Linux-Advocacy Digest #203, Volume #27 Tue, 20 Jun 00 07:13:03 EDT
Contents:
Re: Claims of Windows supporting old applications are reflecting reality or fantasy?
(John Wiltshire)
Re: Claims of Windows supporting old applications are reflecting reality or fantasy?
(John Wiltshire)
Re: How many years for Linux to catch up to NT on the desktop ? (mark)
Re: Claims of Windows supporting old applications are reflecting reality or fantasy?
(JEDIDIAH)
Re: Number of Linux Users ([EMAIL PROTECTED])
Re: An Example of how not to benchmark ([EMAIL PROTECTED])
Re: Number of Linux Users (JEDIDIAH)
----------------------------------------------------------------------------
From: John Wiltshire <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Claims of Windows supporting old applications are reflecting reality or
fantasy?
Date: Tue, 20 Jun 2000 10:22:57 GMT
On Mon, 19 Jun 2000 23:52:01 -0700, <[EMAIL PROTECTED]> wrote:
>
>John Wiltshire <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> On Sun, 18 Jun 2000 09:23:28 -0700, <[EMAIL PROTECTED]> wrote:
>>
>
>> Fair enough. Ignore my other reply to Colin then.
>
>Done.
>
>> My impression is that Microsoft really didn't know what the hell it
>> was going to build when it was working on NT, so it covered all bases.
>
>Legal bases as well, the varations of the promised user interfaces was a
>result of the not knowing what would happen in court in their case agains
>Apple.
Makes sense.
>Some of the comments I have seen in this thread had me thinking that the
>silly season was here. Like Windows NT could not have had a interface based
>on Windows 3.x, because Windows 3.x was not yet out the same for the 486 and
>the devlopment of the 586 (Pentium). As well as the remaining comments
>against my positions and statements in this thread.
>
>Most of my statements in this thread are supported by documents provided by
>Microsoft itself. Not supported by the information available on-line from
>their web sites, rather supported by the hardcopy documents supplied by
>Microsoft at the time of the development of Windows NT.
>
>Website documentation can be altered to creativly modify history remove
>references to promises not delivered, but the hardcopy documentation is as
>good as being set in stone.
>
>A copy of the book "Inside Windows NT" published by Microsoft Press that was
>provided as a handout during a Microsoft Wndows devlopers seminar. The
>seminar took place before the release of Windows NT.
>
>We were instructed to develop software for Windows NT on Windows 3.1 hosts
>using the wins32 addon, so that our software could be available when NT is
>released. How could this be possible if Windows 3.1 was not yet available
>until aftr Windows NT? The book cites the 80386 and the 80486 as existing
>platforms. Other hand outs cited the future of Windows with the 586 How if
>the 486 was not yet out and the 80586 was not yet being planned?
I said that the development of NT was started before Windows 3.x was
release (read my post again). I have read "Inside Windows NT" (both
first and second editions) many times and know what is in there. The
facts are:
i) NT development was started in 1986/87.
ii) Windows 2.x had not been released until November 1987.
iii) The 486 was released in 1987, not in widespread use until 88/89.
Now, I stated: "As the NT design was started *before* Windows 3.x, it
is hard to imagine how a Win3.x shell was going to be part of it.",
whereas you stated "There was going the be a Windows 3.x windowing
system...". You are clearly wrong.
NT 3.1 was released in September 1993. By this time a lot had
changed, including Windows 3.1 being released, Win32s being developed
and many other things. I've still got the betas of NT3.1 and the
original SDK and 32 bit compiler.
>The book and the handout promise support for ALL windows programs. The book
>limits the suppor for Dos programs to those that do not need direct hardware
>access. BUT, a handout that more current that the book, (titled "Providing
>hardware access to MSDOS programs under VDM") outlined methods that HAD BEEN
>implemented to permit Dos programs that need direct hardware access to run
>on Windows NT, providing that it is implemented on computer with the
>particular hardware actually present. This was supplemental to the virtual
>hardware access that is provide for some feature today by VDM today.
Read the DDK - it is available on the web. It is possible (still) to
emulate any hardware you like in the VDM. You just seem to think it
is Microsoft's responsibility to do this and not the
hardware/application vendors. I see no reason for this.
If you'd ever written an NT device driver you would understand the
absurdity of VDM programs writing directly to hardware.
Also, I'd like you to tell me the page in Helen Custer's book that
promises support for ALL windows programs (including those with custom
VxDs and that write directly to hardware). I seriously doubt you'll
find it unless someone's written new stuff in your book.
>There were more promises such as: full support for unix programs written to
>the POSIX standards once recompiled for the NT platform. Full support for
>OS/2 programs in binary form on Intel processors and source code
>recompilations on other processors. Future support for other API and
>emulations when needed, including MacIntosh programs, NetWare NLM's, BSD,
>Unix, Vax, VMS, and PDP's. (About the BSD unix, Vax and VMS, that is not my
>error, it comes from Microsoft.)
Sure. I won't deny that they did promise a lot more than they
delivered. Directions change for every system. I've been coding for
Windows (Application and Driver) since v1.0 and OS/2 since 1.1 and
really haven't missed much of the stuff you mention above. Fact
probably is it wasn't worth Microsoft's money coding all those things
in when what people really wanted was a better version of Windows 3.1.
Did they promise more than they delivered? Sure.
>Almost everything that people have taken exception to in my position are
>supported by these documents.
I've read the documents. I'm not disagreeing with them. I'm
disagreeing with your somewhat distorted view of them.
>The exceptions are the opensource drivers by
>hardware companies and Microsoft distributing Window's 1.x and 2.x as
>freeware.
Both of which are your opinion.
>I am supprised to the reacation against open source hardware drivers. At
>one time no hardware devices came with any drivers. The manufacturer would
>supply the documentation of the interaction of the device with the host
>computer, often it was included with every copy of the device sold. It was
>then the job of the operating systems programmers and other interested
>parties to develop what every driver they would need to handle the device.
>After a while some manufactureres started writing drivers by way of ROM's
>containing the Bios extensions for the devvice. Then the manufacturers
>started providing device drivers and some where along the line they stopped
>providing the documentation of their hardware as they used to. Also
>Microsoft came out with their "assistance to the hardware manufacturers"
>which "required" non-discloseure and non-competition clauses.
>
>We often hear about the high cost that hardware manufacturers have had to
>endure to provide device drivers. With what I was proposing would relieve
>the hardware manufacturers of this burden or at least reduce it for them to
>a level as low as they prefer. The cost savings could be passed on as lower
>prices for the consumers or they could take the money they are saving and
>plow it right back into hardware R&D, resulting in more and better and more
>economical hardware for all of us.
You deny that drivers have value in their source. Your proposal
offers to strip hardware companies of the thousands of dollars
invested in a driver that gives their product an advantage over the
competition. Compare the nVidia GeForce to the S3 Savage2000. Chips
are about the same power. Drivers excelled on the GeForce and sucked
on the S2000. Why should (as you propose) nVidia give up this
enormous commercial advantage?
The simple fact is your proposal does not hold water when you consider
the value which is contained in the driver. See my response elsewhere
in the NG for more.
>The Freeware distribution of "olderware" is not unheard of. In fact it
>often has beneficial results for the company that does this. I say: If it
>was a marketable software product and it still has value why not sell it?
>If it is "valuless" what is there to loose by giving it away? Why lock it
>away in a vault and benefit no one at all?
Sometimes there may be value in not giving the product away? What if
a company stands to lose by giving a product away, but simply not gain
by trashing it? You fail to consider issues of technical support,
sales channel distribution and other non-software costs.
>> When Windows 3.x skyrocketed in popularity they decided to run with it
>> and we now have the monopoly we know and love today.
>
>Love? The only monopoly I love contains Park Place, Boardwalk, and Marvin
>Gardens.
Never heard one of those. I've got one with Mayfair and Park Road
though. Yours must be a cheap imitation. ;-)
John Wiltshire
------------------------------
From: John Wiltshire <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Claims of Windows supporting old applications are reflecting reality or
fantasy?
Date: Tue, 20 Jun 2000 10:24:15 GMT
On Mon, 19 Jun 2000 20:47:24 GMT, [EMAIL PROTECTED]
() wrote:
>On Thu, 15 Jun 2000 23:35:00 -0700, Stephen Edwards <[EMAIL PROTECTED]> wrote:
>>
>>
>>
>>This is incorrect. WindowsNT runs a modified
>>microkernel design. Windows9x is a tad more like
>
>Windows NT is anything but a microkernel. Since 4.x when the GUI was
>merged into the kernel, it has been a bloated pig of a megakernel.
Read the post. NT runs a *modified* microkernel. In fact, NT is
still based around a microkernel with a modular architecture around
the core. Stop showing ignorance.
John Wiltshire
------------------------------
From: [EMAIL PROTECTED] (mark)
Subject: Re: How many years for Linux to catch up to NT on the desktop ?
Date: Fri, 16 Jun 2000 11:41:45 +0100
In article <[EMAIL PROTECTED]>, Pete Goodwin wrote:
>[EMAIL PROTECTED] (2:1) wrote in <[EMAIL PROTECTED]>:
>
>>> Why would I want to run more than a dozen processes?
>>Many, many reasons.
>
>Name one.
I run more than a dozen processes all of the time, because I can.
>
------------------------------
From: [EMAIL PROTECTED] (JEDIDIAH)
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Claims of Windows supporting old applications are reflecting reality or
fantasy?
Date: Tue, 20 Jun 2000 10:29:50 GMT
On Tue, 20 Jun 2000 09:50:14 GMT, John Wiltshire <[EMAIL PROTECTED]> wrote:
>On Mon, 19 Jun 2000 17:06:21 GMT, [EMAIL PROTECTED] (JEDIDIAH)
>wrote:
>
>>On Mon, 19 Jun 2000 11:13:55 GMT, John Wiltshire <[EMAIL PROTECTED]> wrote:
>>>On Sun, 18 Jun 2000 16:20:31 GMT, [EMAIL PROTECTED] (JEDIDIAH)
>>>wrote:
>>>
>>>>On Sat, 17 Jun 2000 04:36:23 GMT, John Wiltshire <[EMAIL PROTECTED]> wrote:
>>>
>>>[snip]
>>>
>>>>>The NT 3.x is quite similar in design to Linux/XF4 if you look at it.
>>>>>Wonder how long it takes them to move X into the kernel to improve
>>>>>speed? ;-)
>>>>
>>>> Even if that happened, our kernel is modular. The wise sysadmin
>>>> would still be able to rip it back out if necessary. So, the
>>>> whole point is moot.
>>>
>>>Not necessarily. Just as you practically need perl, python and half a
>>>dozen other scripting packages to run a full featured Linux
>>>installation, so you will probably end up needing the GUI.
>>
>> This is simply false.
>
>Have you ever installed a minimal Red Hat configuration (ie less than
>100M install) and then figured how much junk you need to get XFree86
>4.0 with Gnome or KDE to run? I know it's not false because I've done
So? This still does not demonstrate the necessity of the
GUI on an arbitrary Unix machine.
>exactly that a few times and been left each time wondering how Linux
>zealots can call Windows bloated.
Actually, you can get a completely functional Redhat 6.x
'development workstation' install in about 300M. Also,
KDE, GNOME and X all have components that are entirely
optional.
However, that is also irrelevant as one is not merely limited
to Redhat or 100M disks. Even in the 386 era, harddisks were
larger than that.
>
>>>>[deletia]
>>>>>>>That's because real-mode never made its way into Windows v3.x.
>>>>>>>In order for Microsoft to move forward, they had to leave some
>>>>>>
>>>>>> It doesn't matter what the excuse is. The 386 was out by then,
>>>>>> they had more than enough information to plan ahead with. They
>>>>>> just chose not to.
>>>>>
>>>>>The 386 was out, but the target platform was the 286. Windows 3.x
>>>>
>>>> So, they could have designed it with both the future and the
>>>> present in mind.
>>>
>>>How? You obviously have some idea, don't you?
>>
>> Are you implying that I am some sort of CIS genius and that
>> if I can't come up with the solution than somehow Microsoft
>> is magically absolved of any responsibility for not coming
>> up with a solution of their own?
>
>No, but if you don't know if a solution is possible then how can you
>criticize Microsoft? By your very criticism you are implying that you
The whole job of an Operating system is to abstract the hardware
details away from the applications. I can't imagine why a graphical
application would even need to be so tied to the underlying interface
that it would break on a future revision of the same microprocessor
family.
>believe a solution is possible. If you can't detail that solution,
>then from what does your criticism stem?
>
>It's like saying Henry Ford should have built a better car, but having
>absolutely nothing to show that it was possible at the time.
...except there are examples of others in the same era being
able to deliver the sort of genuine legacy support I am
describing: most notably IBM.
>
>>[deletia]
>>>>>> No, they should design for the future more than the have
>>>>>> been (in the case of Microsoft). Software doesn't wear
>>>>>> out and OS vendors shouldn't be essentially sabotaging the
>>>>>> capital investments of both companies and home users.
>>>>>
>>>>>So Linux should have a standard binary driver API, or do different
>>>>>rules apply to different systems?
>>>>
>>>> You're trying to change the subject.
>>>
>>>No, I'm just saying that what's good for the goose is good for the
>>>gander. If you want old binaries and DOS apps (which required
>>
>>You are trying to support the fallacy that Application == Device Driver.
>>
>> This would be a good example of the "false strawman argument".
>
>Tell that to the Debian/HURD team, or the Mach guys.
They're not a part of the argument. That's why it's a false
strawman. You're merely bringing it up to distract the
discussion from the Windows fault in question.
[deletia]
--
If you know what you want done, it is quite often more useful to
tell the machine what you want it to do rather than merely having
the machine tell you what you are allowed to do.
|||
/ | \
Need sane PPP docs? Try penguin.lvcm.com.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Number of Linux Users
Reply-To: [EMAIL PROTECTED]
Date: Tue, 20 Jun 2000 10:40:38 GMT
Andrew <[EMAIL PROTECTED]> writes:
>Or the reverse. I have at least 10 distros sitting on CDs on my bookshelf and
>another 3 at work. The only one I have installed is a copy of RH 6.1. (yeah, I
>know, I know)
Could someone explain to me why people seem to have this weird urge to
collect Linux distributions on their bookshelves?
Bernie
--
Any fool can make a rule, and any fool will mind it
Henry David Thoreau
American writer, 1817-62
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: An Example of how not to benchmark
Reply-To: [EMAIL PROTECTED]
Date: Tue, 20 Jun 2000 10:40:39 GMT
[EMAIL PROTECTED] (Pete Goodwin) writes:
>OK, so what happens if I run the official version on Windows with Threshold
>of 25, as the Linux would run with by default (and as did my VC version).
>I get 28 minutes 30 seconds with the official Windows version.
That doesn't make *any* sense. That would mean that going from Threshold 3
to Threshold 25 adds less than 6 minutes under Windows on your Windows
installation.
Meanwhile, on my generally faster Celeron400, it adds more than 11 minutes.
Do you have an explanation?
>>Only if that "ought" to be what I am out to prove, which would lead me
>>to post a whole shitload of numbers in the hope of confusing the point
>>that I still haven't done a proper comparison.
>Are you implying that's what I was trying to do?
Well, Pete, don't take it the wrong way --- but it certainly appeared that
way. You *were* throwing a whole load of numbers around, which, once
properly inserted into the table, revealed mainly that you hadn't done
(and reported) any proper comparison.
And, quite frankly, I don't buy that 28:30 as being the result of the
official Windows version, running the same render with the same settings.
You'd have to come up with a pretty impressive explanation for why
it wouldn't fit the usual "slightly slower than the C400" pattern....
>And I've come to the opposite conclusion, which leads me to suspect there's
>something else going on here - how come your Linux version on a Celeron 400
>is running faster than mine on a PII 400MHz?
It's not just the Linux version --- it's also the Windows version, don't
forget that; 20:58 vs 22:51 for Threshold 3.
>In the end I did do comparisions with like settings and I get very
>different results from you.
You allegedly did *one* comparison (of the obvious two), and as I said,
I don't buy one of the timings --- it falls completely outside all the
other timings.
>>>My long experience with VC tells me it is stable.
>>
>>Then how do you explain the comment the good people who maintain Povray
>>put into the Readme?
>I don't. It depends what they're building - I mean a full blown GUI version
>or a CLI version. I'm very surprised they consider it unstable.
I see. In other words, you have no indication whatsoever that the problems
they allude to don't exist in your compile.
Bernie
--
Computers are useless.
They can only give you answers.
Pablo Picasso
------------------------------
From: [EMAIL PROTECTED] (JEDIDIAH)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Number of Linux Users
Date: Tue, 20 Jun 2000 11:04:46 GMT
On Tue, 20 Jun 2000 10:40:38 GMT, [EMAIL PROTECTED] <[EMAIL PROTECTED]>
wrote:
>Andrew <[EMAIL PROTECTED]> writes:
>
>>Or the reverse. I have at least 10 distros sitting on CDs on my bookshelf and
>>another 3 at work. The only one I have installed is a copy of RH 6.1. (yeah, I
>>know, I know)
>
>Could someone explain to me why people seem to have this weird urge to
>collect Linux distributions on their bookshelves?
Mebbe they feel that they are too wealthy...
[deletia]
--
If you know what you want done, it is quite often more useful to
tell the machine what you want it to do rather than merely having
the machine tell you what you are allowed to do.
|||
/ | \
Need sane PPP docs? Try penguin.lvcm.com.
------------------------------
** 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 (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
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
******************************
