<p><em>75 Oct 16:</em>
(On JCL:)  I've always been able to find some grad student
who has been able to understand all this, and I've never
had to learn any of it.
<p><em>75 Oct 21:</em>
This is all done for you by mirrors.  I think it's kind of cute.
<p><em>75 Oct 28:</em>
RPG is a shame.
<p><em>76 Oct 07:</em>
For anybody else that'd be silly, but not for Dijkstra.
<p><em>76 Oct 07:</em>
(To Gries:)  You're cheating, but we still haven't figured out how.
<p><em>77 Sep 13:</em>
This is instantaneous, which is pretty fast.
<p><em>80 May 09:</em>
I am not a chronic dumper on IBM, but here they need it three colors.
<p><em>80 May 12:</em>
I can't quit without telling you about the lunatic fringe of Information
Retrieval.
<p><em>80 May 12:</em>
I wouldn't want to be quoted, but I can cover all the interesting facts
of 635 in \(12-hour.
<p><em>80 Sep 23:</em>
Don't be seduced by a lovely query language.
<p><em>80 Oct 30:</em>
Undo is an abomination.
<p><em>80 Oct 30:</em>
Try that on your old IMS system.
<p><em>80 Nov 11:</em>
Ullman describes fourteen of the basic seven.
<p><em>80 Nov 11:</em>
(On COBOL:) Its procedures are absolutely lousy.
<p><em>80 Nov 13:</em>
(On IMS:) Protects against small children and honest men.
<p><em>80 Nov 20:</em>
(On reading chapter 5 of Ullman:) If you're really masochistic, go ahead.
<h2>Alan Demers</h2>
<p><em>77 Sep 28:</em>
The key to this is that -- is that -- oh, my ...
<p><em>78 Feb 14:</em>
It's not as good a semi-colon as PL/I.
<p><em>78 Mar 07:</em>
Student: Constants do matter in the real world.

Demers: Well, especially since infinity here is 4800.
<p><em>78 Mar 19:</em>
Jesus H. Bald Christ!
<p><em>78 Apr 20:</em>
I don't want to sound like I'm avoiding answering your question;
I just want to avoid answering your question.
<p><em>78 Apr 20:</em>
Jesus!  How do you give a talk like this with Reynolds in the audience?
<p><em>78 Apr 27:</em>
One reason you might want to say, ``Why not?'', or rather one
reason you might want to say, ``Why?'', or actually, a reason you'd say ``not.''
<p><em>78 Sep 13:</em>
Let me suppose, and this is a big assumption -- it gets an asterisk --, ...
<p><em>78 Sep 15:</em>
Student: Isn't that an inspired construction?

Demers: Oh, it is.  When I thought of it I was insufferably
proud of myself for hours.
<p><em>78 Sep 18:</em>
Let me say what I'm not going to do at the moment.
<p><em>78 Sep 28:</em>
After looking at that for some time, the i and the i are
not the same i, are they?
<p><em>78 Sep 29:</em>
Not only is it not obvious, I claim in general it's false.
<p><em>78 Oct 02:</em>
That's in Knuth, volume tree.
<p><em>78 Oct 04:</em>
By the way, if all you people understand this proof it will all
but double the number of people in the world that do.
<p><em>78 Oct 04:</em>
When I do it, it comes out that way.  When Tarjan does it,
it comes out that way.  So proof by intimidation.
<p><em>78 Oct 16:</em>
Let's take this whole section of the board and write a
big ``dubito'' over it.
<p><em>78 Oct 16:</em>
The precise definition is motivated by that bow tie.
<p><em>78 Oct 16:</em>
The reason I can't say that is because it's not true.
<p><em>78 Nov 06:</em>
I sure wish I had a snappy comeback.
<p><em>78 Nov 06:</em>
It's a commutative ring, so I can do a whole lot of cavalier things.
<p><em>78 Nov 08:</em>
It's not in general obvious how to do it, but I was hoping that would slip by.
<p><em>78 Nov 13:</em>
This may be impossible.  At least it's hard.  I don't know how to do it.
<p><em>78 Nov 15:</em>
When I divide b by n what do I get?  Well, I get b over n, obviously.
<p><em>78 Nov 17:</em>
A k-tape non-deterministic Turing machine is this umpteen-tuple.
<p><em>78 Nov 17:</em>
Demers: Here is the promised Fast Fourier Transform example.
It worked for the two examples I tried, so I'm fairly sure it's correct.

Student: Proof by exhaustive testing?

Demers: Well, it sure exhausted me.
<p><em>78 Nov 17:</em>
No, I did not prove it correct.
I sat at a terminal at midnight and fiddled with it until it worked.
<p><em>78 Nov 20:</em>
It depends on something that I really don't want to get into --
it depends on honesty.
<p><em>78 Nov 20:</em>
It just wouldn't do you any good to apply the pumping lemma to the
168 -- it doesn't stay up that long.
<p><em>78 Dec 01:</em>
It spaces me out every time I look at it.
<p><em>78 Dec 06:</em>
I can write the expression down in polynomial space and time --
in fact, I'm going to write it down in about three seconds on the board here.
<p><em>79 Feb 22:</em>
You put handcuffs on yourself and try to swim the English Channel or something.
<p><em>79 Mar 27:</em>
Demers: Why did we switch from call-by-reference to call-by-value,
when we switched from procedures to functions?

Donahue: I don't know.

Demers: I guess that's as good an answer as any.
<p><em>79 Aug 02:</em>
I don't want to remove temptation from malicious people.
I want them to be malicious and then have the pleasure of killing them.
<p><em>80 Jan 21:</em>
You don't solve NP-complete problems in your compiler.
<p><em>80 Jan 23:</em>
I think right linear.
<p><em>80 Feb 06:</em>
PL/I is a disaster, right?
<p><em>80 Feb 20:</em>
The Algol-68 phenomemon is a little bizarre.
<p><em>80 Apr 02:</em>
(After Bowen's annual spring haircut:) That is really far out.
I don't know if I should deliver a eulogy.
<p><em>80 Apr 09:</em>