Linux-Advocacy Digest #460, Volume #25 Wed, 1 Mar 00 18:13:11 EST
Contents:
Re: Free Internet denied to Linux users (Mig Mig)
Re: Microsoft migrates Hotmail to W2K ("Drestin Black")
Re: w64k - the bugs are being found ("Drestin Black")
Re: My Windows 2000 experience (George Richard Russell)
Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K) ("Mr. Rupert")
Re: Microsoft migrates Hotmail to W2K (George Marengo)
Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K) ("Chad Myers")
Re: Why waste time on Linux? (Matt Hucke)
Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K) (Stefan
Ohlsson)
Re: Microsoft's New Motto (Nick Manka)
----------------------------------------------------------------------------
From: Mig Mig <[EMAIL PROTECTED]>
Subject: Re: Free Internet denied to Linux users
Date: Wed, 1 Mar 2000 22:18:35 +0100
This is nonsens.....
As long as you have the necessary data you can connect to anything with
Linux/FreeBSD/Whatever.
Find the necessary info on the CD if techsupport does not want to help you
- its certainly there
Greetings
John Culleton wrote:
> I received another free internet offer today, this time
> from Juno. As usual Windows 95 or later is required. It's not
> like I haven't paid my dues to Bill Gates. I bought Windows 3.1
> years ago. It won't work and of course Linux won't work. Thus far
> I have free cdroms from Juno, Freei.net, and bluelight.com (KMART).
> Freei.net has Apple software, the other two are Win9x or Win2k only.
>
> The first free provider to include Linux as a supported or at least allowed
> user operating system has a monopoly on the Linux market for this kind of
> service. Whay hasn't any of the bright boys and girls who run these services
> woken up to that fact?
>
> John Culleton
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
------------------------------
From: "Drestin Black" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Microsoft migrates Hotmail to W2K
Date: Wed, 1 Mar 2000 16:25:08 -0500
"George Marengo" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Wed, 1 Mar 2000 14:10:38 -0500, "Drestin Black"
> <[EMAIL PROTECTED]> wrote:
> <snip>
> >I think you would be wrong. Again, as the TPC benchmarks show,
> >using less processors and less machines, the Compaq/MS solution
> >smoked various Sun solutions.
>
> You're right, but that's a single benchmark.
No, more than one. Many more, but it just happens that both of the two most
recent submissions BOTH beat every other machine ever submitted. And I do
see every other OS/hardware represented.
Meanwhile, if we look
> at http://www.top500.org/lists/TOP500List.php3?Y=1999&M=11
> we find that a Linux Beowolf cluster called CPlant is number 44
> on the list.
>
> Maybe I just missed it, but I didn't notice any Windows based
> machines on that list.
No, I don't see any either. Perhaps none have been submitted.
This reminds me of a talk I had with someone who refused to list TPC
benchmark results because there aren't any linux results. The threads in
here regarding TPC constantly reminded us that there were no Linux results
in TPC so it's impossible to make any comparisons to how Linux _might_
perform until there is an actual test. SO, unless you are saying Windows
machines TRIED to make the top 500 list and failed - it says nothing that
they are not there if they never tried. I do not see the OS type listed at
this URL and didn't take the time to check them all. How did you determine
there were no Windows clusters in the list?
Until I know of a windows cluster that ran the linpack benchmark and posted
a result - I can't say what I think of this benchmark.
------------------------------
From: "Drestin Black" <[EMAIL PROTECTED]>
Subject: Re: w64k - the bugs are being found
Date: Wed, 1 Mar 2000 16:53:41 -0500
<[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> From bugtraq:
>
> > I'm running Windows 2000 Pro on a FAT32 file system. I've got a
> > program that creates a very large number of temp files in a single
> > directory. Somewhere between 25000 and 30000 files the system
> > reboots. After the reboot I noticed file system corruption. (It
> > toasted my mailbox files. This may have been due to having eudora
> > running at the time of the crash.)
There is 1... maybe. not much detail ... but, ok, 1.
>
> and another one:
>
sounds like another one, kinda. I mean, what kinda weirdo creates zillions
of 0 byte files to expand to 768 bytes?
But... ok, 2.
Wow, GREAT! You're up to 2 in 2 weeks!
------------------------------
From: [EMAIL PROTECTED] (George Richard Russell)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: My Windows 2000 experience
Reply-To: [EMAIL PROTECTED]
Date: Wed, 01 Mar 2000 22:09:52 GMT
On Thu, 2 Mar 2000 06:18:42 +1000, Christopher Smith <[EMAIL PROTECTED]> wrote:
>
>"JEDIDIAH" <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> On Thu, 2 Mar 2000 04:10:11 +1000, Christopher Smith <[EMAIL PROTECTED]>
>wrote:
>> >
>> >> See, everyone keeps saying this and I actually have never once
>> >> seen this happen.
>> >
>> >Fire up X with a reasonably complex WM and some svgalib program like
>squake.
>> >Flick between the X and squake VTs until the machine locks.
>>
>> This is a contrived example that merely demonstrates that
>> trying to bit bang the same hardware with two root mode
>> apps concurrently is a stupid idea.
>
>The issue being whether or not X could crash the system IIRC. It can.
>
>Despite your accusation of "contrived", it was something I stumbled upon
>quite innocently switching between squake and X.
fwiw, its fixed in later svgalib and xfree versions.
At least, I could do it on Slackware 3.0, but not now.
George Russell
--
One ring to bring them all, and in the darkness bind them.
Lord of the Rings, J.R.R.Tolkien
Hey you, what do you see? Something beautiful, something free?
The Beautiful People, Marilyn Manson
------------------------------
From: "Mr. Rupert" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K)
Date: Wed, 01 Mar 2000 16:24:23 -0600
Both Chad's and Drestin's rebuttal argument to Joseph's post is
absolutely moot. MS is runnig UNIX at their hotmail site.
That speaks volumes for NT and W2K. End of story!
---
The always tidy, always good smelling, and highly presentable,
Mr Rupert
Drestin Black wrote:
>
> "Joseph T. Adams" <[EMAIL PROTECTED]> wrote in message
> news:89jl4a$alg$[EMAIL PROTECTED]...
> >
> > I am a Linux user and advocate, and a firm critic of Microsoft (I
> > consider it a criminal organization, and do not voluntarily use any of
> > its products).
>
> With such an introduction (criminal?!) we take note of your serious bias and
> prepare ourselves:
>
> >
> > I too am amused by Microsoft's failure thus far to successfully port
> > its Hotmail application to NT or W2K.
>
> <grin> Smile and the world smiles along with you, someone once said.
>
> >
> > However, I don't think it's reasonable to insist that the reason for
> > this lack of success is that NT/W2K can't handle the load. I think it
> > could, but only at a very high cost, which Microsoft has deemed
> > unacceptable given the relatively minimal benefits that a succesful
> > migration would have offered.
>
> Your first sentence is spot on. Then you turn to pure speculation and then
> presume to speak for MS and their motives and thoughts. With MS being one of
> the richest corps. in the known universe, I find it funny that you should
> choose to examine this based on cost - something MS has never shown any
> hesitation to ignore and proceed. Cost is something MS is not afraid of.
> Cost is the last thing that would hold them back. All of this, in my humble
> opinion of course. I wouldn't presume to know what someone/thing else
> thinks.
>
> >
> > Clusters of NT boxes manage some extremely busy sites, with acceptable
> > reliability and performance (since NT clustering solutions take into
> > account NT's tendency to fall down under load, and arrange for others
> > to take over with minimal disruption if one should barf).
>
> Well, you tried to muster some truth but couldn't help but fall back to an
> attempt at some insult. It weakens you point and I won't waste time
> educating you. This is old ground, already covered. BUT, safe to say we both
> agree that NT boxes manage some extremely busy sites (I might even say, some
> of THE busiest. In fact, at a recent demonstration (W2K rollout), they
> showed Windows 2000 running a site delivering 1.6 billion hits/day - more
> than all the e-commerce in 1999 squished into 2 days)
>
> >
> > But NT is very different than Unix. While porting well-written apps
> > from one Unix variant to another is relatively simple, porting those
> > apps to NT (or, as many are starting to discover, porting NT apps to
> > Unix) is not. Even for projects of relatively modest size, a totally
> > new design and a total rewrite of all of the code are usually
> > required.
>
> To go from Unix to NT usually requires a very significant if not complete
> rewrite of code. I'll grant you that. However, MOST of the time when you
> convert from one OS to another completely different one - you'll do so not
> only for the OS change but the opportunity to rewrite your app (using
> knowledge you've gained and to incorporate new features and new advances in
> programming languages and technologies). But - yes, NT is very different
> from Unix and porting between the two requires a of rewriting. I could
> equally point out that going from NT to Unix would usually involve a total
> rewrite of all the code. If going only from Unix to NT was a problem, then
> you might suggest that NT (as the end OS) is the troublesome partner - but
> since it's the same in both directions your point is rather a wash and moot.
>
> >
> > Also, NT and W2K run only on a single hardware platform, one designed
> > primarily for workstation rather than server use.
>
> Well, "primarily" for a workstation? True. However, you cannot deny the
> server class hardware created for NT and more so for W2K these days. HP,
> Unisys, IBM, Compaq and others are creating some serious machines for W2K.
>
> > The
> > cost-effectiveness of x86 servers tends to decline dramatically as
> > the load per box grows past a certain point; improvements in the
> > platform over time continue to push that point of diminishing returns
> > upward, but at the present time, neither NT nor W2K, nor any other
> > x86-based OS, is capable of competing on the high end against
> > platforms that are optimized for better hardware.
>
> I must strongly disagree. There are benchmarks, most recently the very huge
> success of the TPC benchmarks and SAP benchmarks and peoplesoft benchmarks
> showing HUGE HUGE leads using x86 servers on W2K over hardware of every
> other vendor and *nix in every blend.
>
> How can you make your statement in the face of x86 wins of over 67% better
> performance using hardware over 66% less expensive!
>
> >
> > The bigger the application, the more costly PC-based solutions become.
> > PCs are not and probably cannot be competitive as high-end servers,
> > because of hardware limitations that are intrinsic to the platform.
>
> I disagree. What intrinsic limitations?
>
> > You can only get lots of MIPS or TPS by adding more boxes, and as you
> > do, the cost of communicating and coordinating work done by those
> > boxes can grow almost exponentially unless the app is specifically
> > designed for that kind of architecture, so even adding more boxes
> > doesn't make clusters of PCs viable as high-end servers UNLESS - and
> > this is they key, the reason why both Linux and NT clusters can either
> > succeed or fail - unless the app is designed, from the ground up, to
> > run on the specific OS, hardware, and network configuration that is to
> > be used.
>
> Well, you are talking about the app here, being specifically designed for
> the task. Who can argue with such a thing. I may disagree with a part of
> what you wrote here but we can let it stand as it doesn't affect the OS
> discussion I thought you were having.
>
> >
> > OTOH, you can buy or build UNIX machines to be as big and fast and
> > reliable as you want them to be, and the cost increases only linearly
> > with performance (maybe even less), not geometrically.
>
> I very very strongly disagree with you assesment of costs. Can you show me
> anywhere a Unix solution that came in cheaper than a x86 solution? I mean,
> again, back to our TPC benchmarks. Look at those prices for Unix boxes that
> can't even produce the same performance. Do you not see that ALL 40 of the
> top 40 positions for price/performance are filled by x86 machines (running
> NT and SQL Server)? How can you claim UNIX boxes to be faster and less
> expensive in light of hard data that says exactly the opposite?
>
> >
> > Thus, my best educated guess is that when MS took a *realistic* look
> > at what porting Hotmail to NT would involve - a total, ground-up
> > rewrite starting yesterday, plus the ongoing cost of maintaining the
> > world's largest parallel server farm, one dozens if not hundreds of
> > times bigger than the one that runs www.microsoft.com - they decided
> > that the cost simply wasn't worth it.
>
> Back to school: #1) microsoft.com generates way more traffic than
> hotmail.com - they can do and have been doing it forever using NT on a
> server farm far less grand than "world's largest" without a hiccup. #2) they
> have no problem with cost, this is MS who has more *cash* than many of it's
> competitors have in total equity. #3) Why would a total rewrite frighten MS?
> It would give them something else to boast about: "The Unix app couldn't
> handle it so we had to rewrite it using our own products"
>
> I do not pretend to know the real reasons why MS did not immediately take
> teh property that was already working and rip it up and port it to another
> OS and language. I can _guess_ that they said: "Hey, we bought a hot
> property. It works. Cool, let's let it do it's thing and see if it survives
> and becomes something we like. Then we'll consider converting it, if it
> needs it." I can *guess* that recently they thought, "Hey, hotmail is doing
> well. W2K is being released. Say, let's port over to W2K and announce yet
> another high volume high profile site running W2K." And, as I've said
> before, I expect to see Hotmail.com running W2K before 2000 is out. Those
> are my educated guesses. But, they are only guesses, which, like yours,
> could very easily be right or wrong. Only time will tell.
>
> >
> > In other words, Unix was the more cost-effective solution. As tends
> > to be the case for *most* enterprise-class applications.
>
> That does NOT follow for many reasons, some of which I've identified in my
> reply. Unix has not been the more cost-effective solution for some time. My
> company bids against Unix solutions constantly. We are ALWAYS the lower bid
> and usually do not have to worry about defending our pricing but instead
> have to prove why our non-unix solution should be picked over the more
> expensive unix solution. See, it's intersting. Many of our clients are
> actually willing to pay more for a solution if it's better. They are not
> driven solely by price. But, it sure helps when we have a product which we
> can say: "does everything theirs does (usually more) AND costs less." I've
> not yet walked into a meeting where a potential client said: "These other
> guys have a unix based proposal that costs less, why should we pay more for
> NT?" Never happened. Ever.
>
> >
> > But this does not mean NT or W2K couldn't have handled the load. I'm
> > sure that given enough cost and time, it could and would have done so.
> > And that basically is my point.
>
> I think you've strayed from your point but, yes, I agree that NT (W2K is
> NT5) could (and will) handle the hotmail.com load but I disagree that it
> would be excessively costly.
>
> >
> > In the future, Windows-based solutions will probably become more
> > cost-competitive, especially at the low end. Indeed, if it weren't
> > for Linux and the free *BSDs, NT would be the most cost-effective
> > platform for building small, data-centric Web sites today.
>
> The price of a NT webserver and a Linux webserver is very very similar. The
> price of the OS barely affects the bottom line (especially for a web server
> which doesn't need lots of CALs)
>
> >
> > But Microsoft solutions have never been terribly scalable, and as the
> > size of a task grows larger, Unix-based solutions become more and more
> > attractive.
>
> I disagree.
>
> >
> > Hotmail is one of the busiest sites in the world. It is thus the
> > *last* place I would expect to find an all-NT solution.
>
> Hotmail is not THAT busy and isn't it true you always find what you are
> looking for in the last place you look? :)
>
> >
> > Microsoft can't be blamed for not doing the migration. They can of
> > course be blamed for promising to do so - a promise which clearly
> > didn't reflect very much thought - and they certainly are not only
> > wrong, but criminally wrong, for advertising NT or W2K as being
> > reliable, scalable, enterprise-class platforms.
>
> You are wrong. Especially repeating your inane "criminal" claims. NT has
> already proven itself to be reliable, scalable and enterprise-class by being
> and doing these things in real companies in the real world today. You
> personally may not have wittnessed such things but fortunately a lot of
> others have and many run their entire business on the reliability of NT
> servers and are still here to prove it. It's hard to claim criminal
> misrepresentation when you can bring in witnesses who can certify and
> demonstrate the claims to be true.
>
> > By the standards of
> > any competing platform, they certainly are not. There are several
> > market niches that they do fill very nicely, but these are at the low
> > end, not the high end.
>
> Your opinion, I do not agree.
>
> >
> > For those willing to even consider a non-Microsoft solution, Linux and
> > the *BSDs are killing NT at the low end; commercial Unix kills it at the
> > high end. Microsoft is well aware of this, and is working on carving
> > out a niche somewhere in the middle that it can defend. It would be a
> > mistake to underestimate Microsoft's ability to adapt to changing
> > conditions. Once it realizes that it can't make a one-size fits all
> > OS and that nobody really needs another one anyway - UNIX in all its
> > flavors is already as close as anything needs to be - it will find and
> > do its damnedest to fill whatever market niches it thinks it can. The
> > results should be interesting.
> >
>
> Your opinion, I disagree.
------------------------------
From: George Marengo <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Microsoft migrates Hotmail to W2K
Date: Wed, 01 Mar 2000 22:33:29 GMT
On Wed, 1 Mar 2000 16:25:08 -0500, "Drestin Black"
<[EMAIL PROTECTED]> wrote:
>
>"George Marengo" <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> On Wed, 1 Mar 2000 14:10:38 -0500, "Drestin Black"
>> <[EMAIL PROTECTED]> wrote:
>> <snip>
>> >I think you would be wrong. Again, as the TPC benchmarks show,
>> >using less processors and less machines, the Compaq/MS solution
>> >smoked various Sun solutions.
>>
>> You're right, but that's a single benchmark.
>
>No, more than one. Many more, but it just happens that both of the two most
>recent submissions BOTH beat every other machine ever submitted. And I do
>see every other OS/hardware represented.
Yes, there are several TPC-C entries, but the point is that TPC-C
isn't the be-all, end-all benchmark --- it's only one of several.
>Meanwhile, if we look
>> at http://www.top500.org/lists/TOP500List.php3?Y=1999&M=11
>> we find that a Linux Beowolf cluster called CPlant is number 44
>> on the list.
>>
>> Maybe I just missed it, but I didn't notice any Windows based
>> machines on that list.
>
>No, I don't see any either. Perhaps none have been submitted.
Perhaps. Perhaps no Linux cluster have been submitted for TPC-C
>This reminds me of a talk I had with someone who refused to list TPC
>benchmark results because there aren't any linux results. The threads in
>here regarding TPC constantly reminded us that there were no Linux results
>in TPC so it's impossible to make any comparisons to how Linux _might_
>perform until there is an actual test. SO, unless you are saying Windows
>machines TRIED to make the top 500 list and failed - it says nothing that
>they are not there if they never tried.
That's right, it says no more about NT than TPC-C says about Linux.
We don't know if there were even any submissions.
>I do not see the OS type listed at this URL and didn't take the time to
>check them all. How did you determine there were no Windows clusters
>in the list?
By looking at the hardware and for ones that might be running Windows,
going to the site and getting any listed specifics.
For instance, the #1 spot is held by Intel's ASCI Red so I think,
hmm... that might be running Widows, so I go to the site. Nope,
it's running Unix.
#2 is running on a 604e processor, #3 is running on SGI, etc.
A possible is #34, running on a Compaq Alpha... hmm... maybe
that's NT.
I go to the web site and an article about their supercomputer
mentions the Tru64 OS, not NT.
There are only a few on the list that aren't running SGI, Cray,
NEC, PPC 603/604e -- #265 is another self-made cluster.
NT maybe? Nope, another Linux Beowolf cluster.
As I said, I could have missed a result obtained on NT, but I
did look for it.
>Until I know of a windows cluster that ran the linpack benchmark and
>posted a result - I can't say what I think of this benchmark.
Maybe it's already been done, but it just didn't make it into the
Top500 list.
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K)
Date: Wed, 1 Mar 2000 16:36:18 -0600
"Mr. Rupert" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
>
>
> Both Chad's and Drestin's rebuttal argument to Joseph's post is
> absolutely moot. MS is runnig UNIX at their hotmail site.
-- because that's what it was running when they bought it.
> That speaks volumes for NT and W2K. End of story!
It says that MS wasn't stupid enough to try to pick up a
half-way working system to move it to NT, only to move it
to Win2K again in a few months.
-Chad-
Have you recompiled your kernel today?
------------------------------
From: [EMAIL PROTECTED] (Matt Hucke)
Subject: Re: Why waste time on Linux?
Date: 1 Mar 2000 22:38:27 GMT
In article <89htuc$oel$[EMAIL PROTECTED]>,
proculous <[EMAIL PROTECTED]> wrote:
> When there are so many great windows and mcintosh programs out their
>what is the point of wasting time on a build it as you go along system?
For the same reason some people fix their cars themselves, or use
the manual controls on their cameras, or make a roux the old-fashioned
way - it's fun. Linux fans enjoy learning about the machines and
becoming familiar with how the OS works.
--
Matt Hucke ([EMAIL PROTECTED]) |
Cynico Network Consulting | Microsoft delenda est.
Author, "Graveyards of Chicago" |
http://www.graveyards.com/ |
------------------------------
From: [EMAIL PROTECTED] (Stefan Ohlsson)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Fairness to Winvocates (was Re: Microsoft migrates Hotmail to W2K)
Reply-To: Stefan Ohlsson <[EMAIL PROTECTED]>
Date: 1 Mar 2000 23:46:02 +0100
Chad Myers wrote:
>"Mr. Rupert" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
>> Both Chad's and Drestin's rebuttal argument to Joseph's post is
>> absolutely moot. MS is runnig UNIX at their hotmail site.
>-- because that's what it was running when they bought it.
>
Why doesn't matter.
>>That speaks volumes for NT and W2K. End of story!
>It says that MS wasn't stupid enough to try to pick up a
>half-way working system to move it to NT, only to move it
>to Win2K again in a few months.
>
It says one or more of several possible things.
Is it that MS deemed it too expensive to port?
Is it that MS realized that is wasn't possible?
Did they test and NT couldn't handle the load?
The questions are many but MS (maybe wisely) keeps their mouth shut.
Either way, it's bad PR to run a site with competitor's OSes, because
whatever the exponation might be it gives the impression that their
OS can't handle it.
>-Chad-
>Have you recompiled your kernel today?
>
My kernel's in ROM..
/Stefan
--
[ Stefan Ohlsson ] · http://www.mds.mdh.se/~dal95son/ · [ ICQ# 17519554 ]
Have you succeded in dodging the 63000 bugs today?
------------------------------
From: [EMAIL PROTECTED] (Nick Manka)
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.unix.advocacy
Subject: Re: Microsoft's New Motto
Date: 1 Mar 2000 22:45:24 GMT
Reply-To: [EMAIL PROTECTED]
In article <89jhp9$9c9$[EMAIL PROTECTED]>,
"Neil" <[EMAIL PROTECTED]> writes:
>> Only one active login per machine which is completely tied to the
>> GUI console's context
>
> With vanilla NT, true enough.
> NT TSE, or W2K with Terminal Services, slightly different.
Too bad that's not what's actually be shipped to people in the
general case. Regardless, user context is married to a GUI context
unless you specifically write something as a service.
>> No support for multiple instances of user data and applications in the
>> Registry
>
> Can you expand on this bit?
I have two completely separate instances of a program being run --
if that program does not explicitly provide for that, it will step
on it's registry keys. I can't have Office installed twice. Moreover,
there is an ungodly amount of per-user information stuck in HKLM
and a bunch of user-irrelevant information stored in HKCU, and
without specific "session" support kludged into TSE, HKCU isn't
unique because it's tied to a GUI context.
> Do you mean that different versions of say Office tend to use the same
> hierarchy? 'cos some of that is version specific.
But not all of it, and it certainly isn't instance specific. If
NT was really "multi-user," you could do things like run multiple
domains off a domain controller and be able to run multiple instances
of services.
>> A single flat event namespace in the kernel (they had to do major
>> hacking on TSE to get around this, and still haven't, really)
>
> Can you elaborate on this a little?
Find a book on the architecture of NT. _Inside Windows NT_ is oft
recommended, but by no means complete. There is a flat namespace
that kernel events exist in, which is why TSE has a horrid kludge
to support "sessions" built into it.
>> More-or-less complete trust of the currently logged in user
>
> I guess that rather depends on the rights of the currently logged-in user,
> and ACLs on things that support ACLs.
There is an assumption built into NT that the user at the console
is special relative to any other user at any given point in time.
>> Expectations by programs that they are priviledged (more policy
>> than than design, but still a factor if you actually
>> want to run anything on your NT box)
>
> Surely this is purely an app thing? An issue with the app design, or the
> vendor's implementation?
Like I said, it's more a policy issue, but it still exists. That
developers can solve the problem doesn't mean that users can.
>> Heavy use of needlessly exclusive locks coupled with minimal COW behaviour
>
> Again could you elaborate on this? I'm aware of some of the compromises of
> necessity (well pretty much) within TSE.
In a multi-user system, where multiple concurrent access to resources
are the norm, you can't do what most Windows programs and NT services
do, which is lock files exclusively (even if just for reading).
That NT has virtually no copy-on-write support for most objects
doesn't help any.
>> No user accessible global namespace and lots of indeterminable mappings
>> of resources to namespaces within a user context
>> (ie, network drives and printers)
>
> Can you give some examples of this ie the problems?
Network drives and printers are a per user thing, and there is no
way to specify an absolute path to a generic resource. UNC paths
sort of can, but once again the program has to explicitly support
those by not using any of the hundreds of library calls that don't,
and UNC paths aren't extensible to random backing objects.
> I understand where you're coming from with some of this, but most people
> implementing TSE (IME) knew well about this problems from previous NT 3.51
> implementations (Winframe and the other derivatives), and already had a
> workaround.
Workarounds are bad. Workarounds mean the underlying problem still
exists. I don't pay out the nose for software for the priviledge
of working around its failings. These failings are present in both
single-user NT and TSE, and are likely never going to go away.
Don't get me wrong, NT is an acceptable if not particularly desireable
desktop OS (especially considering the win32 alternatives are Win9x
and WinCE), but not a thoroughly multiuser system.
--
Network Samurai http://www.syncronym.org/~nick/
------------------------------
** 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
******************************
