Re: Stupid MS challenge

Riley Williams (rhw@MemAlpha.CX)
Sun, 16 May 1999 15:29:35 +0100 (GMT)


Hi Richard.

>>> I think that if Linux does go ahead with the rematch, we
>>> add the following added requirements:

>>> Finally, the test should be run in continous mode for *1
>>> month*...

>> Be sensible!!! "1 month" depends on who's measuring it, and
>> when. The test should run in continuous mode for 5 weeks,
>> which is a much more precice measurement of time...

> NT doesn't run anything for more than a day. It stays up, but
> the applications always stop. This is considered "normal" by M$.
> Even M$'s C++ environment needs to be rebooted between compiles
> or it won't link.

You think I don't know this? AUCC in its wisdom expects its students
to use NT to do their computing work, and few of the systems stay up
for more than 90 minutes without locking up. AUCC's answer to this is
to limit students to not more than 60 minutes at a time...

> Running a benchmark for a month will guarantee that even a CP/M
> machine will beat NT in overall performance, even if the database
> is on 96k, hard-sectored, single-sided floppies that have to be
> changed by hand.

I look at it a different way: Requiring the benchmark to run for a
decent period will, in my honest opinion, result in one of two things:

1. Microsoft producing a version of NT that is stable enough to run
for at least the designated period of time.

2. Microsoft participating in the benchmark and losing - and being
unable to hide the fact because of the amount of publicity the
said benchmark session has already had.

3. Microsoft refusing to participate since they know that they can't
win in such a scenario.

I personally believe that Microsoft are incapable of doing the first,
and can't afford the second, and therefore that the last is a given,
but I can also see Microsoft using the fact that the phrase "a month"
is imprecice to make it appear that different time measures were being
used to bias the results against them and they withdrew because of it.

THIS is the reason that I ask for a PRECISE time measurement to be
used, rather than an imprecice one like "a month". I'm not worried in
the slightest which one, any of the following, along with any other
precice term, is acceptable, providing it's at least 28 days long:

28 days 4 weeks
29 days
30 days
31 days

35 days 5 weeks

91 days 13 weeks

Best wishes from Riley.

+----------------------------------------------------------------------+
| There is something frustrating about the quality and speed of Linux |
| development, ie., the quality is too high and the speed is too high, |
| in other words, I can implement this XXXX feature, but I bet someone |
| else has already done so and is just about to release their patch. |
+----------------------------------------------------------------------+
* ftp://ftp.MemAlpha.cx/pub/rhw/Linux
* http://www.MemAlpha.cx/kernel.versions.html

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/