Linux-Advocacy Digest #893, Volume #32 Mon, 19 Mar 01 16:13:02 EST
Contents:
Re: the mismeasure of scale ("Aaron R. Kulkis")
Re: NT vs *nix performance (Rex Ballard)
----------------------------------------------------------------------------
From: "Aaron R. Kulkis" <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy,soc.singles
Subject: Re: the mismeasure of scale
Date: Mon, 19 Mar 2001 15:21:25 -0500
The Danimal wrote:
>
> Anonymous wrote:
> > aaron wrote:
> > > Anonymous wrote:
> > > > then there are those who are in business and understand economies of
> > > > scale. not to mention the cost of paying a headcase unix guru to be snotty
> > > > and obnoxious whilst smelling up the office and dripping twinkie crumbs on
> > > > the server and making rtfm sounds with his porcine cakehole.
> > >
> > > It takes a minimum of FIVE Windows adminstrators to get the same productivity
> > > of ONE Unix administrator.
>
> That's because Unix administrators don't have to cope with the
> same types of users. The daunting complexity of Unix selects for
> highly self-sufficient users. This is like comparing the productivity
> of two physicians, one who treats terminally ill elderlies and
> the other who treats healthy young people. You'd be naive to think
> the doctor with the higher patient death rate is inferior. You have
> to compare them on the same patients.
I'm not talking about the productivity of Windows USERS vs Unix USERS.
I'm talking about the number of ADMINISTRATORS needed to keep a
site up and running.
Due to the fact that in Windows, users are not prevented from fucking
up the system....you get, naturally, a lot of systems that get fucked
up by
a) the users themselves
b) side-effects of programs being run by users
On top of that, there is the INCREDIBLY HUGE design flaw called
"the Windows Registry", which is
a) not human-readable
b) a single-point-of-failure
c) easily corrupted
d) difficult, if not impossible, to repair when corrupted
>
> The low level of real-world compatibility in Unix systems discourages
> people from connecting them to random hardware devices and relentlessly
> installing new kinds of application software as is routine with Windows
> computers. This reduces the number of incompatibility headaches for
> administrators but it also means Unix computers are not as generally
> useful. Unix computers are for the most part more specially useful---that
> is, people tend to use them for smaller numbers of more specialized
> tasks, rather than as general purpose information appliances.
This is a red herring. Are you saying that there are businesses which
require that at least one of EVERY PRINTER ever manufactured be installed
and working on site?
Of course not.
When you buy equipment, you go out and buy something that is compatible
with your operating system. And yes, there are printers that work a hell
of a lot easier with Unix than with Windows.
>
> > > The last time I worked for EDS, a mere TWENTY Unix administrators did
> > > ALL systems administration for approximately 15,000 of Unix machines
> > > throughout General Motors, all over the country.
>
> What kinds of users were using those computers, and what were they
> doing with them?
That's immaterial.
UNIX USERS ****CANNOT**** FUCK UP THE SYSTEM CONFIGURATION, regardless
of how intelligent they are.
>
> Suppose an engineer is running some specialized technical application.
> The engineer is probably technically competent. The engineer probably
> knows a lot more about that particular application than any Unix
> administrator does. The engineer's job requirements include being
> able to keep his/her own tools running. The engineer is not afraid
> to read manuals. The engineer has probably taken some programming
> courses, and may even be a pretty good programmer in his/her own
> right.
You would be amazed. By the way, the vast majority of Unix users
in GM are detailers...who tend to hover in the 100-105 IQ range.
>
> But most of all, the engineer spends most of his/her time running
> that one application. Once s/he learns how to do that, there's little
> for the administrator to administer.
And the typical Windows user spends most of his time running that
one application....Once s/he learns how to do that, the system continually
gets corrupted, providing lots and lots of work for administrators.
>
> Administration problems generally indicate how unfamiliar the users
> are with what they're doing at the moment. Which means, essentially,
> how many users are working at the edge of their abilities, trying new
> things, and experiencing the high disaster rates typical of any sort
> of exploration.
No. ADMINISTRATION problems indicate how vulnerable the system
is to service outages caused by
a) users re-configuring the machine
b) infrastructure changes.
What you're talking about are APPLICATIONS TRAINING, which is NOT
an issue for administrators. If the user can log into the machine,
fire-up the application, and it isn't crashing, but doesn't know how
the fuck to use his software, that is NOT an administration issue,
it's a USAGE issue.
For example, if you don't know which button to click to get right-hand
justificication, does that mean that your machine is broken and an
ADMINISTRATOR means to look at it? Or does that mean that you merely
don't know how to use the program, which is a SOFTWARE TRAINING issue.
Administrators do NOT deal with software training problems.
>
> Crime rates are high in boom towns and low in retirement communities
> where old people go to die.
>
> > > Conversely, twenty Windows administrators have a very hectic time
> > > keeping ONE 1,500 user site running properly.
>
> Of course. That reflects the lower average technical competence of
> a Windows user. If you moved all those people over to Unix, your
> administrative costs would certainly not decrease. It's more likely
> that your administrative costs would explode, at least initially.
No...it reflects the fact that Windows has poor system security,
because *ANY* user can fuck around with the systems settings.
>
> What percentage of the general population do you think has the
> mental capacity to master things such as the Unix shells,
> regular expressions, the 'find' command, and vi on their own?
> Consult the IQ chart:
Spot the red herring.
Every system I've worked on in the last 5 years has a GUI version
of the find command available for GUI users.
Vi.....not the only editor on the block. There are dumbed-down
"notepad"-like editors on EVERY unix system I've ever seen for
over 5 years now.
>
> http://www.sciam.com/specialissues/1198intelligence/1198gottfredbox2.html
>
> Unix is suitable for people in the IQ range under "Gathers, infers
> own information." According to Gottfredson's article:
It is also suitable for the lower IQ's...those people who NEED a
GUI because they can't deal with a command line.
>
> http://www.sciam.com/specialissues/1198intelligence/1198gottfred.html
>
> a person with a low IQ requires close supervision and up to
> six times longer overall to learn what a person with a high IQ
> picks up on his or her own. Considering that even a smart person
> requires several years to become reasonably competent with Unix
> the prospects for it ever becoming a truly mass phenomenon in
> its classical form are not promising. Is there any chance of
> putting the bulk of the population through 10 years of compulsory
> Unix education?
No need. If you take two groups of novice users, and plop them
down in front of a modern Unix system, there is the same initial
shallow learning curve.
The difference is...the Windows users have ZERO facilities to get
past the point-and-drool phase.
Meanwhile, Unix users ALWAYS have the command-line available for
doing more sophisticated tricks.
The particular user HIMSELF doesn't necessarily need to to know
how to write shell scripts, by the way....just as long as there
is a knowledgeable person available to them to do that.
Add one such person to a 20-person department, and you can very
well DOUBLE the productivity of that department, as s/he writes
customized shell-scripts to replace highly-repetitive, point-and-click
intensive tasks with simple shell scripts that do EXACTLY what is
needed to one or more files at once.
For example, maybe there is some post-processing with expense
reports that is rather labor intensive.
You can give the users the OPTION of doing that post-processing
by hand with 4 expense reports...or running a script which
automates it, so that all four are done more quickly than it
takes to type the command line:
$ clean_exp_report dallas1.exp dallas2.exp atlanta.exp manteno.exp
Further develop it by putting a graphical front-end on it, and
ta-dahh, you've helped everybody, even the point-and-drool user
who can't handle a command line.
>
> Plus the high level of real-world compatibility in Windows systems
> encourages people to purchase every imaginable type of hardware doodad
> and attempt to plug it in.
And allowing users to put various pieces of random junk, and install
drivers that corrupt the DLL's of the system helps your business
how, exactly?
> Toss in the world's largest software library
Most of which is utterly redundant.
It matters little if there are 50 billing programs available vs 5,000,
you're only going to run ONE.
> and Windows systems have BY FAR the greatest scope for multivendor
> incompatibility problems.
Primarily due to the cult of most Windows applications writers using
binary file formats instead of applications keeping as much as possible
as human-readable text files.
>
> If Linux became equally popular it would have a comparable number
> of multivendor headaches. At the moment, Linux probably has *MORE*
> multivendor problems than Windows, but most of these are so
> outrightly fatal that they discourage users from trying adventurous
> things that generate support costs.
This is an utterly false statement.
>
> In other words Windows costs a lot to support because it's so good
> that people think it's better than it is.
Dan, you're talking like somebody who has NO EXPERIENCE with ANYTHING
outside of Windows, DOS, and maybe MacOS stuff.
>
> > > When I was at Kmart headquarters, a 2,500 Windows-users site, they had
> > > close to 100 Windows administrators.
> > >
> > > If this was running Linux or Unix, the necessary support staff for desktop
> > > computers would be under a dozen people.
>
> Sure, because most of those Windows users would simply give up
> as soon as they looked at their first man page.
>
What part of "Users are NOT supposed to be configuring the system"
do you not understand?
*MOST* Windows administration problem tickets (like 90%) are caused by
one of the two following activities:
1) User fucking around with the system configuration
2) Windows doing its usual self-destruction routine.
Thus, the majority of a Windows administrator's day is spent treading
water, with the weight of "no system security" on one foot, and
"crappy kernal programmers in Redmond" on the other.
Windows admins have very little time to do improvements during
business hours. This is why windows shops have to have all of their
administrators come in on weekends and holidays to perform any sort
of on-site upgrades....it's the ONLY time they can work without
being buried under a waterfall of trouble-tickets from systems
that quit working, for no apparent reason, within the course of
the business day.
> To make a good case you need to analyze a representative set of
> Windows technical support incidents and explain how Linux solves
> those problems without introducing more problems.
Most of the problems that generate trouble-tickets on Windows machines
CANNOT OCCUR on a Linux machine because:
a) Linux and Unix do NOT allow the user to change system settings such
as network address, DNS server, etc., etc., and all the bazillion other
ways that Windows users find to completely hose their own, supposedly
"intuitive" system.
b) Linux and Unix do NOT allow a user-installed program to change
system settings.
c) Linux and Unix do NOT have a problem with Library files (DLL's
in Windows are library files) creating conflicts when one program
wants version5 of a library file, and another program wants version6.
In fact, it has been DOCUMENTED that Microsoft ***KNOWINGLY***
sabotages competitors products (such as Word Perfect), by releasing
new DLL's that SPECIFICALLY target certain products using the
previous version of that DLL.
Since there is not the same sort of incestous relationship between
the library writers and the applications writers, no such problems
exist in Unix, Linux, and other platforms.
>
> By far the best way to reduce upgrade costs is to make sure your
> new system does everything the old system did, in addition to whatever
> new features you're buying the new system to get. That way the new
Again, this is where Unix and Linux are complete winners, and
Mafia$oft product can't even get into the arena.
Try getting the source code to a DOS program from 1985, compiling
it, and then try running it on your Windows 2000 system.
I can GUARANTEE you that it will not work.
Try using a Windows app written for Windows3.1, and doing the same
thing on Windows2000. Once again...IT WILL NOT WORK.
Conversely, get some source code for a Unix program written in 1975,
and compile it on your Linux box (notice that Linux even INCLUDES
COMPILERS WITH THE DISTRIBUTION)...and LO AND BEHOLD...your 25-year
old Unix code runs!!!!
Why? Because backwards compatibility is a primary consideration
in the Unix design....to the extend of writing NEW libraries with
the KNOWLEDGE that, eventually, they will be superceded by newer
versions of the same functions, with larger datasizes.
The 16-bit and 32-bit libraries in Windows are UTTERLY incompatible
with each other.
Conversely, the 16-, 32-, and 64-bit libraries in Unix land are
strictly backwards compatible...such that, often times, all a programmer
needs to do to migrate up to a new datasize is ... run the compiler
program again.
NO MODIFICATIONS TO THE CODE IS NECESSARY.
This is why the Windows camp is continously re-inventing the wheel,
and often, with much fanfare that sounds litterally as stupid as
"New MS-WHEEL 3.0...now with EIGHT sides"...because the current
crop of programmers at microsoft seem to display an INCREDIBLE
lack of recognition of basic computing principles...which are very
much the equivalent of someone who doesn't realize that proper
wheels are CIRCLES, not polygons.
> system does not call attention to itself. It does not get in the
> way of the users. It does not destroy the massive existing value
> of pre-existing knowledge. By far an organization's most valuable
> computer-related asset is the knowledge of its people.
>
The "'massive' existing value of pre-existing knowledge" at Windows
sites is all novice-level.
You've got windows users who think they are "power-users" because
they can push the limits of the "designed for--and meant to keep
everybody as--novices" GUI of Windows....or because they've figured
out how to do "mail merge" in MS-OFFICE.
Frankly, the functionality of the GUI is the ***LOW*** end of computing
sophistication, and especially, productivity. And mail-merge, which
*IS* a true productivity boost...is the same kind of batch-processing
which you "there shall be nothing other than a GUI" folks belittle.
> Esperanto is a much more efficient and logical way to communicate
> than English. It has a simple, regular grammar and is vastly easier
> to learn and use correctly than any language that evolved spontaneously.
> Anybody who recognizes Linux as being better than Windows should
> be all over Esperanto.
False analogy.
Nobody uses Esperanto for anything of importance. Conversely,
I would defy you to find even ONE automotive of aerospace part
designed in the last 8 years that wasn't designed on a Unix system.
I defy you to find ONE bank or securities firm that keeps their
records on servers running Windows.
You're argument is based on the idea that Windows must be superior
because so many people WHO HAD NO OTHER OPTION on cheap Intel hardware,
bought it, as opposed to not getting any affordable computers.
***NOT*** that there have not always existed competing Intel-based
Operating Systems....only that Microsoft negotiated ILLEGAL anti-
competitive contract to ***PREVENT*** PC manufacturers from even
offering non-Mafia$oft operating systems to their customers.
*ILLEGAL*
*ILLEGAL*
*ILLEGAL*
For example...if a Chrysler dealership want to start selling
Lincolns and Cadillacs...he is FULLY allowed to do so...and if
Daimler-Chrysler, Ford, or GM try to do ANYTHING to punish him
for offering competitors products...they will be in court so fast
it would make your head spin.
Microsoft's contracts were similarly illegal...and they KNEW IT.
They specifically abused the Non-Disclusure Agreement mechanism
to specify that regardless of whether one accepted a Microsoft
contract, one could NOT disclose the terms of the agreement to
ANYBODY without prior approval of Microsoft's lawyers...INCLUDING
LAW ENFORCEMENT PERSONNEL.
Fortunately, in the most recent Justice Department prosecution,
the OEM's were able to get some of those contracts SUBPEONEAD,
which takes precedence over the NDA.
>
> So why doesn't any large company order its people to communicate
> only in Esperanto? Because the cost of dealing with the
> inefficiencies of English is microscopic compared to the cost
> of translating everything you're doing now into Esperanto. Doh!
When you write a document in English....do you expect that, 24-months
down the line, it will be fatlly corrupted....or that 10 years down
the line, the English API will no longer be available, let alone
supported...rendering all of your English-langauge documents as
so many sheets of firestarter...
2
DOH!
>
> Learning is the most costly human activity. It's the largest
> component of capital stock. Throwing away your knowledge is like
> burning down your factories, only worse because at least when
> you burn down an old factory you can get rid of a lot of worn-out
> equipment. Knowledge does not degrade with use; it gets better
> with use (although it can become irrelevant if external requirements
> change).
Most Windows GUI knowledge is IMMEDIATELY transferable to any other
GUI.
the difference is...on Windows...that GUI is the boundary WITHIN
WHICH the user must remain confined....on Unix and Linux, the GUI
is the STARTING point, from which the use, ***MAY***, if he/she
***CHOOSES***, advance to more skillful use of the system.
>
> For Linux to become a serious competitor to Windows it must
> be compatible with Windows on at least an application level.
>
That criterion was met three YEARS ago, Dan.
> Perhaps the high cost of Windows administration suggests that people
> are trying to do TOO MUCH with their computers, but this is not a
No. It indicates that Windows is a house of cards, which has
a) no security...not only from malicious hackers, but from
authorized users as well.
b) corrupts itself frequently.
I won't even go into the problems with the poorly-designed NT filesystem,
which pretty much FORCES Windows Neutered Technology to be re-installed
every 6 months, because it doesn't reuse a key component of a file
structure after the file has been deleted....which means that eventually,
YOU RUN OUT OF THEM....and thus, are unable to create *ANY* new files.
This situation was documented over two years ago.
Has it been fixed by Microsoft?
a) yes
B) HELL NO!
> Windows problem. "Too much" in this context would mean attempting to
> combine hardware and software from multiple vendors in ways those
> vendors did not anticipate, or by users who lack the ability to
> do what they are attempting. If users are pushing the envelope
Why does *NO* other operating system suffer from this sort of
problem on *ANY* level.
it's ONLY Windows that has a problem with two randomly-selected
programs corrupting each other.
> they are going to have problems, no matter what OS they are running.
> The question then becomes why are users doing this? To me it sounds
> more like an industry infrastructure problem than anything to do
> with the OS. Vendors need to come up with a better way to tell
> users which combinations of their products will work.
Your ignorance in this regard is showing, Dan.
In all NON-Windows Operating Systems, this is a non-issue, because
on other Operating Systems, application installers are NOT allowed
to touch ANY libraries used by other applications.
>
> There's also a massive need for better error diagnosing. Computers
> are tools for handling information, and the most critical information
> is whatever information you need to keep the computer itself running.
> Every error message should include a unique identifying code (with
> every application having its own distinct error code namespace,
Compared to the massive deluge of fundamental design flaws in
Windows, this is of trivial importance for Windows users.....
> much like I.P. addresses) and should allow the user to click on
> a hyperlink that will automatically retrieve the latest accumulated
> knowledge about (a) the conditions that cause the error and
> (b) how to fix it. Currently the ridiculously poor organization
The basis for such systems are already commonplace at many Unix
sites, and have been for a couple of years now.
> of the industry forces thousands of people to independently
> rediscover the solutions to the same problems. It's as if the
> industry is in denial about the nature of errors and what kind
> of social organization is necessary to deal with them intelligently.
No..that's just Microsoft. The rest of the industry's operating
systems crossed such hurdles LONG ago, Dan. The only reason that
you consider these ideas to be "something to wait for" is because
you stick with Microsoft.
>
> --- the Danimal
--
Aaron R. Kulkis
Unix Systems Engineer
DNRC Minister of all I survey
ICQ # 3056642
K: Truth in advertising:
Left Wing Extremists Charles Schumer and Donna Shelala,
Black Seperatist Anti-Semite Louis Farrakan,
Special Interest Sierra Club,
Anarchist Members of the ACLU
Left Wing Corporate Extremist Ted Turner
The Drunken Woman Killer Ted Kennedy
Grass Roots Pro-Gun movement,
J: Other knee_jerk reactionaries: billh, david casey, redc1c4,
The retarded sisters: Raunchy (rauni) and Anencephielle (Enielle),
also known as old hags who've hit the wall....
I: Loren Petrich's 2-week stubborn refusal to respond to the
challenge to describe even one philosophical difference
between himself and the communists demonstrates that, in fact,
Loren Petrich is a COMMUNIST ***hole
H: "Having found not one single carbon monoxide leak on the entire
premises, it is my belief, and Willard concurs, that the reason
you folks feel listless and disoriented is simply because
you are lazy, stupid people"
G: Knackos...you're a retard.
F: Unit_4's "Kook hunt" reminds me of "Jimmy Baker's" harangues against
adultery while concurrently committing adultery with Tammy Hahn.
E: Jet is not worthy of the time to compose a response until
her behavior improves.
D: Jet Silverman now follows me from newgroup to newsgroup
...despite (C) above.
C: Jet Silverman claims to have killfiled me.
B: Jet Silverman plays the fool and spews out nonsense as a
method of sidetracking discussions which are headed in a
direction that she doesn't like.
A: The wise man is mocked by fools.
------------------------------
From: Rex Ballard <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: NT vs *nix performance
Date: Mon, 19 Mar 2001 20:25:24 GMT
This is a multi-part message in MIME format.
==============734CE4C68D55D778F0467304
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Chad Myers wrote:
>
> "Bob Hauck" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > On Sat, 3 Mar 2001 21:42:08 -0500, JS PL <js@plcom> wrote:
> >
> > > Only 95% are forced to buy windows? Why not 100%?? What is different
> > > about the five percent that they are exempt from being "FORCED" to buy
> > > windows?
> >
> > Presumably they bought Macs instead of PC's.
>
> Which is another reason why MS does not have a monopoly. They
> can never establish price controls because Macs are a reasonable
> alternative to the PCs and consumers could reasonably switch to the
> Mac
Keep in mind that Microsoft holds key leverage positions in the form of
preinstalled Microsoft Office for Mac. Net effect, Mac or Windows, you
will pay Microsoft.
> if the PC/Windows realm became too pricey (and indeed they
> have as we have seen with the recent iMac v. PC price wars).
Mac has one major OEM whose production capabilities are limited.
Apple's
distribution channels are also limited (due to Microsoft). Third party
software availability is also limited (also due to Microsoft), and many
of Mac's competitive intellectual property advantanges (Quicktime video)
have
been driven out of the market (due to Microsoft). Finally, Microsoft
refuses
to adhere to established standards, which means that Apple must pay
Microsoft
for the licenses.
Microsoft has exclusive agreements with the OEMs who produce most of the
desktop
machines sold worldwide.
UNLESS you purchase a machine preinstalled with Linux or UNIX, you are
forced
to pay money to Microsoft and feed the Microsoft monopoly machine.
> -c
--
Rex Ballard
It Architect
http://www.open4success.com
==============734CE4C68D55D778F0467304
Content-Type: text/x-vcard; charset=us-ascii;
name="rballard.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Rex Ballard
Content-Disposition: attachment;
filename="rballard.vcf"
begin:vcard
n:Ballard;Rex
tel;cell:908-723-4008
tel;work:973-723-4008
x-mozilla-html:FALSE
adr:;;;;;;
version:2.1
email;internet:[EMAIL PROTECTED]
fn:Rex Ballard
end:vcard
==============734CE4C68D55D778F0467304==
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
