OT Windows Vista Warning

[Bob Eager]

There still has to be an underlying OS to make it all work. Try running an X Window system without one.

[Andy Hall]

Which is a reasonable indication.

On Unix based systems it is common to have both an X server and X clients, although an application/file server typically won't need to have X display server components installed and a firewall certainly shouldn't.

On Windows one finds and uses X servers but rarely if at all X clients. (use of correct nomenclature that an X server is the display component that is normally on a desktop. I know that you and Huge know that but others might be confused by the apparent reversal of roles)

Nonetheless for an application and file service platform, there is no need to have a local GUI display component, especially since, as is the case with Windows, it compromises the security and integrity.

[Huge]

You have the point back to front. Try running Windows without a GUI.

[Bob Eager]

That can be done. And it's perfectly usable as such. OK, not with GUI applications, but it's still an OS.

[Andy Hall]

Like Windows, it was a monitor

[Huge]

Nope. It was a program loader and monitor.

[Bob Eager]

From where are you getting these definitions?

[Andy Hall]

CP/M = Control Program/Monitor

or Control Program for Microcomputers

Even Gary Kildall didn't have the audacity to call it an operating system in the way that Gates described MSDOS, even though functionally they bore a lot of similarities.

Microsoft has a long history of dishonesty in this regard.

[Steve Firth]

s/ in this regard//

[Bob Eager]

Oh, I know that. I was using back in the 1970s. I mean your assertion that Windows is not an OS.

[OG]

I'm not aware that you owned the definiton of an Operating System.

[Andy Hall]

You're probably not aware of a lot, but leaving that aside, an operating system can be described as having a number of characteristics and offering a number of services.

Against many of the key metrics, Windows fails so miserably that it doesn't warrant the title of an operating system, but is more akin to a monitor/debugger with a GUI.

[Dave]

Oh, I don't know. There will be the equivalent to version of the Fylde Cost edition of XP within a few days :-)

Dave

[Andy Hall]

I haven't given any "definition"

However, I have highlighted numerous areas, which can be described as characteristics of an operating system where Windows is so seriously deficient that it barely meets the characteristic, if at all.

For example:

- Protection against badly written code or malware compromising the integrity of the machine and data is poor because of the many compromises required to achieve meaningful performance and run existing applications.

- Vulnerable out of the box in order that users can install applications

- Exposure of a large number of opaque network sockets which are security vulnerabilities

- Poor to non existent self housekeeping resulting in a degradation of performance

- Opacity in terms of how to optimise and maintain the environment

- Poor management of service processes

- Poor networking implementation vulnerable to denial of service attacks

It is interesting, but not surprising to note that the top 5 out of 7 operating software security vulnerabilities listed by the SANS Institute in 2006 are in Windows and associated MS applications.

[Bob Eager]

He doesn't. But he has invented his own, which is at variance with most (if not all) standard textbooks on the subject.

I agree that Windows may not be a secure operating system. But it is one, nonetheless. There are plenty of operating systems that have little or no security, because in their environment they do not need it, for example. But they still provide key services such as resource management, virtualisation of hardware interfaces, filing systems, etc. Thus they are generally classed as operating systems. Again, provision of networking is not part of the core definition of an operating system, although Andy seems to think that it is.

[Andy Hall]

I haven't invented a definition, because either that has to be so generalised to encompass anything that might be labeled as such as to be useless as a definition for any specific product; or is so detailed that none would meet all or most of the characteristics.

Therefore, one can only reasonably look at the subject by taking the characteristics that would be reasonably expected to be relevant to the intended use.

Generally they would be classified as embedded operating systems or in some cases real time operating systems; the assumption being that the programmer knows what he's doing and is taking responsibility in the coding quality of the applications for the integrity of the complete system.

One can subdivide further and say that there are desktop and server classifications of operating system, and it would be in those categories that Windows would aspire to be.

In the most general of senses it is not, but it certainly is in terms of the context and of a classification of operating systems into server and desktop. A server operating system these days would be fairly useless without some form of networking. It could be argued both ways for a desktop, although for the use to which most desktops and notebooks are put, some form of networking would be expected by almost all users.

When one puts Windows into the expected context of an operating system for a desktop or a server, it fails so miserably on most of the characteristics that I think it's entirely reasonable to say that it is not an operating system. One could let it off the hook by saying that it meets some of the broadest characteristics, but the definition then has to be so superficial that it is not worth considering. One can't even excuse it by saying that it's OK that it doesn't meet security requirements because real time and embedded operating systems sometimes don't, because Windows doesn't aspire to those market sectors.

Textbook definitions are interesting, but what actually matters is the context and usability for the intended purpose.

[Bob Eager]

characteristics that I think it's

No...it fails to be fit for purpose, but it is still an operating system.

[Bob Eager]

But they are still operating systems. Only now are you trying to subdivide the definition.

There are accepted generic definitions of what an OS is, and even CP/M satisfies that (and so do all versions of Windows based on the NT kernel).

One can then select particular characteristics based on intended use, of course. But your original bald statement did none of those things.

That's all.

When I'm talking about these things, I generally start off with two or three possible generic definitions. Then, when that's established, I look at the subdivisions.

I've been doing this for a long time, but the fundamentals haven't changed. If you wish, search for ISBN 0333598482 at Amazon, and select the Search Inside link. Then look at the 'Excerpt'!

[Andy Hall]

I would have thought that that would have been apparent from the original context of the thread - i.e. a desktop environment was under discussion.

Certainly there was no mention of embedded, real time or other environments, and it would be if one wanted to include those, that a generic definition would be appropriate.

Except that they are so generic as to be worthless when discussing a particular environment.

There was no need because the context (i.e. desktop environment) had already been established.

I prefer to cut straight to the point.

No they haven't, but the fundamentals are so generic to make them of limited value when applied to a specific product.

[Andy Hall]

Sort of like a car without an engine, steering wheel, exhaust system, wheels and brakes is still a car?

Q: Why do I buy a car?

A: To get me from A to B

Q: Will the above achieve that?

A: No.

If the "car" had been delivered to the customer never having had those things, it would never be described as a car but as a body shell.

If it had been delivered to the customer having originally had them but they had been removed, then it might be described as a broken car.

The point is that the general definition is academic.

Now apply the same principle to Windows. There are aspects that are fundamentally broken and which have never worked properly - it isn't as though it once worked and someone broke it or removed bits.

Ergo, it is reasonable to say that Windows is not an operating system in the same way that a car body is not a car.

It goes without saying that it's not fit for purpose.