Date: Tue, 05 Nov 1996 21:15:09 GMT
Server: NCSA/1.5
Content-type: text/html
Last-modified: Thu, 12 Sep 1996 23:37:55 GMT
Content-length: 8170
Andy "Krazy" Glew's Favorite Quotes and Sayings
Andy "Krazy" Glew's
Favorite Quotes and Sayings
This collection started out as a set of IBM computer cards pasted
to the wall of McGill's McConnell Engineering Computer Facility,
where I operated the RJE site and provided programming advice,
was pasted to the glass wall of my office at Gould,
and has followed me around.
Summus non Americae, eh?
-- The Canadian National Motto
BUGS: it's the PIPELINE, stupid!!!
-- Andy Glew
-
Expressing my observation that many bugs in P6 were caused by incorrect
design of one dimensional pipelines, not the multi-pipeline decoupled nature
of P6.
-
The title of a talk I gave at the IEEE Asilomar Conference in 1994.
-
We need more formal methods for the straightforward design of pipelines.
Bypasses - just say no.
-- Andy Glew
- An Architect
- A Micro-Architect
-- Robert Geva
(You will not be able to see this without HTML 3, supporting
the "BIG" and "SMALL" HTML commands.)
Measure twice, cut once.
-- "This Old House"
Virtual - something that isn't there, but appears to be.
Transparent - something that is there, but doesn't appear to be.
"In times of profound change, the learners inherit the earth, while the
learned find themselves beautifully equipped to deal with a world that no
longer exists."
--- Al Rogers, Global SchoolHouse Network, speaking at CUE
- Glew comment:
-
I wonder if there is some correlation between "liberal"
and "ability to learn and adapt to new things",
so that liberalism is a trait that has a survival advantage
in times of rapid change,
but which is of less advantage during the long intervals
of less change.
The computer industry may act as an example:
profoundly liberal when it started,
conservative now that the time of great innovations and new ideas has passed.
The 3 most important facts in programming:
- There are N numbers in the set 0..N - not!.
---paraphrased from Stan Kelly-Bootle
- Any programming problem can be solved by adding a level of indirection.
---Unknown
- Any performance problem can be solved by removing a level of indirection.
---Mike Haertel
---Compiled by Mike Haertel.
Q: What is the single most important fact in Computer Science?
A: The fact that there are N+1 elements in the set {0..N}
---Stan Kelly-Bootle
Names are a scarce resource.
---Andy Glew
If it ain't broke, find something else to fix
-- Tim Olson, AMD
The worst maxim around is
"Don't fix it if it ain't broke".
You'd better be improving it,
because your competitor is.
---Wayne Calloway, Pepsico CEO
---FORTUNE April 10, 1989
All great ideas eventually degenerate to hardware
---Peter Drucker?
(The poor ideas remain in software)
Computer memory requirements are growing by 1 address bit / 18 months
---IBM, 1987
If your users don't cost more than the appropriate computers,
you should get better users
(unless they are grad students)
---andy@polya.stanford.edu
(Justification for running emacs on Cray)
ALWAYS code to be reused.
Because then your code can be reused.
Saying that "this code will never be reused"
and coding that way is a self fulfilling prophecy.
---Andy Glew
There has never been a good product designed by a committee.
Keep a decade behind.
I'm all for simplicity.
If it's very complicated I can't understandit.
Parity is for farmers.
---Seymour Cray
Now that I have taught you how to avoid division,
the next time I will teach you how to avoid multiplication.
---Seymour Cray, talking to numerical analysts.
Let's make smart machines to run slower software faster!
---Bob Campbell, uw-beaver!tikal!bobc
-
Flon's axiom:
-
There does not now, nor will there ever,
exist a programming language in which it is the least bit hard
to write bad programs.
Development is maintenance.
---Brian E. Marick
"Do you really want to do this?"
is a question that should never be asked
unless there is a chance that the answer is "No".
Otherwise, if you are asked this question 9 times a day,
answering "Yes",
the 10th time will have you answering out of habit.
-
The 10th axiom of Systemantics
-
When a fail-safe system fails,
it fails by failing to fail safe.
---John Gall, Systemantics
Scientific applications grow a factor of 100 every 10 years
---Giles
Perfection is the enemy of the dissertation
---Ivan Sutherland
Perfection is the enemy of completion.
---Andy Glew
If all else fails, follow the instructions.
-
No! No! No!
Read the instructions, bearing in mind
they're probably fucked up anyway!
---Andy Glew
Indirection is the right direction.
---Andy Glew
Never put off until run time what can be done at compile time.
-
You wouldn't believe the shit I got from late binding advocates
the first time I posted this to a Gould newsgroup!
---Andy Glew.
Never put off until execution time what can be done at decode time.
---Andy Glew
Recurses! Called again...
Recurse you, Mad Baron!
It's easier to apologize than to ask permission.
---Grace Hopper
We need a coalition to stamp out smart I/O devices.
---Chris Torek
Most "smart" I/O controllers are, in fact, "smart-ass" I/O controllers.
---Rob Pike (paraphrased)
An educated man knows of a synonym for every concept.
-
The Titanic Effect
-
The severity with which a system fails
is proportional to the intensity of the designers'
belief that it cannot.
In the early 1970s 8K of ROM was equal to 8 bits of register space.
What a reason for microcode!
---Info from Patterson, 1985
Instructions should not bind together operations
which an optimizing compiler might otherwise choose to separate
in order to produce a more efficient program.
---Brian Case
It's just a shift to the left
And then a shift to the ri-i-i-i-ght
Get a load of those MIPS
And the code's real ti-i-ight
This lockstep pipeline
Really drives me insa-a-a-a-a-ane
Let's do the pre-charge again!
---Brian Case
Measurement of repetition is not equivalent to repetition of measurement.
---Eugene Miya
The best is the enemy of the good.
- RISC
-
Relegates Important Stuff to Compilers
The maxim
Nothing avails but perfection
may be spelled
P A R A L Y S I S
---Winston Churchill
Turning off the type and range checking after the program
is debugged is like throwing away the lifeboats after the maiden
voyage.
(Or like travelling without a convoy through a submarine war zone.)
The biggest (and most common) mistake in computer design is not providing
enough address bits.
---Bell and Mudge, "Evolution of PDP-11"
Rules of Thumb
-
Every 3 years transistor account on a chip increases 2X.
-
Every 3 years chip speed increases by 2X.
-
Every 3 years DRAMs increase in size by 4X.
-
Every 10 years DRAMs get 30% faster.
-
Every 3 years Disks increase in size by 2X.
-
Every 10 years Disks get 30% faster.
-
1MB of main memory is required for every MIP.
- No way! Disproved on microcomputers!
-
1Mb/s of bandwidth is required to feed every MIP.
-
Programmers increase the size of programs 2X every year.
--- UNIX International, May 7, 1990
- Rule for debugging:
-
When in doubt, print it out!
Buying clothes more than once a decade is a waste of time
--- Andy Glew
$Header: /u/g/l/glew/public/html/RCS/sayings.html,v 1.8 1996/09/12 23:34:44 glew Exp $