Windows / Unix server question

This is not another post to start a bashing contest between Windows and Linux adepts, but something from one of my classes... I'm detecting some BS here, but I can't find anything "unbiased" to prove it... Anyone have any opinions?

Here's the post in question:

Quote:

---

While testing Operating Systems for the Justice Department, I worked with engineer's from Microsoft, Data General, BSD, and Sun.

White papers showed that Windows only emulates multitasking, and that is the reason that they only have a 10% share of the server market. The majority of servers are Unix based.(even Microsoft has Unix Servers)

For performance we tested how many web pages each could handle per minute...Windows choked at 50,000 hits per minute, and Unix could handle in access of over 1 Million hits per minute.

People much smarter that me agreed that this was a multitasking issue.

---

I don't buy the "10% share of the server market" for Windows... afaik, it's much higher. For web servers, it's somewhere around 20% according to Netcraft. Makes me wonder how old his "research" is.
And 1 million hits per minute? In a test? How? And what server?...
Anyone care to debunk this? Or confirm so I can go give him some credit? :D

There could be some truth in this mess of information here:

Windows was designed not to be multi user and wasn't designed with security in mind. (Click cancel, PWNED).

NT was originally not designed for networks.... well sort of. Nt wasn't internet ready, it had to me made so. Get the oldest NT you have anywhere and try getting online.

So by this you get, Windows 9X and NT lines at first were not designed for the net.

Unix.... Heh, BSD is what made the internet, and TCP/IP. Linux is a child of the net.

The first Unix was never designed for security. At the time it wasn't needed.

If you think how Microsoft uses the NT line for 2000, XP, 2003, and Even next generation, it becomes a problem if they are just being made for use on a net.

Windows 2000 went a step ahead, it was in fact design for networks. XP was too. But they were still based on NT.

So there could be truth in some of this. Microsoft has innovated virii though. The first Virii were on Unix machines, but Microsoft really took the torch and ran with it here. Unix virii were really just pranks anyway, but they count. All in all, I doubt many people here work on Windows and know for sure about the emulated multi process ****. Making you use a GUI for one would slow down results. On Unix just shut off or don't install X, leave the RAM in your box to answer requests instead of looking pretty.

maybe after some more replies I can talk more but I'm still waking up. Didn't have the greatest day to start.

Well, one of the things you have to consider in any 'study' like this is the configurations that were used....were they straight out of the box, turn the system on and test, or were any attempts made to change how the system works (to either speed it up or secure it). Either way, in alot of comparisions I have seen, you get more of a comparison of how the systems were setup than of the actual performance of the system ... a good example would be that in Solaris version < 8, there were no default tweaks of the TCP/IP stack... and a web server (Apache) running on a default configuration would literally choke under minimal load...a few stack changes later...the performance was orders of magnitude better...Sun's attitude at the time was we prefer to leave it to the system admin to configure the system how they wish...some of which Sun has changed in its later versions...

Regardless, at least with the Solaris systems...alot of the tests I saw at the time were not really reflective of the server's capability...

Can't specifically back up the numbers there, but did want to mention some related things I have personally seen.

But the article isnt talking about Windows, is talking about how Web Server (IIS pehaps?) handle the transactions: using O.S. multitasking or using a monolitic thread (Internal multitasking).
Supposing that is IIS, does anyone knows that is its thread model?
BTW, the usage of O.S. multithreading or "internal" multithreading cant prove per se that a system is better than other. Some Mainframes transaction servers (CICS) is single thread and still rocks :)

Hi,

I was always under the impression that Windows supported multitasking and multithreading.

DOS on the other hand does not.

It seems that there is a lot of information missing...............what applications (16 bit 32 bit?), what processors. what RAM etc. When was the test conducted? I can remember when Sun kit would hammer a PC running Windows, if only because it had a much faster processor. We used to use them for CAD applications.

I would also have thought that MS has a bigger share, particularly of applications servers?

My guess is the article refers to about 1996?

:)

Quote:

---

Hi,

I was always under the impression that Windows supported multitasking and multithreading.

DOS on the other hand does not.

---

My understanding was that Windows didn't support true multithreading until NT4 and Win95, and that even then the implementation was lacking (changes that Win2K/XP/2003 have improved significantly upon)...which would BTW put your timing dead on :)

I also agree, there is quite a bit of information missing ...

My understanding is Windows 2003 and the Windows.NET applications\technology are true multi "tasking".

Although could all be just sales hype too.


MLF

Morgan:

2003 is the best version of Windows I think I've ever used for things like production. It's very fast as well and noticably faster than 2000 on the exact same box. without tweaking. So hype or not it does have something the others don't.

Yes ...I agree with you gore...it is a HUGE improvement.

And very stable....have only had very minor issues easy to resolve...although alot of "hotfixes" :rolleyes: ...

Have a 2000 server...that I am trying to convince the customer to upgrade as we speak...so I will keep my eye on this thread.

2000, heh, great for a desktop but server? Not in my opinion. One might argue it's been tested enough to be stable now and now that it's 5 years old it's got most bug worked out. But then again, when the support runs out....

This is why I always recommend SUSE Linux, it's much cheaper and gives peopel a chance to work in a Unix. Heh and then there is that look on a face you get when someone first looks at YAST. From DOS to Windows to Be to Unix to Linux and any OS, it's the best tool I've used, and even easier than Windows.

Quote:

---

My understanding was that Windows didn't support true multithreading until NT4 and Win95, and that even then the implementation was lacking (changes that Win2K/XP/2003 have improved significantly upon)

---

I always thought Windows 95 did not have anywhere near a multitasking system. It emulated, sure, but thats about it though. Infact, im pretty damn sure they never really attempted to make any real steps at all towards a real multitasking OS intil CE came out.

Quote:

---

Originally posted here by ¤The¤Spe©ialist
I always thought Windows 95 did not have anywhere near a multitasking system. It emulated, sure, but thats about it though. Infact, im pretty damn sure they never really attempted to make any real steps at all towards a real multitasking OS intil CE came out.

---

No, it did multitasking...at least far better than its simulated mutlitasking in 3.1... I remember playing with it at the time that it came out...I remember being floored that I could play Doom and download something over my 14.4K modem at the same time...where it had been impossible before in 3.1

:) Ok I just dated myself :)

EDIT: I had seen multitasking work on other OS's at time (solaris and OS/2 warp come to mind), that was just the first time that I saw it work sort of right :) Like I said, there have been dramatic improvements since then...

Pfft, I have a laptop with a 3.06GHz P4 M processor, 512 RAM and a 32 MB Nvidia Ge Force FX GO 5200 card and I play doom every day on it. I can beat quit a bit of it on Nightmare.

From what I've learned, Windows became true multitasking with the release of Windows 2000.
I think the confusion comes from the fact that with a single processor, "true" multi-tasking is not possible (since the processor can't do two things at the same time; it will switch back and forth between two tasks so fast that it seems as it is, though). Since Windows-based computers traditionally only had one processor, it was thought that these systems cannot trule multitask.

And about the server-share: from what I can find, Windows has a 50% market share in the server market (all servers combined), and *nix derivates around 40%. For web servers, it's around 20% for IIS and 70% for Apache, but the majority of Fortune-1000 companies run IIS...

This isn't entirely true...NT4 supported up to 4 processors.

Yes,

Windows 3.1x was basically a GUI riding on the top of DOS. Windows 95 was an attempt at 32 bit multitasking.

As I recall there were two problems:

A lot of Windows programs were non-reentrant, so you would get a rather staccato multitasking. Secondly, if you ran any 16 bit applications it really ruined the multitasking performance.

This is possibly why some people considered it to be "emulation" when it was in fact true multitasking that didn't always work too well.

:)

Quote:

---

Originally posted here by Negative
From what I've learned, Windows became true multitasking with the release of Windows 2000.
I think the confusion comes from the fact that with a single processor, "true" multi-tasking is not possible (since the processor can't do two things at the same time; it will switch back and forth between two tasks so fast that it seems as it is, though). Since Windows-based computers traditionally only had one processor, it was thought that these systems cannot trule multitask.

And about the server-share: from what I can find, Windows has a 50% market share in the server market (all servers combined), and *nix derivates around 40%. For web servers, it's around 20% for IIS and 70% for Apache, but the majority of Fortune-1000 companies run IIS...

---

Neggy:

Most version fo Windows until.... 95 or NT.... Used Cooperative Multi Tasking, which would explain **** performance. Giving apps control and telling them to play fair is stupid. This could have something to do with the romour that it wasn't true multi tasking. If your OS Kenel can't control the apps, you have a problem.

Special:

iddqd

idkfa

Doom 3:

God

give all

Quote:

---

From what I've learned, Windows became true multitasking with the release of Windows 2000.

---

Nope, NT was the first preemptive multitask O.S. Win9x uses preemptive multitasking too but with a **** model, so you can just ignore tha family.
Also, NT supports multiprocessing (more than one physical P.U.)
Starting from NT (ignore that 9x b.s.), Windows is a preemptive multasking O.S. (there is not "true" or "false" - its a preemptive multitask or cooperative (?) multitask), and supports also multithreading - where one task (Adress space, process or whatever you name it) can create child tasks and explore a multiprocessor HW.

Quote:

---

Anyone care to debunk this? Or confirm so I can go give him some credit?

---

It's complete BS.

He need to define "windows" in fact this reads circa 1995

Here's his explanation:

Quote:

---

The year was 1998. The testing was for new technology for the next generation of Enterprise servers for the Justice department...

"ALL" of the technology was beta technology I witnessed at each of these companies R&D center, to determine the next generation of systems. Now at that time Microsoft was beta testing NT 5, which later was renamed "XP". The statistics showed that Windows was no match for Unix as a server...As a matter of fact Microsoft uses Unix servers, and is developing their own version of Unix to be a player in that market.

Now Windows has grown leaps and bounds since 1995...and could be the IT solution for a lot of companies, but I have been blessed to work on a lot of projects, where money was no object, and the price difference between systems was justified by performance and usage.

Windows does not have a OS that would run on the equivalent of a Sun E-10000...Which is one of the most powerful systems...And they still have not found a solution for systems with over 32 processors. They are 100% idle, the OS can not access them.

Now A lot of companies can not afford a server of this quality or price, and could use Windows and their products for their IT solution, but it won't be on government servers...cost analysts say for that kind of traffic, you will use Unix.

---

IIRC NT5 is Windows 2000. Windows XP is NT 5.1.

Preemtive multitasking means the OS designates which process (or thread or whatever) runs on a particular processor. This scheduling is the whole trick. Scheduling on XP still sucks IMO. Why? I keep hitting the same thing.. Processor indicates 99% idle but the OS doesn't respond to clicks, keys or whatever... Until after a few seconds then everything happens in the order I clicked... I never ever experience that on my BSD machine..

Hotmail used to run completely on BSD I believe. Then MS bought it. They tried to migrate to Windows 2000 but had to stop because W2K couldn't handle it. AFAIK now only the front-ends are windows machines but the backend is still bsd.

Ah. Almost forgot. Locking is another big issue.

I know FreeBSD. They used to have a giant lock. Meaning if 1 processor accessed the kernel the other proc had to wait if it also wanted access to the kernel. With 5.x (SMPng) they're using a more finely grained locking mechanism. This has improved the performance on SMP systems dramaticly.

So the biggest factors IMO are scheduling and locking, from an OS point of view.

So Negative,

He was comparing a non-server beta os to a unix deployment? And a sun at that. OK XP loses. lol. I don't think anyone would argue the speed of a unix installation versus a windows installation on similar hardware back in what was it... 1998? Traditionally Unix machines where faster because they ran on high end hardware as compared to small deployments of Windows. Unix was developed down one path, windows down the other. Server versus workstation. Shoot the fastest decently priced mini frames were Sun and Data General Unix. Both extremely expensive compared to Windows. Hundreds of times the cost. No make that thousands.

Now when we go into this realm of a supercomputer with 30 some processors? Well that's apples and oranges. Since moors law is in effect perhaps some research into a modern reality would be better considering it's 2005. :D Outside of the super computer realm where I would wager unix is king or some of the grossly proprietary OS', when you compare Unix and windows on the same harware they vary in performance based on what you are doing.

Once you cross a threshold though Unix is appropriate without rival. But you pay out the ass for that. In fact many older unix deployments often get replaced by clusters of windows machines that rival the speed at a fraction of the cost.

And as I said ;) :

The Internet : Built for Unix,

Windows : Built for the Internet

Unix and Unix like OSs are king in this area. Networking, TCP/IP, and almost every protocol, was written for Use on a Unix platform. Windows has caught up pretty quick, but they still aren't my first choice for servers.

Costs:

Think of something like this:

You buy a barebones computer and load it up with like 4 GBs of RAM, and dual Xeon Processors. You spent a lot on this and need an OS to go on it.

You can download Slackware or pay for Windows Server. Which one makes that total cost lower? When Microsoft say TCO is smaller, it's because they assume people need to be trained to use Linux, they assume they already know Windows, but what if they already know Linux? Then they have to be trained in XP.

Let me tell you, Slackware, SUSE, and Free BSD would be FAST on that hardware. Though Free BSD MIGHT need to be tweaked, the default Kernel is for 486 CPUs and older hardware, when you run it on high end you have to do another.

gore gore gore....

You and I both know the latest hardware will have issues with Suse. So in respect to cutting edge, faulty drivers on linux could cause more than a performance loss. It could corrupt data. Same goes for windows when driver support is lacking, but in general a harware manufacturer would be better at supporting new windows drivers versus linux. And while on the subject linux is not Unix. TCP/IP was built to be independent of the OS layer. And the internet was not built for unix, it was built ON Unix. That's all there was. We were all experts at the time. You couldn't even connect to the internet with early versions of Windows. Billy missed the boat originally and windows sockets that supported TCP were 3rd party. Trumpet Winsock was the king. And windows did catch on quick, it was 1993 or 1994 the MS release their own TCP/IP sockets, but 3rd pary drivers were better.

Windows was built so that a CPA or Author or Small Business owner could use a computer. Unix was only affordable by large corporations and even then they often outsourced the expense, that is how Ross Perot made his money, selling data like payroll, customer databases, you know infromation storage The nice tools that come along with Linux now were nowhere to be seen. Both have come a long way. What made windows so popular? You could put in a card and actually use it with only a little tweaking versus spending a week installing specific and propietary hardware. It was "open", anyone could make stuff for it. Unlike Data General and Sun and Apple. Isn't that ironic? We applaud Apple now but they would not release any compatabilty hardware UNLESS you licensed it through them versus the PC which had no restrictions because MS was free to provide the OS from harware vendors, especially IBM. Which was the corner stone of Unix mainframes colored Blue and sold everywhere.

When MS speaks of TCO they are speaking to "Total", meaning support, hardware maintenance, licensing, application development, etc. Out of the box, Linux is cheaper and in general performs better. But there is a higher overhead of continued maintenance depending on the application set. Now if you already have in house expertise then that cost is lower so there are many variables. But then you limit yourself to a single point of risk. In your example windows would be fast as well. But definitely would cost considerably more, unless you had to spend hundreds of hours on the linux box tweaking it.

Even the initial install is more, if you are paying someone to do it like a technician. Probably 2 or 3 hours more per box. In most cases people will need to be trained to use linux. It's just a fact that many people have never even seen it let alone use it. You would have to analyse your help desk calls to see if the TCO would surpass what is currently the intalled base.

But all in all, especially in the security realm on the server end, away from end users I deploy Linux for one reason... its cheaper. :D It takes more time but time can be washed away easier than initial licensing cost, and in reality I have seen and built linux systems that end up being more expensive of the life cycle of an object.

Quote:

---

Originally posted here by RoadClosed
gore gore gore....

You and I both know the latest hardware will have issues with Suse.

---

My laptop is brand new, and SUSE 8.2 is running on it right now. That's almost 2 years old but works out of the box on a brand new laptop. It depends on what hardware you buy, this is a discussion about servers, SUSE Enteprise is amde for servers, the hardware shouldn't be a problem.

Quote:

---

So in respect to cutting edge, faulty drivers on linux could cause more than a performance loss. It could corrupt data. Same goes for windows when driver support is lacking, but in general a harware manufacturer would be better at supporting new windows drivers versus linux. And while on the subject linux is not Unix. TCP/IP was built to be independent of the OS layer. And the internet was not built for unix, it was built ON Unix. That's all there was. We were all experts at the time. You couldn't even connect to the internet with early versions of Windows. Billy missed the boat originally and windows sockets that supported TCP were 3rd party. Trumpet Winsock was the king. And windows did catch on quick, it was 1993 or 1994 the MS release their own TCP/IP sockets, but 3rd pary drivers were better.

---

IBM is a good friend of Novell, and even Dell and HP support Linux for server purchases. The only one I can think of that I'm not sure about for Linux is Gateway, and who the hell uses them?

And for Windows, their TCP **** was taken from BSD ;)

Quote:

---

Windows was built so that a CPA or Author or Small Business owner could use a computer. Unix was only affordable by large corporations and even then they often outsourced the expense, that is how Ross Perot made his money, selling data like payroll, customer databases, you know infromation storage The nice tools that come along with Linux now were nowhere to be seen.

---

But when Windows first came out, there was no Linux.

Quote:

---

Both have come a long way. What made windows so popular? You could put in a card and actually use it with only a little tweaking versus spending a week installing specific and propietary hardware. It was "open", anyone could make stuff for it.

---

Heh, like when Microsoft tried taking out Netscape for making a browser that worked on Windows and wasn't IE?

Quote:

---

Unlike Data General and Sun and Apple. Isn't that ironic? We applaud Apple now but they would not release any compatabilty hardware UNLESS you licensed it through them versus the PC which had no restrictions because MS was free to provide the OS from harware vendors, especially IBM. Which was the corner stone of Unix mainframes colored Blue and sold everywhere.

---

Before the PC, Microsoft sold Xenix.

Quote:

---

When MS speaks of TCO they are speaking to "Total", meaning support, hardware maintenance, licensing, application development, etc. Out of the box, Linux is cheaper and in general performs better. But there is a higher overhead of continued maintenance depending on the application set. Now if you already have in house expertise then that cost is lower so there are many variables. But then you limit yourself to a single point of risk. In your example windows would be fast as well. But definitely would cost considerably more, unless you had to spend hundreds of hours on the linux box tweaking it.

---

What if it's a company who uses legacy things on 486 machines? Windows won't run on that anymore.

Quote:

---

Even the initial install is more, if you are paying someone to do it like a technician. Probably 2 or 3 hours more per box. In most cases people will need to be trained to use linux. It's just a fact that many people have never even seen it let alone use it. You would have to analyse your help desk calls to see if the TCO would surpass what is currently the intalled base.

---

Umm, you can install on every machine at once. Linux and BSD can be installed from anywhere. All you need is an NFS server, or like with SUSE, you can do the installs over SSH or a VPN and all that needs is 256 RAM. I'd think Windows would take long you actually have to be at each machine from what I understand.

Quote:

---

But all in all, especially in the security realm on the server end, away from end users I deploy Linux for one reason... its cheaper. :D It takes more time but time can be washed away easier than initial licensing cost, and in reality I have seen and built linux systems that end up being more expensive of the life cycle of an object.

---

Don't forget to tell those companies when the next big Worm is laying hell across Windows they don't need to worry ;)

Quote:

---

My laptop is brand new, and SUSE 8.2 is running on it right now. That's almost 2 years old but works out of the box on a brand new laptop. It depends on what hardware you buy, this is a discussion about servers, SUSE Enteprise is amde for servers, the hardware shouldn't be a problem.

---

OK I'll give you a taste of some real world shark bites since I am loading my 3rd Win2k Server this week...

SuSE enterprise corrupted data through a new off the shelf SCSI controller and took out a critical database that encrypts some pretty cool **** to prevent fraudulent documents, that cost my ass and reputation and forced an outage that cost me 10 years off my life. But I still use SuSE. :D It's not a mecca of perfection by any means.

On another project a SuSe network probe keeps locking up after about a week of packet sniffing. Reboot fixes it and to get the entire node up initially, it took about a week of tweaking MySQL and fixing dependencies in YaST. Today I loaded Win2k and SQL and had it up in about 4 hours. In this instance I had to kill a linux box to keep the application up and servicing. So one wonders was it the OS, the Application, the hardware layer or just bad configs. It's all relative. So its hard to state the OS is the cause or the cure. In any modern circumstance.

Quote:

---

IBM is a good friend of Novell, and even Dell and HP support Linux for server purchases. The only one I can think of that I'm not sure about for Linux is Gateway, and who the hell uses them?

And for Windows, their TCP **** was taken from BSD

---

Windows has taken a lot of **** from Linux. Gzip is the latest, behind the scenes in the latest OWA release. At the same time linux copies windows look and feel and methodology within modern interfaces. Damn thing looks and feels like Win9x. Which I like. Of course you can pick the interface and sometimes I like the big eyes and multiple consoles (you know what I mean).

Quote:

---

But when Windows first came out, there was no Linux.

---

When windows first came out there was unix and early versions of linux were nothing more than old unix basic clones. At least the way one interfaced them. I used the first release I could get my hands on, downloaded it from an old FTP server from U. Mass. I had a constant habit of forgeting to switch to binary mode to dload apps. I was pointing out that windows was never designed to be a server. That came later when MS realized the potential of the internet in the server market. Before that Bill cared less. It pisses me off when people compare Windows and Linux within a time period that Linux didn't even exist. LOL. Hell now that I think about it, windows boxes didn't connect to Unix either to get "connected" they connect to Novel. lol. Unixe just served up TCP/IP based services and many of them had to be retrofitted to do that.

Quote:

---

Heh, like when Microsoft tried taking out Netscape for making a browser that worked on Windows and wasn't IE?

---

MS pissed me off with the netscape fiasco. I remember the service update that did it. I deleted windows and restored from backup just because of it. Didn't use IE until years later when interfacing ASP based business partners forced me too. But MS never made it so Netscape would not run and they never tried to sue Netscape, it was the other way around. I have never stopped using it, although I like Mozilla better. But AOs spell check appelet has bee crashing it. So I have been using IE for the past week or so.

Quote:

---

Before the PC, Microsoft sold Xenix.

---

The entire history of MS is interesting, which key moments turned them into the biggest company on earth.

Quote:

---

Umm, you can install on every machine at once. Linux and BSD can be installed from anywhere. All you need is an NFS server, or like with SUSE, you can do the installs over SSH or a VPN and all that needs is 256 RAM. I'd think Windows would take long you actually have to be at each machine from what I understand.

---

True for windows too if you add in additional components like an NFS server for Linux. Neither is fool proof and ONLY works if the hardware is pretty damn close in all cases. Unless you go to some terminal services instance where the image is held on a server. I was refering to actually getting something to work on it like from Zero install to a full blown database server. Windows wins unless you get a full blown and configured distro image. But that would be the same for windows as well.

Quote:

---

What if it's a company who uses legacy things on 486 machines? Windows won't run on that anymore.

---

Dude, if modern Windows won't run a legacy 486 windows app, linux aint going to do it either. By the way I have a few DOS apps! That are in live production. They are running on 486 and 386 machines with Okidata printers. However linux makes use of a great way to use old machines to run linux apps.

Quote:

---

Don't forget to tell those companies when the next big Worm is laying hell across Windows they don't need to worry

---

He he, and you don't think I am not worried about all the MY SQL server databases and my own personal PHP Nuke. The level of worry is equal, and in the major MS cases it was patched months in advance. But MS is a larger target so a firewall comes in handy. Since MS is smacking down their installs the next big worm could just be cross table My SQL exploits, worse PHP problems and in case you don't check up on Apache I don't know how many patches I have applied this year. Lickily, unlike the past linux is making that easier and easier every year.

Heh, for a walk down memory lane:

For the record. Netscape came BEFORE IE, Billy copied netscape's feel but instead of making it better, they made it worse. It sucked in fact so why the hell would I let MS force me to use it? Damn that thing sucked back in 1992 or whatever.

IE was made from Mosaic ;)

Hmmm,

I won't get too involved, but I would like to make a few comments............

Quote:

---

What if it's a company who uses legacy things on 486 machines? Windows won't run on that anymore.

---

Errrrrr, yes it will.............not XP or Win2K but we are talking legacy APPLICATIONS, which are probably DOS/ASM anyways?.........they won't run...........nothing to do with the hardware (apart from ASM to some degree):) This is all to do with the "computing experience" B/S........like they try to stop you loading stuff onto what they think is a "slow machine"?

And you can run the APPS on P4s with a suitable OS (Win98SE/ME and to some extent NT4.0...........remember there are 4 .bas apps in the Sys32 folder of NT4.0?..........now those are "quickbasic" and that shipped with DOS 5.0)

Now, those "486 legacy apps" won't run in a *nix environment............things were a lot different then, and *nix apps tended to run on RISC chips that peed all over intel stuff (and Cyrix.........don't think AMD were playing back then?) they were written for the kit.

So my point is I don't think that *nix gives better legacy support?

:D

Well to some extent you're right on the head with that one. Linux and BSD do have some great support that's being used now and being worked on each day like DOSBOX, and Dosemu and even WINE.

Of course they aren't complete and won't run every app, but they have a lot of potential and most apps are doing great more and more each day.

Hi Gore,

I think that we are at the same place but I am taking hardware history and you are taking software history?

Quote:

---

Of course they aren't complete and won't run every app, but they have a lot of potential and most apps are doing great more and more each day.

---

I put that down to people such as ourselves taking an historical interest? Let's face it, there is no real need for a *nix OS to support DOS?............but it is rather cool?

OK guys say to me "johnno, that OS isn't supported anymore"............and I just smile at them and say "oh yes it is" Hell, if you can support Billy Windows you shouldn't have a problem with DOS?

And 10 ubuntu CD packs came today.................... :D

DosBox is cool, it's faults lie in legacy driver support. People write em when they get bored. It runs on Winders too. I have to retract and earlier statement. I pay as much attention as the linux guys who listen to one another. Win2k Server will deploy desktops right from the intall if you load the Remote Installation Service. No extra server required. Of couse in either case, linux or winders. The remote box must support and have enabled remote booting.