Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!snorkelwacker.mit.edu!stanford.edu!leland.Stanford.EDU!fangchin
From: fangchin@leland.Stanford.EDU (Chin Fang)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Quoting PC Magazine and much more(was Re: Comparing 486 to 386 Systems)
Message-ID: <1991Apr9.213259.17287@leland.Stanford.EDU>
Date: 9 Apr 91 21:32:59 GMT
References: <40409@netnews.upenn.edu> <84050014@hpfcmdd.hp.com>
Organization: AIR, Stanford University
Lines: 91

In article <84050014@hpfcmdd.hp.com> edm@hpfcmdd.hp.com (Ed Moore) writes:
>The 12/25/90 issue of PC Magazine, page 169, claims that a 386/33 gives more
                      ^^^^^^^^^^^^^
>bang for the buck if a math coprocessor isn't needed.

I don't want to belittle or slight a well established, much read and popular 
computing mag.  But I would like to point out one fact -> PC Mag's editors
are humans beings too. => they would make mistakes just like the rest of 
us.  Sure, they have big marketing clut, if they say yes, then a product 
immediately becomes hot, and vice versa.  This has both good and bad side
when someone relies on it soly for info.

Two examples:

(1).

One big editor (hint, who writes the Inside Track) sometime earlier this year
(? I don't read PC Mag anymore) claimed that someone TOLD him that MIT's X
Windows can be only run on 486 class machines.  So he want on and said that
he considered X just another passing fad!

I personal run X on a 3yrs old 20Mhz 386 without cache (8megs however) on a
daily basis, right now I am typing this message from a xterm emulator on my
box in the X environment.  The display speed is on par with SUN SPARCstation 1
at medium load.  An he (or whoever talked to him) considered this SLOOOOOW???

I pity him, because he obviously is not exposed to one of the most exciting 
public domain softwares, X386 server wrtten by the brilliant German Thomas Roell
for 386/486 boxes running 386 Unices.  Thomas himself uses a 33Mhz 8meg memory
and he can always achieve a composite display performance BETTER than SPARC 1!!

And this PC Mag editor calls this SLOOOOOOW?

BTW, X11R4 does not require a math coprossor, Thomas doesn't have one.

(2).

Recently, IDEK 17" monitor was picked as PC Mag Editor's choice.  I played with
a brand new one right out of box extensively using Thomas Roell's configurable
X server to give it a though work out at display res from 640x480 all way up
to 1152x900, other than the last one, I always used 256 color, non-interlaced
resolution.  I will say my impression:

If I were to pick a 17", almost flat screen monitor for extended period of high
resolution use, I WON'T pick IDEK!!  

The worst shortcoming of this monitor?  -> narrow horizontal sync frequency 
range, it's upper limit is ONLY 50 Khz!  That's bad! (see below for reason)

Because a display's resolution is defined the portion of time of the a electron
beam sweeping from left side to the screen's right side and back to the left 
side.  So, if such a sweeping takes 1280 clock tics of the current vedio 
driving clock timing, and if you somehow can take 960 clock tics out of the
1280, then you be definition have a horizontal resolution 960.  Vertical res
is defined similarily.  This is all in layman's language, so technical precise-
ness is not claimed here.

BTW, YOU HAVE NO REASON TO USE 640X480 800X600 OR 1024X768 by the above 
definition.  That's why I used an unothordox 960 as example. Suffice to say 
if you want to achive HIGH SCREEN REFRESH RATE, you need to have board 
(eg high limit) horizontal sync frequency range, with upper limit of it 
perferable in 65Khz ~ 70Khz or beyone if you can find one.  The boarder/
higher, the better.  For more details, see X386 release notes for a tutorial
on vedio timing basics.  Also if you want to buy monitor, please heed my
advise here too.

In addition, I am SICK of seeing montor ads claiming non-interlaceness.
Noninterlaceness DOES YOU NO GOOD if the screen refresh rate is low, 
like 60Hz or below.  One can have all non-interlace h/she wants but with a low
image refresh rate, defined as the driving clock (eg. 65Mhz) divided
by the time required to sweep a electron beam from left to right back to left 
AND from top to bottom back to top, then your eyes will suffer badly!

Go with high refresh rate, like 70Hz or even higher, NOT just non-interlaceness

It's sad in the world of Bill Gates, 640x480, 800x600 or all VESA suff
are all that are available to people.  If IBM had not come up XGA, I am
afraid that MSDOS users would be trapped in 1024x728 for much longer.

Enough said.  I just want to point out the potential fallacy of quoting 
PC Mag (or even scientific publications sometimes) without checking out your
quote carefully.  Please don't construe my comments in this article as flame
to PC Mag or anyone. I only hope my reminder would bring in some new 
perspective to this discussion.

Regards,

Chin Fang
Mechanical Engineering Department
Stanford University
fangchin@leland.stanford.edu