[Beowulf] Re: Linux laptops, and M$ advertisement

Peter St. John peter.st.john at gmail.com
Tue Apr 17 09:43:02 PDT 2007


David, Jim,

On 4/16/07, David Mathog <mathog at caltech.edu> wrote:
>
> From: "Peter St. John" <peter.st.john at gmail.com> wrote
> > ...The testamonials all seem to be equivalent to "I"m a PhD in Rocket
> Science but setting up a Linux Cluster is too complex, so using MS saves me
> time to do Rocket Science".


Then Dave:
>

That's the way it sort of looks on the surface, but read a little deeper
> and it comes out differently. ...


...

where Matlab is the first application listed.  Matlab is
> available for linux too, but ESP is not required to guess what
> happened:
>
> 1.  Matlab too slow for some analysis on desktop Windows.
> 2.  Users beg for more power.
> 3.  Wow, we can do this on a windows cluster, then our Windows
> administrator won't need to learn anything new.
> 4.  Buy a windows cluster.


Contrast this to what Jim wrote:
>
...A more accurate model for this market would be "I'm a rocket scientist
with a Windows desktop application that provides me with a familiar
interface to a back end that might be a cluster or might not. I don't want
to have to care what's running back there in the closet, because I'm a
rocket scientist, not a sysadmin.  *I* don't much care who provides me with
a turnkey solution, I'll be happy, independent of what's under the hood. But
I'll be darned if I'm going to download ISOs off the web, burn them and
build my own cluster.  Talk to the computing infrastructure people and stop
bothering me.  And if you're selling something they don't like, you're
doomed."
<

I'm thinking that Jim's model is more like the way things ought to be (defer
to expertise), and it's that way alot in alot of places, but I think Dave's
interpretration of a buying decision (I'm using this, I have budget, I'll
buy what my own vendor sells me to solve my problem) is a challenge in a lot
of places.

I think we have seen a problem like this in several guises on this list, and
I propose a common theme.

Instances:
1. Person that has budget and problem solves problem in way immediately
accessible to himself; he can solve the problem himself by buying from his
existing vendor. This can be like catering your wedding at McDonald's
because you have lunch at McDonald's every day. So MSWin desktop user who
needs more computing and has budget buys MSWin cluster.

2. Programmer building widget for FC+=Omega needs a tool or utility that
doesn't work right so he builds his own (because he can program!) without
waiting for the volunteer contributors to the Widget Tools and Utilities
Project to do it for him, and thereby creates a dependency or strand of
spaghetti in the source tree for the next distro, and makes RGB one notch
nutsier.

3. Mathematician or Physicist discovers he needs huger number crunching and
builds his own damn computer; von Neumann in 1947 or RGB in 20007, same
thing, and a very good thing, why the impulse to solve problems for oneself
is both a boon and a bane.

4. Physicist who builds his own damn Supercomputer agonizes for days over
the emergency shut off switch wiring, because he can, he teaches
electricity; but he could just walk one hundred yards down Science Drive and
get an EE grad student to do it for him.

The theme is maybe something like marketing: exporting expertise from
sources to consumers. It's the very purpose of universities, it's the very
purpose of this list; but we're mostly exporting to each other (a great
feedback system for propelling invention), while corporations grab the vast
majority of spenders, with a bigger direct effect on the world's economy.
...

>
> David Mathog
> mathog at caltech.edu
> Manager, Sequence Analysis Facility, Biology Division, Caltech
> _______________________________________________


Thanks,
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20070417/089b536d/attachment.html>


More information about the Beowulf mailing list