RE: Excellent Interview
- From: "Chris Hofstader" <chris.hofstader@xxxxxxxxxxx>
- To: <programmingblind@xxxxxxxxxxxxx>
- Date: Tue, 29 Apr 2008 07:26:55 -0400
I think Knuth's single mindedness, whether he is right, wrong or somewhere
in between is a cornerstone of his success. He may be 100% wrong in his
ideas about parallel systems but he is not the only professor/theorist on
Earth nor is he closed to debate, discussion and criticism as long as it
comes in a polite and respectful manner - a courtesy he usually affords
those with whom he disagrees.
No one in the world of theory and discovery will ever be right all of the
time. Chomsky, the greatest linguist of the past century, was wrong about
other primates having a capacity for language; Minsky and Rodney Brooks
continue to debate about the direction robotic research should go; McCarthy
and Minsky will still debate over minutia in the differences between
Stanford Lisp, scheme and Common Lisp, etc. No one will say that all of
these guys who have made incredible contributions to our field are "stupid"
because we disagree with them on a particular issue in their world view.
cdh
-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of Sina Bahram
Sent: Monday, April 28, 2008 8:51 PM
To: programmingblind@xxxxxxxxxxxxx
Subject: RE: Excellent Interview
By that definition, everything is not universally useful. And of course it's
not, so there is no need for one to claim that it is.
I just think that he's being rather finite about it because he's even
referring to it as parallelism. That's a current conceptual crutch we need
so that we can move on in our understanding of the current model of
computing. True Turing machines are parallelizable ... Tens of processes
happen simultaneously already in all of our machines and have for years and
years. The fact that we make believe that things are parallel with context
switches and time sharing, or whether we actually do have parallelism is a
purely hardware advancement that should simply help things instead of hinder
them, but folks have been relying upon parallelism being only pseudo true
for so long, that it's a shocker now that, oh my God, a race condition
really does happen, even though it never occurred when I ran this program on
a single core machine.
For example, do you think that quantum computing experts call their model
parallelizable? The smart ones might use that term to help classical folks
understand what's going on, but of course it's parallelized, and so what? It
doesn't matter ... It simply helps you and allows a much wider array of
solutions to be explored at the same time. I just think it's silly to
attribute so much controversy to this concept. It's not that one is better
... It's the reality that silicon aint getting all that faster for the short
term, and we have some amazing solutions that have some implicit benefits.
It's not the time that a process takes that annoys users. It's the fact that
after that one is done, then three others that don't' even depend on it have
to now go do their thing ... Well, those could have been going at the same
time. I'm talking OS layer, so forget about whether you use it or not in
your programs ... The language, virtual machine, or other abstraction layer,
in 10 years, will probably just handle all that anyways, but the act of
actually being parallelized I think can only help, given our current
situation.
Again, it's just disappointing that he's not being rather generalized about
this with respect to his view, and instead he chooses to dismiss it so
superficially.
Take care,
Sina
-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of Ken Perry
Sent: Monday, April 28, 2008 4:32 PM
To: programmingblind@xxxxxxxxxxxxx
Subject: RE: Excellent Interview
Actually I think he was bang on, on that part of the interview but I promise
everyone not to get into a flame war over who is right who is wrong. If you
noticed with what he said he did not say it was useless. I just think a lot
of people are pushing parallelism into places where it don't need to be. I
think I will now go back to writing my simple parallel version of
tic-tac-toe and be quiet.
Ken
-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of Sina Bahram
Sent: Monday, April 28, 2008 12:58 PM
To: programmingblind@xxxxxxxxxxxxx
Subject: RE: Excellent Interview
Unfortunately, he seems to be becoming more and more out of touch with
reality.
I was quite disappointed at his rather short term, rather uninformed in some
ways, view of parallelism.
Take care,
Sina
-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of Chris Hofstader
Sent: Monday, April 28, 2008 1:53 PM
To: programmingblind@xxxxxxxxxxxxx
Subject: RE: Excellent Interview
I'm glad you guys enjoyed it as much as I did. When we did the brief on
behalf of Borland in the Supreme Court case Lotus v. Borland that tried to
claim copyright on a user interface, Knuth was one of our signatories.
Unlike the rest of the group, he had some additional demands we needed to
agree to before he would sign on. Specifically, all communications
regarding the matter must come to him by email from Stallman and no one else
and, including rms, no one was to call him on the telephone. As he says in
the interview, he is a strange guy.
cdh
-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of Jared Wright
Sent: Monday, April 28, 2008 11:29 AM
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: Excellent Interview
Agreed. I saw this come across Reddit last week and was very engaged by it.
Thanks for giving me an excuse to read it again.
Jared
Ken Perry wrote:
> Very nice Chris.
> Ken
>
> ----------------------------------------------------------------------
> --
> *From:* programmingblind-bounce@xxxxxxxxxxxxx
> [mailto:programmingblind-bounce@xxxxxxxxxxxxx] *On Behalf Of *Chris
> Hofstader
> *Sent:* Monday, April 28, 2008 4:24 AM
> *To:* programmingblind@xxxxxxxxxxxxx
> *Subject:* Excellent Interview
>
> Hi,
>
> At: http://www.informit.com/articles/printerfriendly.aspx?p=1193856,
> you will find a somewhat lengthy interview with Donald Knuth.
>
> For those too young or clueless to know who this incredible
> contributor to computer science and algorithmics is and the
> contributions he has made:
>
> . Shortly after rms "invented" free open source software, Knuth
> release TeX under GPL 1.x which later grew into LaTeX, a system still
> widely used by blind and other people looking to express mathematical
> and other interesting expressions in an unambiguous manner. Along with
> emacs, this represents one of the very pioneering steps in free software.
>
> . Knuth has published many of the most important books and articles in
> the history of computer science. He has invented many of the most
> important algorithms as well as incorporating the works of many others
> into his nearly canonical volumes on computer science.
>
> . Knuth is a terrific professor at Stanford University and is one of
> the most accessible individuals in the field.
>
> Enjoy,
>
> cdh
>
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________ NOD32 3060 (20080428) Information __________
This message was checked by NOD32 antivirus system.
http://www.eset.com
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________ NOD32 3062 (20080429) Information __________
This message was checked by NOD32 antivirus system.
http://www.eset.com
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
Other related posts:
