Discussion Locked

The Great Myth - Windows that lock properly

0 helpful
28 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

Microsoft Windows has been out in its many variants for over 20 years. The first versions were overlays on DOS and had very good security compard to the current versions. Why does it seem impossible for Microsoft to write a new version without including most, if not all, of the vulnerabilities that were found and patched in the previous versions?

Some think this is laziness, other see a conspiracy. I don't know which, but it is bloody annoying to have to spend a lot of time keeping up with the vulnerabilities and loading the patches.

Yet each version is touted as being the most secure yet, then within weeks we have the patches for previously known faults. Soon we are told we can expect to see the first fruits of the MS involvement in Trusted Computing and that this new software will be totally secure and have a big brother tie that will make Orwells 1984 look liberal and open minded.

This leads me to ask some questions:

1. Does anyone really think that MS will get this right with no known vulnerability types at time of launch?

2. Does anyone really expect that they can trust MS with the centralised data that they say they need to make the Trusted Computing to work?

3. Is there any real corrolation between the current lock down on software etc by MS and the movement by small business to Linux/Unix systems?

Thoughts and comments on these points are most welcome. Yes I know I have thrown the cat amongst the pidgeons on this, please try not to throw too many more in.

28 replies

by jaqui

no.
microshaft is not capable of producing anything worth using.
and they are incapable of producing secure code.

by deadly ernest

Since you are in Vancouver I may be able to get away with saying that some people like M$, like those guys a few years back that just LOVED to play with MS Flight Simulator whilst learning to fly into buildings - seems the simulator did not include the instruction 'Always park on the runway'.

Less seriously I always found it funny that certain games manfuacturers made the PC version of the game to run on MS yet if you bought the network version it was designed to run on a Linux server.

by ~omega~

Honestly, I'm getting irritated by all these people who tout their opinions as truths. Windows has time tested and popular support from consumers to back up the claims it makes. Linux has nothing but a bunch of people who like to fling mud at a good product.

Of course, windows has flaws. If you wrote a product with the same capabilities, under a deadline, you'd probably come up with some flaws as well.
The difference is, while Linux supports a product for a few months, then throws it's code to the consumer to troubleshoot, fix and recompile, Microsoft *EMPLOYS* people to do that for you. I seriously doubt Linux is flaw free... It just have fewer people to exploit, so it isn't the center of attention for hackers.
Not to mention that MS has a huge balancing act it has to maintain between security and usablity. The two are relativly mutually exclusive.
I am the first to point out that Microsoft has flaws, but I'm not so quick to throw it in the trash.
This is not to say that I don't think Linux is a great product as well. There is room for more than one operating system in this world, and I personally don't want my customers using Linux. I would be over there everyday teaching them how to use their computers!

by hal 9000

The really juicy things to attack are the Internet Nodes the main 13 of which are on US soil and not one of them uses a Windows Product on it.

Then there are the financial institutions which have billions of $ in different currencies flow through their systems on a daily basis again none of these server units run Windows or any M$ product for that matter.

The very first case of Hacking was into the Melbourne Universities Mainframe which of course ran Unix and from there those individuals could run rampant as there just was no effort placed on security back in those days and what made it even better was that there where no laws being broken.

But when you look at what happened back then they spoofed valid user accounts and then created their own so that they could have permanent access to the network.

As for this "Windows has time tested and popular support from consumers to back up the claims it makes. Linux has nothing but a bunch of people who like to fling mud at a good product."

With the bulk of the software houses only writing apps for Windows machines what else can you expect? Ask any of these users if they have had any down time and they will reply "YES but that was required by the System to do Such & Such."

Ask them have they ever experienced a intrusion and they will reply sometimes "Yes but that was because I didn't do such & such."

Ask them has a Windows patch or Hot Fix ever broken an OS or application and their answer is always YES!

"Of course, windows has flaws. If you wrote a product with the same capabilities, under a deadline, you'd probably come up with some flaws as well."

I just love the bit about Deadlines when was the production version of Longhorn supposed to be a production reality? M$ deadlines are nothing short of a joke I can still remember some guy getting really upset with me for being Flippant when I suggested that he not wait for Longhorn to hit the markets as he was being serious and I was joking at his expense, that was now about 18 months ago!

I'll leave the rest of your rant alone as really it is at best not worth the effort and at worst ill informed. But I do have to say this I attended the Launch of 2003 the very first M$ Product to be released after their much touted "Trusted Computing" was announced and we where all told that this was a new product from the ground up and complied with every specification of the "Trusted Computing" that M$ had introduced so that their customers could have faith in their products. Within 2 weeks there was a Patch available for 2003 ES and as it was a "Critical" one it sort of left a sour taste in my mouth after listening to their "Hard Sell" about just how great it actually was and just how much they had improved the Base Code.

Now part of the problem is that M$ writes code to be used on as wide as possible array of hardware and even XP has a basic requirement of a 300 MHZ CPU with 128 MEG of RAM which by any standard is positively geriatric so why are they supporting this old hardware when it isn't possible to buy now days new? The OS that ran perfectly on this stuff is no longer supported and that was either NT or 98 if M$ considers this obsolete software why is the hardware still supported? Why does each release of Windows get bigger and have more eye candy but doesn't address the very basics that so much time and money have been spent on addressing on previous versions?

Why are M$ Product Life Cycles so short? Currently we will soon be reaching the 5 year limit on XP where full patching will no longer be supported but right at this point in time there is no replacement for XP! The 5 year limit should arrive just about the same time as whatever Longhorn is called is released and there will be no time to allow us to fully test the new offering so we can come to grips with it before deploying it on a wide scale Business base if the Business actually want it. They didn't want to move away from 98 Se to Y2K or even XP and while 98 and XP where being sold side by side 98 was the one winning in the sales area this came directly from a M$ Partners meeting and was their excuse for stopping production of 98 but then allowing the retro installing of 98 if we sold a copy of XP Pro. I really have no idea of just how many XP Pro pieces of software that I have sold that will never actually be used but I'm doing as I'm told and fitting the XP Pro COA to the box and loading it with 98 or NT4 depending upon the need of the business.

With the release of XP the activation was a real pain for me as I have a unit there that I test new Hardware on and I was constantly having to reactivate all the M$ products loaded on it every time I changed or even fitted for a few hours of testing 3 or more new products. That was a nightmare as at least once a week I was reactivating all the software even though the computer had been returned to its original configuration, since then I now only use Volume License stuff and even buying that isn't an easy task as I have to order it in and then wait at least 1 week before it arrives so I can then pick it up or wait another few days until it eventually gets freighted out to me.

While it fixed one problem it has introduced several more that where unexpected and unwanted. Now every time that I install a Service Pack I have to ring M$ and get a new Product Key as for some reason the ones that I've been given just are not acceptable to the Service Packs, at least with SP1 you where told about the problem before you actually installed the pack but with SP2 that didn't happen and you only found out about the problem the next time you logged onto M$ Web Site to down load a bit of something and you where denied the opportunity because the M$ Server wouldn't accept the product key that I had obtained from M$ before applying the previous Service Pack.

It is really expensive, time consuming and a waste of time having to ring M$ for a new Product Key wait at least 1 week while they E-Mail you one after putting you through the grinder to prove that you really bought the thing in the first place and generally making you feel like someone who is stealing from them and then you have to change the Product Keys on all the computers which is another thing that M$ doesn't pay for but which they cause.

Anyway Rant Finished no malice intended! :)

Col ]:)

by jaqui

to look at the open source model if you think it's supported for a few months then tossed to the people to deal with.
it's always the people, and a new version of the os every 6 months.

hmm, a 6 month product cycle.

ms uses a 5 year cycle..

how does linux / *bsd, irix, unix do it?
beat ms by 4.5 years in product cycle?
ms pays 20 thousand programmers at most

open source has 20 million.
( 20 million linux users, add the other open source and the number climbs )
take a look at the newer releases, all the features winders has, with very few vulnerabilities.

guess those millions of prgrammers get it done better than ms thousands do.
more yes fixing code, 24/7 progress on code quality.

ms can't win in a quality battle.
they can't afford the same number and time for programmers that the open source operating systems have.

by apotheon

That's a very good summarized presentation of the most common, and least valid, arguments against Linux use. Good job.

It's clear you're no Linux expert. It was clear in the second sentence, when you said Windows "has time tested" support to back up its claims, as though Linux does not have the same time in service. From there, it just went downhill.

by deadly ernest

The last version of Windows that was prepared and released as per the deadline was the original PD-DOC licenced to IBM; since that day MS has never met an original announced deadline with an O/S. In some cases they just announced the release with no pre advertising.

Windows is used by people simply because MS pay many millions in advertising and use their market muscle to leverage their apps into prebuilt systems givin uneducated users no choice or option.

I have been around a looong time and appreciate what MS did to improve the market in its early days, but since they stole (yes stole) the Windows concepts and code and started building total Windows based O/Ss with Win 95 their quality has dropped very badly.

People usually do not mind if a new O/S or app has a fault or two that turns up after it is released, happens with the best software. The big problem with MS is that they constantly tell us that the new O/S is better, stronger and more secure and within a few months have to release a patch for a fault or vulnerability that was known about two or more versions back. Simply put they are not fixing problems with the kernel when they arise but patching after release only if they get found out. If a fault is noted in Win 95 it should have been fixed at the kernel level by Win 2000. But that does not happen, even when MS say that they have rewritten the kernel from scratch, they include the old faulty code - why - laziness, sloppiness and just bad programming.

Tell me you have fixed the problems, tell me it is better and then charge me a lot for the same faulty crap and anyone will complain.

BTW most Linux distributors employ people to write the interfaces etc for their software, most of what is done by the open source community is the coding for the new kernels and coding for special apps like Appache etc.

I agree we need more than one O/S and they all have their uses, but lying to people re quality is not a way to make them happy.

by stress junkie

I've worked with M$ products for a long time. I was introduced to MS-DOS because my employer's main customer had MS-DOS. The employer had been using CP/M and Tandy TRS-DOS, both on Tandy Model 2 and Model 16 computers. I preferred either to the MS-DOS. This employer also had M$ Xenix and DEC VMS so I was exposed to a lot of different technology at that job. As time passed M$ Windows came on the scene. Windows 3.11 was almost good enough to use in business. Then M$ added a free TCP/IP stack to Windows and called it Windows for Workgroups. Since my employer at THAT time was paying $1,000 per client for a third party TCP/IP stack I thought that Microsoft making a free driver available was great. NT v3.5 came along and I saw that a TON of excellent system administration tools were included. I was also running DEC VMS at the time and you would pay tens of thousands of dollars for system administration tools like the ones that came free with NT. Then I was a REALLY BIG fan of M$. Plus with NT you didn't see the General Protection Fault that was so common in Windows 3.11 and its predecessors. M$ NT v3.5 gave you exellent networking interoperability. I used an NT machine to create network shares that could be accessed from Macintosh, VMS, and Windows at one location. It was great in unifying a fragmented network storage scheme. Plus I found that the IPX/SPX driver would completely circumvent Novell security. I could connect to a Novell v3.5 server without having an account or any kind of authentication and I would have complete access to all of the files on the server. That's when I lost interest in Novell.

However as time marched forward Microsnot products were found to be poorly designed. As networking became more prevalent in businesses problems in Microsoft security were found that were related to or enabled by networking. Then businesses started to connect to the Internet. Microsnot software then became a liability to business security. Since I've always been a security nut I decided that Microsnot product quality was too poor to use responsibly in business. Unfortunately Microsnot had become very popular and there were few reasonable options. You could use OS/2 Warp if you could find it. Various Unixes were not user friendly enough for people to adopt for home use, if you could find them. So we saw the beginnings of the current situation.

It has taken over ten years for Linux to be considered a possible desktop replacement for M$. Ten years ago there were articles in tech magazines asking if you would be fired for recommending Linux be used in work. Back then there was ONLY the XFree86 GUI and very little end user software.

So things are getting better as far as considering Linux as a business tool. People do not ask if you would be fired for recommending Linux. Many small businesses consider Linux to be one of two possible options to use on servers. The only real resistance is that businesses don't want to have a fragmented computer environment. They don't want to have one person for servers and another for desktop and neither knows how to do the other's job. But Linux is being deployed in servers.

Now end user software for Linux is better. I don't know if OpenOffice.org or Sun Microsystems' Star Office can be considered drop in replacements for M$ Office Suite. File formats are a problem. If you have to share documents with people who use M$ Office Suite then the Linux software isn't a good choice. If you don't have to share documents with people who use M$ Office then the Linux products appear to me to be pretty good. I'm not a document guru though. A good secretary might disagree with my assessment of the Linux document software.

So here we are. The most popular software has serious design flaws around separation of privileges, separation of communication between processes, and coding bugs that allow buffer overruns and other bad stuff. The Unixes are much better about separation of privileges and communication between processes. Other problems like buffer overruns are a problem with Linux products. A high profile Linux supplier like Red Hat or SuSE do a good job of finding and fixing bugs and making those fixes easy to install over the Internet.

Lately Microsnot products have been increasingly disabled before registering the product with Microsnot. I recently purchased the student edition of Microsnot Office. It won't do anything until I register it with Microsnot. That's a problem in my book. Microsnot is making products that require you to register with them. However when I go to the store to buy a hammer I don't have to register my purchase with the manufacturer before I can use the hammer. As far as I'm concerned Microsnot does not have any right to know my name or anything else about me just because I purchased their product. You don't find this sort of Orwellian facism in open source products. Naturally I can put false information into the product registration, and I do. The point is that Microsnot should not require that you register with them before the product works. I paid for it; it should work.

So, for system security it seems that any Unix would be a better choice than Microsnot operating systems. End user product quality is an issue on any platform but patches appear to fewer and more effective in open source land. A good Linux distributor will do a good job finding, creating, and releasing patches for newly discovered software problems. That includes software from other sources that they included on their Linux installation disks. Other Unixes are similar. the three BSDs are very stable and so have few new problems. Sun Solaris is good but their product line is confusing, expensive, and they keep making changes in their software strategy. However Sun's operating system is genuine licensed Unix and it is good quality software.

There are some other players but they are not really worth mentioning.

So that's the way that I see it. Most small businesses want Microsnot software so we support it. If a business wants to try something else then we can offer one of several Unix and Unix like products. The problem is that they are not as user friendly and not compatible with Microsnot software data files for the most part.

And now you know the rest of the story.

LSMFT

by deadly ernest

Actually I have been using MS apps since DOS 2 and have been using Linux since 1998. I can relate to people wanting to stick with a familiar product and thus stay in the MS rut. However, the latest versions of Open Office and the Linux distributions like Mandrake 10 etc have gotten over that. If switching from MS Office 2003 to MS Office XP or the latest Open Office the difference in feel is the same and most people will pick up on the variations easily and quickly. The latest Linux distributions is as easy, often easier, to load and set up than the latest MS software; also they come with the security default settings as on not off.

Over the eyars Novel is about the only stuff I have not used. But have usually stayed with MS as that is what my client base used, mostly small businesses. But lately more of them have started asking me about switching to Linux, wanting to know the costs, the learning curve etc - the big thing for them is the lost down time whilst patching faults.

I am also building a Mandrake 10 box so that I can run my current MS apps and my older Win 96/98 based apps on the same system within WINE without having to use a dual boot system.

Microsoft got a hold because they produce cheap software for the masses - they still do but try to act like they produce quality software - something they have never done.

by jaqui

is a poor cousin to crossover office.

crossover office is not, as the name suggests, an office suite, it's the functional version of wine.
it even runs ms autorun to start installing ms based software into linux.

Xandros linux comes with this commercial application. I've seen it install borland's builder 6 into a xandros installation without a single problem.

by htmlman

If I am not mistaken. Reading the license agreement to Windows software you are not buying the software you are getting permission from micro$ to use the product.
I haven't checked all the document associations to Open Office 2 but the ones I used can open and reconise office documents and viseversa.

by lastchip

1. No

2. No

3. Yes.

And something that seems to be overlooked. As I understand it, once purchased, a Linux Distro of your choice, may be installed on as many machines as you like.

Hows that for the bottom line?

by hal 9000

You don't have additional License fees imposed for exceeding a set number of units attached to it.

Then a really important thing is that the Linux Server products are far more scalable than anything that M$ has available. Currently you can run 1K CPU blade arrays on one Linux installation which is just not possible for any Windows Server Product currently available.

Then the really important thing is that you are not constantly rebooting the Linux systems every time you add new software to make it work properly and with the current Blade arrays and Thin Clients you really effectively have a small Main Frame at a fraction of the cost of what they used to be, in a fraction of the space and at a fraction of the running costs as well.

Col ]:)

by hal 9000

You realize this. :D :D :D :D :D

Is that enough or do I also have to add a "Sarcasm Alert" as well?

Col ]:)

by hal 9000

1 No Way in Hell. I've actually been known to say that something is about as likely of happening as it is that M$ releases a product that works the way that they claim.

2 NO. NO. No. No. NO. NO. NO. NO. to the 100000000000000000 power!

3 Most defiantly.

Since its inception M$ has constantly shown that they are untrustworthy in any form whatsoever.

There is only one certainly with anything M$ it will be bigger require more system resources and have even more "Undocumented Features" than the previous version of whatever it is that you are looking at at the time that comes from M$.

Even if you where stupid enough to actually Trust M$ with your information and they only used it for their own purposes which would be forced upgrades and constant residuals you can not trust their infrastructure as it is unsafe and has been hacked on numerous occasions so any data that you are silly enough to send them should be considered as Public Domain.

The data that I provide M$ as a Certified Partner is from my prospective advertising only so I only give out what I would place on a business card.

Does that go some way to answering your question?

Col ]:)

by deadly ernest

kind of confirmed what I thought - but will have little effect on what happens.

What does worry me is that when the Trusted Computing concept was first floated by MS under another name (Paladium I think) they stated that any machine using the new system would be safe to communicate with any other machine using the system and that they would ensure that no contanimated systems would infect the system by making it impossible for the older systems to talk to the new systems and you could not turn off the checking system. This caused a furor and the Trusted Computing concept came out saying that you would be able to turn it off, but again it would only talk to a 'trusted' machine whilst TC is turned on.

This still leaves me wondering how all those millions of poorer people who use older systems and software will be able to use e-commerce since they probable wont have 'trusted' systems until about 7 years after their release (based on current trends. Thus about 80% of home uses wont be trusted by the big companies with the new systems. Methinks many millions being spent on hardware and software capabilities that will only be of use to a few govt depts - thus they should not be built in but add ons.

by hal 9000

Is very conservative as I still am building new 98SE Boxes as the companies don't want to upgrade their existing software to the newest one out. So a lot of the new box that I sell has a nice XP Pro COA on its side and is loaded with 98 SE I don't think that a single one of these Volume License products will ever be actually used unless they decide to use XP Pro when Vista hits the market.

The whole thing is a mess and it's getting harder to deal with every day.

Col ]:)

by humbletech99

They're both good and bad products, depending on what you look at.

If linux user base approaches microsoft's, I'll bet anyone on the planet my entire life savings and even my life (I'm so confident that this is true) that there will be more security vulnerabilities and in increased frequency found in Linux AND the software that runs on it (don't forget that).

Give credit where its due, they're both good enough to have millions upon millions of users, servers etc. so they can't be that bad, there's just always room for improvement, being the greedy humans that we are, we want more... forever.

Let's give the software developers on both sides a pat on the back (they're excellent people) and say keep up the good work and we'll get there!

Positive encouragement, please, people...

by apotheon

It's a good thing I'm a nice guy. I won't require you to give up your life when your prediction doesn't turn out to be accurate.

In fact, you're too late: it's already inaccurate. Take a look, for instance, at the comparison of Linux webserver systems as compared with Windows webserver systems. Linux/Apache systems make up something like 60% of the webservers, and Windows/IIS about 20%, as I recall. Somehow, Windows/IIS ends up being the most exploited, most vulnerable webserver implementation on the market.

Maybe we should look into DNS. Linux outnumbers Microsoft there, too, and Microsoft gets the lion's share of exploits.

How about mailservers? There's a greater percentage of mailservers running Sendmail (rather than Exchange) than the percentage of browsers that are IE now, thanks to the influence of Firefox. Exchange is the most-exploited mailserver, though.

Perhaps we should compare Squid proxy servers to MS Proxy Server, or version control system servers between Linux and Windows, or engineering workstation systems in aerospace engineering (sensitive government-related systems), or rendering farm servers in the movie industry, or . . . well, you probably get the point by now. In all those cases, the Microsoft solution is [b]FAR[/b] outnumbered, but always gets most of the exploits.

Windows has its uses. Those uses are not security-related. You really should recant on that "bet my life" bit.

by rkuhn040172

If Linux has so many programmers at their disposal, then why such lame driver/device support from Linux?

And don't give me that proprietary/vendor crap.

6 month product lifecycles are a pain in the rear especially when updates are so frequent yet can't even detect my printer.

Oh yeah, I guess I could read for a few hours and mod something, or I could just let Windows detect and install it for me...

by apotheon

There's awesome driver support. The problem isn't lack of driver support in general, but lack of driver support for certain specific things that are somewhat common. The example to beat all examples is the Winmodem which, as referenced in the name for it, was designed specifically for Windows.

Wireless cards are neatly handled by ndiswrapper. Video cards are well-supported by kernel modules installed when the OS is installed.

Meanwhile, I've lost count of the number of times I've seen problems arise with Windows trying to install the wrong drivers for something, and refusing to let someone simply install a driver from CD.

Finally: Why not mention the proprietary drivers offered by vendors?

by hal 9000

Would you like me to list a range of devices that Windows will not host?

Not the newest ones but the tried and tested things like Video Capture cards. The Life View ones are not detected by XP on install and nor are the newer Sound Blaster Sound Cards. A lot of the Video Cards will result in default installs and require a driver upgrade from the maker not MS.

Col ]:)

by apotheon

There's a big difference between oft-repeated Microsoft marketing copy with little basis in reality and the still unrefuted arguments that explain problems with Windows systems that Linux systems solve easily.

by deadly ernest

I see that we are starting to get into the 'which is better' debate - again. For those who have not been keeping up with the surveys etc, a few points - and no I do not remember the bloody URLs for them so please don't ask. What I have seen during 2005 reinforces the trends from earlier years. NB Market share changes are usually based on percentage of market share - can be misleading as the market is still growing. Often a reduced percentage can still represent a real growth due to a high market growth rate.

In the desktop software area MS is the clear leader but still losing more market share to Linux each year- the rate of change is increasing.

General office apps is mostly MS with them losing more market share each year to Linux and Lotus. Especially so in regards to mail servers.

There are more Linux Internet and general system servers than any other O/S followed by Unix and then MS - common reasons cited being dependability followed by better security.

For special embedded software apps - Linux is the O/S of choice as it uses less resources and is easy to adjust for the specific use; about 75% of the market in one survey.

Some serious vulnerability studies of several Linux distributions and their MS equivalents shows that MS, at time of release, were far worse than the Linux equivalents released at that time. Another point found was that when the Linux suppliers fixed a vulnerability it did not appear again in later versions. MS tend to shoot from the hip with fixes and patches that often cause other troubles; whilst the Linux community takes more time in preparing the fixes and appears to test them better. This results in the Linux fixes taking a bit longer but being a better quality fix, that is longer lasting. the down side is that not all Linux distributors cooperate closely re the making and inclusion of fixes. Up side is that each new kernel includes all the known fixes.

Both camps have problems re drivers - the biggest complaint against Linux is that the hardware manufacturers do not always write Linux drivers for their gear so you have to wait until the general community does it. The biggest problem with MS is that their software automatically thinks that any MS generic driver is far better than anything specially written by the manufacturer and will automatically 'fix' the faulty driver by replacing it with the generic. This 'feature' cannot be turned off in MS but can be turned off in Linux.

People should be closely looking at what they want out of the system and then choosing the best OS to suit their needs and budget - often the needs will revert around compatibility with other software or people, and that is where MS tends to keep market share - people not wanting to change from a known beast.

by lastchip

Today it has been announced, "Second Part To Hell", a virus writer, has the dubious distinction of being the first to publish a virus for Vista.

Trust Microsoft with anything; Dream on !!!

This discussion is closed to new replies.