🚀 go-pugleaf

In <rmcassid-2108951317450001@dante.eng.uci.edu> rmcassid@uci.edu (Robert Cassidy) writes:

>In article <danpop.808872265@rscernix>, Dan.Pop@mail.cern.ch (Dan Pop) wrote:
>
>> In <docjoe-1808952118480001@ip198.wilmington.de.interramp.com>
>docjoe@interramp.com (Joe Ragosta) writes:
>
>> >Which is why published studies demonstrate that the user has a 50% higher
>> >productivity and 50% higher accuracy using a Mac than a Windows machine.
>> >And Mac support costs are thousands less. And so on. Who cares about
>> >computer speed when the real issue is how much work you can get done?
>> 
>> I definitely get more work done on Linux than on MacOS, no matter how
>> much faster the Mac processor is.  Your studies are worth zilch to me.
>
>And I definately get more work done on MacOS than UNIX, which just goes to
>show that your individual cases are worthless. Let's look at the majority
>of users - I'll give them MacOS, you give them Linux - we'll see who gets
>more work done. That's the basis of the argument - that for *most people*,
>MacOS is better. Certainly UNIX is better for some, and DOS, and Windows,
>and AmigaDOS but for most people, MacOS is probably the best choice.

Yet most people are born masochists and deliberately avoid the OS which
is the best choice for them, especially these days, when there is no
significant price penalty.  OK, injecting hard facts when discussing
with an advocate isn't fair play :-)

I thoght I made my point clear enough.  Linux is the best choice for me
and I couldn't care less about what's the best choice for anybody else.
I don't give Linux to _anybody_, except people who come to me and ask for
it.  Even my boss, a long time Mac user, asked me to install Linux on his
PC notebook.
>
>> That's easy.  What OS was the Pentium running?  What applications were
>> involved in the tests?  You can prove anything you want to prove by using
>> "real world" applications.  All you have to do is to choose the applications
>> according to what you want to prove.  This is why the industry recognizes
>> the SPEC suites and not "side-by-side application tests" as a measure of
>> performance.  Get a clue.
>
>But nobody runs SPEC. SPEC is great if you want to design something for
>the hardware, not compare complete systems (apps, OS, and all). So fine,
>PPC is 30% faster than x86 right now. Why is SPEC a good measure for the
>95% of the population that work with a system that SPEC is invalid for?

SPEC is composed of real applications running on real hardware, hence it
compares complete systems.  How about getting a clue about what you're
talking?
>
>> BTW, there is no ideal way to test the performance of a machine as a 
>> general purpose system.  It just happens that the SPEC benchmarks are
>> the best we have right now.  But they show the PPC to be too slow and
>> the Pentium to be too fast for the Mac bigot's tastes, so let's screw them
>> and go back to side-by-side application tests performed with a carefully
>> chosen application set.
>
>Sure there are good tests. Application tests are good. Take the top
>selling 10 software packages that are available commonly for each platform
>and compare them. That's what people are buying and thats what people are
>running. How about it? The users pick the app set, deliver on what they
>buy.

Then, it's up to those users to do the tests for themselves.  I don't use
any of the top 10 selling software packages, hence such a test would be
irrelevant to me.  Some of my applications are similar to those in the
SPEC suite (especially 085.gcc, 026.compress and 022.li), so the SPECint
number is the primary criterion for _me_ when comparing systems.  I couldn't
care less about the performance of Word, Excel or PhotoShop or any 
comparison based on this kind of applications.

Even if PhotoShop or Excel would run 100 times faster on a Mac than a PC,
I wouldn't consider switching platforms if the SPEC data wouldn't show that
this makes sense (and even then, unless the OS I'm interested in runs on
the Mac).  I never installed Windows on any of my computers and I wouldn't
consider installing MacOS either.  Keeps me on the sane side.

Nevertheless, I don't think that a comparison based on the top 10 apps
would be flattering for the Mac.  A large part of those apps would come
from MS, and we all know that the performance of MS apps sucks on non-MS
OS's.  But feel free to try it.

Dan
--
Dan Pop
CERN, CN Division
Email: Dan.Pop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland