Microsoft fixes severe 19-year-old Windows bug found in everything since Windows 95

[Jeff Gaines]

I think 3.1 was the first usable version.

Did you use the Trumpet Winsock with KA9Q? I remember down-loading it form Alice Springs which had a very limited number of 'phone lines.

[dennis]

It uses the drivers that came with the hypervisor and you don't generally need to install those drivers other than those needed to manage the hypervisor from the OS running, you don't need them to run the OS and applications.

[hah]

nothing? Is that a Linux distro?

[Stan Brown]

I stopped reading at this point:

"According to Freeman, the bug relies on a vulnerability in VBScript, which was introduced in Internet Explorer 3.0. Even today, the bug is impervious to Microsoft?s anti-exploitation tools (known as Enhanced Mitigation Experience Toolkit) and the sandboxing features in Internet Explorer 11. "

Sounds like people who use real browsers have no cause for concern.

[Tony Houghton]

It's the lazy person's version of Linux From Scratch.

[The Natural Philosopher]

Nom wi8n 95 worked OK. Not great, but it worked

[Paul]

I like to think of the situation this way.

1) Average defectivity rate is one bug per thousand lines of code. 2) Average OS has 50 million lines of code. 3) Every OS has 50,000 bugs. 4) Easy bugs are found quickly, subtle bugs are not. OSes are released, before we're significantly down that bug curve. (I used to work in a group that released two OSes, that's how I know the situation. I saw the actual bug curve.) 5) In bad companies, code is reused without review. For example, all sorts of idiots reused JPG and TIFF libraries, not knowing there was a mine-field inside them. We ended up with some serious image-related exploits, years after the libraries were in common usage. Witness the mess with SSL as another example (just about every bad practice imaginable). 6) While there are some techniques for reducing defectivity, they're not all that effective. They might reduce bugs by a factor of three. There is an organization that tracks the issue, and delivers "ratings" to companies. I've forgotten the details.

In such a situation, I don't see much point in standing on a soapbox and claiming superiority. There isn't any. Merely a matter of how public some of the more serious issues were handled.

As far as WinXP goes, if you go to Windows Update today, there are no security patches for you. No new ones. But, if you modify the OS, to indicate it is a Point Of Sale system, you can get security patches. And bugs like this one, may actually be patched for you.

Paul

[Huge]

[snippage]

They could.

[J. P. Gilliver (John)]

Yes I remember using Trumpet, though I don't remember downloading it from Alice Springs; it could be that it was supplied by my ISP with their suite (which included a mail client and a newsreader).

[Tim Streater]

OK, ta. No excuse for Windows 1 then.

[michael adams]

It's probably a question of resources. Xerox stars had something like 384k of memory whereas the lowest specced PC's had between

16k (PC 1981) and 128k (XT 1983) which was the year before Windows 1 launched.

At that stage the objective was still to try and sqeeze as much as possible out of minimal resources. Rather than trying to force potential customers into hardware upgrades.

Apparently Gates was much influenced by the Visi On GUI for PC's. Which being feature rich, but resource hungry required a hard disc for virtual memory swaps. As hard discs, like extra memory were relatively expensive at the time this was one of the factors which made Visi ON a commercial failure.

michael adams

...

[Tony Bryer]

You've indentified MS's key problem, maintaining backwards compatibility, and I think they've done a pretty good job. As Raymond Chen identifies in his book "The Old New Thing", any number of corporates have in-house apps or particular hardware that they need to keep running and if a new release of Windows won't run them, then they don't upgrade.

[Daniel James]

I bought a Dell 486 box specially to run the beta of NT 3.1 (which *was* available in 1992, the commercial release was in 1993) ... it had 64MB at the end of its useful life, but I *think* it had only 16GB when I first had it, and it ran NT very nicely. It had a 1.4GB SCSI hard drive, too ... I remember thinking at the time that that was HUGE.

Wikipedia says the minimum spec for NT 3.1 was 12MB, so my recollection of 16MB is probably correct. I upgraded to 64MB when the price of RAM dropped (quite a lot).

.. but that's beside the point ... we're discussing which version of Windows was usable, and for a fair comparison we have to assume that the hypothetical "test machine" in that comparison meets the minimum spec for the Windows version in question. I'm a freelance programmer, and my time is my money ... if I have to buy better hardware to run an OS that crashes less often and saves me time then it's money well spent.

I used (well, played with) Windows 2.03 and 2.11 ... and 3.0 was a HUGE improvement over those. 3.1(1) wasn't too terrible as long as it was run in 386 (extended? enhanced?) mode ... but 16-bit application software was really unreliable (large model code on the 16-bit segmented intel architecture was a black art that most application programmers never quite mastered) and Windows 3.x was not robust in the face of application crashes -- though it was much better in 386 mode than in "real" or "standard" modes. Win9x was better, but still not good, and by that time NT had been out for 3 years

[Vir Campestris]

• about 1000...

I hated 9x. And then the jammed the 9x Explorer on top of NT :(

Andy