MS Security Guy Wants Vista Bugs Rated Down

jcatcw writes "Gregg Keizer reports that Michael Howard, an MS senior security program manager, says that the Microsoft Security Response Center (MSRC) is being too conservative in its Vista vulnerability rating plans. Microsoft's own bug hunters should cut Windows Vista some slack and rate its vulnerabilities differently because of the operating system's new, baked-in defenses."

Comments

Hmmmm. . .

[bplipschitz]

Sounds a little like Michael Howard might be "baked in". . .

Isn't that .....

[edwardpickman]

rate its vulnerabilities differently because of the operating system's new, baked-in defenses."

...half baked?

Re:Isn't that .....

[Anonymous Coward]

No, I believe Michael Howard is totally and utterly baked. He clearly needs to stop hitting that bong.

Re:Isn't that .....

[numbski]

You'd have to be smoking some pretty good weed to go along with this. :P

Let's say on *nix there's a vulnerability that allows for remote ssh access. You can only get in as an unprivileged user, heck, you may even get /dev/null as your shell, but it lets you in. Do you rate down the remote access flaw because of *nix's "baked in " defenses? No! You fix the bug and update.

Just because your system is overall more secure doesn't mean that you don't blow the whistle on the flaws just as hard. It's called VIGILANCE.

Re:Isn't that .....

[ericlondaits]

Just because your system is overall more secure doesn't mean that you don't blow the whistle on the flaws just as hard. It's called VIGILANCE.

Mmmm... while it's true that the price of freedom is eternal VIGILANCE, remember that you can get Vista Ultimate for as little as $399.95.

Re:Isn't that .....

[Gnavpot]

Do you rate down the remote access flaw because of *nix's "baked in " defenses? No! You fix the bug and update.

Did you intend those two scenarios to be mutally exclusive?

Rating a bug low does not necessarily mean that it is fixed slower.

Re:Isn't that .....

[Hierarch]

Let's say on *nix there's a vulnerability that allows for remote ssh access. You can only get in as an unprivileged user, heck, you may even get /dev/null as your shell, but it lets you in. Do you rate down the remote access flaw because of *nix's "baked in " defenses? No! You fix the bug and update.

Well, actually, you do rate it down. This is basic risk assessment, and if it comes to a prioritization of resources — which bug should we fix next? — I want that priority set according to the impact of the problem. Cold, hard, rational assessment, not “ZOMGRemoteAccessExploitWTFBBQOver”

You seem to assume that reducing the rating of a flaw means you don't fix it.

Now, more importantly, from TFA, we have

[The] rating system is clear-cut. If an Internet worm can spread without user action -- the MSRC's definition of "critical" -- on Vista, the vulnerability will be so tagged, Vista-specific security technologies notwithstanding.

This is different from the case you're outlining, and if the bloke in this article is really trying to change these criteria, I've got a real problem with that. If it's the difference between a buffer overrun that allows remote access versus a buffer overrun that allows an outsider to crash that process, I think it's the MSRC that needs to correct their own criteria. Either way, it shouldn't be driven by an outsider, although he can and should make the suggestion to them that certain criteria should be revisited.

Re:Isn't that .....

[Seumas]

Nothing says security like naming your flagship product after the part of a house that is made of thin glass and can be broken with a small rock, stick or an elbow and allows everyone outside to see everything going on inside.

Re:Isn't that .....

[UncleTogie]

whereas Windows users tend to gain at least a basic appreciation for proper security practices.
Don't take this personally, but:

What frickin' planet are YOU on? Most Windows users expect Windows to take care of all that FOR them....and boy, are they surprised to find that clicking that "You're infected! Click here to pretend to fix your computer whilst actually infecting it!" actually DOESN'T fix a darn thing. I'm not talking ALL Windows users, but it's a frighteningly large group.

What MOST Windows users want is a system that doesn't make them THINK.

Re:Isn't that .....

[Miseph]

"That doesn't seem like a very vigilant attitude to me... whereas Windows users tend to gain at least a basic appreciation for proper security practices."

While the first part is true, Windows users (myself included), by definition, are ignoring one fundamental security practice... they aren't using a secure system in the first place. It's like making sure your front door is bolted shut and you've got bars over all your windows, but your house only has three walls (and it's not triangular).

Re:Isn't that .....

[tdelaney]

There's a difference between severity and priority.

A bug may be high severity (e.g. remote access) but low priority (e.g. because it's believed that other factors mitigate the remote access).

Another bug may be low severity (e.g. a user interface quirk) but high priority (e.g. because reviewers have seen it and are talking down your product because of it).

Severities should be based on how much damage may be caused to the *users* of the program. Priorities are usually determined by how much damage the bug causes to the *developers* of the program ...

Re:Isn't that .....

[VertigoAce]

Either way, it shouldn't be driven by an outsider, although he can and should make the suggestion to them that certain criteria should be revisited.

To give some context to who Michael Howard is, he is one of the head security guys at Microsoft. One of his roles is to improve the development process across Microsoft to improve security. So the MSRC responds to actual security vulnerabilities, while Michael looks at why the development team missed the bug and how to avoid it in future products.

If you read what Michael actually said the issue becomes more apparent. A security bug that affect Vista and XP will usually be given the same rating, even if Vista has defense mechanisms that it make it extremely unlikely that it can be exploited. In the security alert they will list any defense mechanisms that make it harder to exploit the bug, but they don't change the rating.

Re:Isn't that .....

[suv4x4]

What MOST Windows users want is a system that doesn't make them THINK.

You're saying that as if it's a bad thing. Do you insist on an OS that makes you think a lot?

While you're thinking on the OS you could be thinking on the next YouTube or something. Why waste so much talent? Anyway, if Microsoft survives Vista (which it'll most likely do), and has success with Vienna, we'll have exactly that: proliferation of managed, secure code and deprecation of binary code (which will run in sandbox) except for a range of professional applications (media processing, database engines and so on resource intensive tasks).

Re:Isn't that .....

[causality]

While you're thinking on the OS you could be thinking on the next YouTube or something. Why waste so much talent?

You say that as though the amount of thinking a person can do is a finite quantity, and that each time you think you decrease this quantity, so therefore the wise thing to do is conserve it as much as possible.

However, it's really more like a muscle -- the more you use it, the more able it becomes. Linux made me think very much when I first began using it, especially considering that this was 1997 so as you can imagine, the automatic graphical/menu-driven installers were not nearly as good as they are now. It took a decent amount of thinking to come to an understanding of why the system works the way that it does, but having done this I can now make related decisions instantly. I also learned a lot about how to find my own answers, which is also a skill that comes more easily now than it did before.

You're saying that as if it's a bad thing. Do you insist on an OS that makes you think a lot?

Depriving me of the ability to make security decisions on my own, on the premise that making those decisions requires thinking and well that's just too hard, is not my idea of the proper role of an operating system. The OS is not "making" me think, because all of the thinking done is inherent to the task performed; for example if I am setting up a network then I am thinking of how I want this done, what steps I will take to secure it, which computers will be used for what tasks, etc. The thinking is inherent in the task, just as sowing seed is implied by wanting to reap a harvest.

The thinking required cannot really be separated from the task; the best approximation is to have the designer of the system (Microsoft, in this case) try to determine in advance what you will and won't do and set defaults that attempt to please everybody. This is, of course, directed more by marketing's idea of what they think most people want rather than the developers' ideas of what is technically superior. That is (imho) the biggest difference between the Microsoft approach and that used by most Linux distributions.

To sum it up, no I do not believe that thinking is a bad thing, and thinking in particular is one of those things that I would much rather do myself than have someone try to do for me.

Re:Isn't that .....

[lanzz]

while it's true that the price of freedom is eternal VIGILANCE, remember that you can get Vista Ultimate for as little as $399.95. perhaps they should release a new edition, Vista Vigilante?

Hal Howard

[Anonymous Coward]

I work at Microsoft, I can get Vista for practically free but I refuse to even touch Vista with a bargepole and dont recommend it to others. They dont need it anyway even if it was "finished" and secure.

Re:Hal Howard

[Anonymous Coward]

I work at Microsoft, I can get Vista for practically free but I refuse to even touch Vista with a bargepole and dont recommend it to others. They dont need it anyway even if it was "finished" and secure.

You sound like a contractor that is bitter you didn't get hired on. Those of us are employees of MS want to make certain that we get the remaining bugs fixed. That isn't going to happen if we point fingers and play the blame game.

I work on embedded devices at MS and we won't have Vista support ready for a couple more months. Once the Visual Studio GDR is released in Apr/May Windows Embedded 6 will release SP1 and it will then be possible to develop/debug embedded devices from Vista and I will update all of my machines. Until then I have machines running Vista and XP. If you did work at MS you would have been interested enough to take a look at it. You would have grabbed the source code for both Vista and XP so that you could compare them, but since you are obviously just a contractor you don't have that ability. If you could do it you would find that the Vista code base is much cleaner. The 70% rewrite that was done was worth it. The new kernel is modular and agile. There is still room left for some performance tweaks, but from an engineering standpoint it is beautiful. Over the next few years this will become very important. It took years to get the embedded version of XP ready, but thanks to the changes to the Vista kernel we should have Vista Embedded ready in less than a year. Once we strip out the shell, the graphics, and most of the managed code we will have a nice version that will run on a fraction of the resources required on the desktop.

When people ask me whether they should switch to Vista I generally tell them to stick with what their computer came with. If they are ordering a new computer then I ask them what they are ordering and recommend Vista if they are ordering a powerful machine. There are currently a few issues left with some applications, and there are many drivers that are not yet available for Vista. However, that situation is changing rapidly, and when Vista SP1 is released many of those problems will be fixed. A large part of the problem is that in order to make security better there were massive changes to the interfaces between user and kernel space, and the entire driver subsytem was rearchitected so that all drivers run in user space. We painstakingly went through and added as much backwards compatibility as we could into the system. However, there are literally millions of Windows programs that have been written, and we do not have copies of all of them. When a user finds a new program that doesn't work we do add it to a list of programs that are known to not work. Developers are constantly working to add back-compatibility support for the applications that don't work based on the popularity of the application.

A rough translation to human speech...

[dyfet]

"Your making us look bad, cant you lie a little, we do all the time..."

This was a public service translation, for those who have trouble understanding Microspeak...

Its about the bug, not the environment

[Anonymous Coward]

This guy is IMO a narrow minded fool. Sure, Vista may have extra security features which can limit the extend of damage which a certain bug can do. But does this mean that these features have any impact on the severity of those bugs? Lets "translate" this to Linux:

Say a new local SSH exploit has been found allowing attackers to gain root privileges. Does the fact that you'd need user accounts which are actually useable by people make any difference on the severity of the exploit? "Gee, cut the homeuser some slack since they won't have any real user accounts to begin with. So stop scaring them and rate the bug as it really is?" ? But... The bug really is what it says to be. In my example its a critical issue, in the case of a Vista bug its Important.

Just because you may benefit from the extra security enhancements doesn't imply everyone else does. So please; cut out the idiocy and the desperate attempts to push Vista forward by focussing on all good points and ignoring the bad points, and simply keep calling things what they are. I for one now question the professionality of this guy.

Re:Its about the bug, not the environment

[NearlyHeadless]

If you've read Michael Howard's writings, he's certainly not a "narrow minded fool". On his blog, he talked about security features in the compiler and linker such as /GS and /SafeSEH. With these in place--and OS-based onese, such as Address Space Layout Randomization and Data Execution Prevention-- buffer overflows still exist, but are much harder to effectively exploit. Yes, the process will abort, so you could still have a denial of service attack, but you've greatly reduced the chance of a more serious remote code execution.

Note that OpenBSD is also adopting similar defense-in-depth strategies, including SSP and N^X. Adoption is much more haphazard on Linux Distros, so you may be at much more risk running an application such as SSH on Linux than on OpenBSD even when it is compiled from the same source code.

Re:Its about the bug, not the environment

[OmegaBlac]

Adoption is much more haphazard on Linux Distros, so you may be at much more risk running an application such as SSH on Linux than on OpenBSD even when it is compiled from the same source code.

SSP is included with recent versions of GCC 4.1 and above. If your specific distro is using GCC 4.1 or newer, then they are compiling with SSP already.

http://gcc.gnu.org/gcc-4.1/changes.html

Re:Its about the bug, not the environment

[hxnwix]

Right, and that's why OpenBSD pretends that remote exploits are warm & fuzzy happy ponies. Because of their "baked in defenses." ...
Errr, NO , this guy promulgating deceptive doublespeek. But perhaps he knows better - perhaps he's just a dishonest jackass and not a retarded jackass. What was your point again?

Re:Its about the bug, not the environment

[penix1]

Because the Microsoft security people are totally conservative when it comes to measuring risk, they assume that every one of these mitigations has been bypassed (or disabled), and measure vulnerabilities accordingly.

And that is a correct assumption to make. If a security "feature" can be bypassed or disabled, you can't make any other assumption. I firmly believe the biggest threat to Microsoft security is Microsoft itself. Policy from one section of Microsoft is fighting policy from another section. The security folk are fighting the "ease of use" folk. The piracy folk are using the critical updates as a means of checking legitimacy. WGA thinks you're not legit? You stay vulnerable making Microsoft a menace to networking. All these are policy fights that make being a Microsoft user less and less attractive.

B.

Re:Its about the bug, not the environment

[driftwolf]

If Vista is so much more "secure", then any flaw should be much MORE serious, not less. After all, aren't they supposed to have worked so long and hard to reduce the flaws in this one? If one advertises a secure system, then any breach is, by definition, important. MS Vista is being pushed as a highly secure system to many businesses. Hence, security issues are that much more important, as they were used to sell the system in the first place.

As we've heard that much (some?) of their vaunted security is actually just optional smoke and mirrors (several of the user security features for instance), I don't think MS Vista should be given any easier ride than any other operating system. Let it be judged independently, on its own merits, and not through re-definition of what is critical or not for political (and of course publicity and monetary) purposes.

Any system that defines itself as "secure", but isn't, deserves to be ranked accordingly. Microsoft (and it isn't alone by a long shot) has a very long history of selling one thing and delivering another. Changing the criteria based on what they are selling isn't warranted until what they deliver matches that in every respect. So far, they aren't doing that with MS Vista either.

Re:Its about the bug, not the environment

[kscguru]

His security features are /GS, /SafeSEH, layout randomization and an execute bit? Okay, he really is full of it.

• /GS. In theory works fine. In practice, you MUST (1) get the software publisher to compile with the switch, (2) cannot use inline assembly (/GS bails out on such code), and (3) must be willing to sacrifice a small bit of performance. In other words, a fair amount of real-world code can't use this. And oh by the way, this doesn't protect against all buffer overflows - it only protects against the easiest category. It's still quite possible to corrupt data with a buffer overflow, and maybe use that data to gain control.
• /SafeSEH. Right ... how many common languages don't have good exception handling? You said C only, right? And how often do you use Windows exceptions in C? Not much, you say? When I've seen SEH code, it's almost always very narrowly scoped and thus easy to get right - in real code, Windows SEH is just a trampoline to get into another exception mechanism. Making it "safer" adds no value.
• ASLR. This one makes generating a sucessful exploit a little more difficult - moves it from medium-easy to medium, because it's harder to hit a "target buffer". Of course, for compatibility reasons, a fair number of apps turn this off (they have assumptions about where code lives, and/or need the wasted address space). It helps - statistically. But a lucky guess is still going to succeed, and I don't trust luck for security.
• DEP. A two-pronged technology, which (1) uses the NX bit and (2) disallows syscalls from data segments. Oh but wait, (1) requires having a fairly recent processor and (2) is fine for some apps, but breaks for anything that does dynamic code (e.g. a Java runtime), so it's also disallowed for many, if not most, apps.

So what do we find out from this list? You get defense-in-depth - IF you are running the latest hardware, IF you use only software built with MSFT's favorite options (some of which are opt-in), and IF you only run apps that embrace all these strategies. How many Joe Consumers fit into those ifs? Datacenters might be closer, but I'll bet even they can't generally say all these hold true.

I'm glad open-source is adopting some of these measures. But let's be realistic - all any of these technologies do is make a sieve less leaky by putting a second sieve underneath. Something is nice, but we would be fools to treat any of these security "features" as more than a speed bump.

Re:Its about the bug, not the environment

[LO0G]

Has he EVER refered to those as "security features"? I'd be surprised, Michael Howard doesn't usually make those kinds of mistakes.

Usually those are described as mitigations, since there are no security guarantees associated with them (since they can be bypassed, they're not security features.

You keep using that word

[Gothmolly]

I do not think that the word "security" means what you think it means.

Or, you're a FUD-peddler whose job it is to convince Gartner that you don't suck... I'm not sure.

New rating for new system?

[Jimbitz]

I can't believe someone known as microsoft security guru would make a statement like that.
An exploit is still an exploit. It doesn't matter if it's found in a brand new OS or the predecessor.

Thank god there are people who doesn't agree with him.

Re:New rating for new system?

[GIL_Dude]

Well, I think the point would be something more like this:

A buffer overflow is found in lsasrv.exe. It's remotely exploitable on Win2k3 server and Windows XP and can run arbitrary code and doesn't require an account on the system (remote wormable). It's only locally exploitable on Vista, requires a local (even if low privileged) account to be logged on an run the code (possibly via social engineering - click here for SomeStarNaked.exe).

He's talking about the rating - a rating should be in relation to something. Otherwise - what does "5 star movie" mean? Is 5 stars the best? Is it 10 stars for the best? So, you need a rating that puts them in relative perspective. In this case, the same overflow should get an "extremely critical" for XP and Win2k3 server. It MAY not deserve as high a rating on Vista though depending on its ability to be exploited and spread. Possibly on Vista it could get just critical or maybe even just important.
I think it is key when rating the vulnerability to take into account how it can be utilized and what is required to exploit it.

Re:New rating for new system?

[rbochan]

Yeah, threat rating: "waaah... security is hard!"

This is not wise

[EXMSFT]

Don't challenge the hackers. It's great that Windows Vista has some built in low-level security protections. It's also great to see that Michael is discounting the significance of UAC. And he should - most people will wind up turning it off. But I think that attempting to say that Vista is fire retardant is most likely going to serve as a method to encourage hackers and script kiddies to try and set fire to it. Saying "because it's Vista means the exploit isn't as bad" is a horrible argument. It's an OS, and an exploit is an exploit.

In short I don't think Michael should assume. When you assume, well, you know.

stop whinning and just....

[3seas]

...fix the bugs.

Re:stop whinning and just....

[rucs_hack]

They can't

Not because of anything so simple as crap coders or Microsoft being shit (lame reasons when there are so many others that can be justified with examples) . They can't because it's too complex, subject to too many attack vectors, and closed from peer review of code.

Time was this refusal to allow external entities to search for and fix bugs in their code was acceptable as normal business practice. Since Linux got more popular, people have started to see that peer review of code is superior when it comes to finding and fixing errors.

I'd be willing to bet that if Linux was closed source it would be as defective as Windows is. That it isn't testifies to the usefulness of open source/bsd style approaches.

Re:stop whinning and just....

[tuzzer]

I'd be willing to bet that if Linux was closed source it would be as defective as Windows is. That it isn't testifies to the usefulness of open source/bsd style approaches. Something being closed source doesn't mean it can't be peer reviewed. We use peer reviews at my job all the time. The rule is you don't check your own code, others do. It helps. A lot.

Re:stop whinning and just....

[HiThere]

It does, however, meean that those who review it will be few in number, and will have a similar perspective. These are both strong indications that the peer-review is weak.

P.S.: Note that OpenSource programs with few developers interested in the code run into this same problem. Good peer review takes lots of eyes in multiple environments over an extended period of time. A structured code walkthrough just isn't the same thing. It helps, but it's not the same.

Missing the point

[UnknowingFool]

Why is it that MS always misses this point: Secure is relative. Advocating that MS can be more lax in its procedures because Vista is more secure is like saying you don't need to train anymore because you didn't finish last in a race. Microsoft may have better security than its predecessors; however, that remains yet to be seen whether or not it is adedquately secure. Given the companies history of boasting about security and then failing to deliver, it would be best if they were conservative when it comes to security. Wasn't there a recent slashdot article on how OpenBSD had an its second security issue in a decade? Compared to that, Microsoft security is a joke.

Tired article on a stupid statement.

[lancejjj]

Microsoft's own bug hunters should cut Windows Vista some slack and rate its vulnerabilities differently because of the operating system's new, baked-in defenses, according to [Michael Howard, a senior security program manager in Microsoft's security engineering group] who is often the public persona of the company's Security Development Lifecycle (SDL) process. Microsoft shouldn't have this guy as the "public persona" of security if he isn't 100% within both the security & public communications loops at Microsoft. "Vista" is supposed to be all about security. Why are they having this guy "chat" about it when he isn't a communications expert and when he isn't representing Microsoft's corporate opinion?

I'm sure we've all said a few things that were externalized "thought experiment" instead of "well thought out conclusions". And I think I can see how his line of thinking was going, although I disagree with his statement. And I wouldn't be surprised that in hindsight he disagrees with his own statement.

Microsoft has inadvertently set this guy up as a fall guy by anointing him as a semi-official spokesperson. Hopefully he won't find himself on the street due to what is a failure of his management.

Obligatory

[dkleinsc]

You are trying to cover your own ass. Cancel or Allow?

That's a hard one.

[pilsner.urquell]

Lets see, Microsoft has been selling crap all these years and now wants to be cut some slack? Yea, right.

A little late for that...

[Jasin Natael]

By this logic, then, shouldn't most of the bugs for Linux and OSX have been rated as "relatively unsafe", while the Windows bugs were almost universally labeled "Über-pWnz0r3d"?

It seems like he wants this just so he can compare turds to turds, boosting the sales of Vista by saying the Windows 98 and 2000/XP bugs of yesteryear were worse because the same bug is arguably less severe under Vista. It may be true, but he should hope that if anyone takes him seriously, they don't start rating severity relative to similar bugs in competing products.

Be careful what you wish for...

More like "half-baked"...

[Anonymous Coward]

"Built in defenses".

Yeah, right. He's been reading too much William Gibson...

softer...

[beando]

Vista making microsoft became microsofter...

It would seem..

[ChePibe]

That Mr. Howard has yet to come to the sad realization that the rest of the Vista-using world has...

MS Security Guy?

[Psx29]

Are you sure it wasn't a PR guy?

Awww

[Centurix]

They're hurting your feelings, come here and rest on my man boobs. There there, that's better isn't it mr security person. What, they're not as soft and comfortable as your moms boobs? Excuse me, I'd like you to rate my boobs better than that, after all, I am a MAN!

Of course!

[RMingin]

Obviously any Vista security bugs should be rated less severe... I mean, nobody's running that OS, right? Minimal impact!

An interesting response

[Trelane]

From a Red Hat hacker

Comment removed

[account_deleted]

Comment removed based on user account deletion

baked in?

[DragonTHC]

in Linux and Unix and Mac's BSD, what's higher than root?

in Microsoft Vista, what's higher than administrator?
    root
          superroot
                supersuperroot

that's right, there are three privilege layers above administrator in Vista.

users cannot access those, but software can.
"Oh, you're a process, here's the keys!"
"Oh you're a user? You want to access your computer, confirm or deny?"

Re:baked in?

[pallmall1]

that's right, there are three privilege layers above administrator in Vista.

Do they have to press a button to get to ludicrous root?

OS bakers poem

[bl8n8r]

wake-n-bake lets all take
a look at microsoft half-baked
hit the bong and sing this song
windows got security wrong
Around we go with disclosure fud
Michael Howard please pass the bud

I think MS needs to talk to a lawyer

[SmallFurryCreature]

Simple send each and every person who works for the company in anyway to a lawyer and tell the obey the first rule.

SHUT THE FUCK UP

Just stop talking, do NOT say anything, remain silent.

MS just can't do that and keeps blurting out things that make it seem extremely silly indeed.

This latest claim is like saying that a grease fire in your kitchen isn't dangerous if you live near a firestation. That getting shot through the chest isn't as much a of a hassle and shouldn't count as an attempt on your life because you happen to be in a emergency room.

A bug, is a bug, a security hole is a security hole. That they are even rated is already bad enough. They should have just one variable "fixed" wich is a boolean.

Claiming that a so called critical bug isn't as severe because the unproven untested OS it runs on has some safety measures, which by the way have been programmed by the same people who programmed the bug, is not exactly raising my opinion of MS.

Had they simply listened to the lawyer they would have kept their mouth shut and not dropped another notch in my estimation.

Perhaps it is all part of a cunning plan with them hoping that humans like computers suffer from wrap around and if they lower my opinion far enough it would wrap around to positive again.

or they are stupid.

But I liked the end, unless Vista picks up it will receive the same non-attention as OS-X, now that gotta smart.

They want to do something about security?

[argent]

The first thing Microsoft needs to do to get ANY credibility at all where security is involved is to take immediate and rapid steps to eliminate the role of the HTML control as an element of the security system.

That means getting rid of "Security zones". All documents displayed by the HTML control must be considered "untrusted".

To do this, start by getting rid of the ability for documents viewed in the HTML control to request the use of ActiveX objects, since no documents are considered trusted, ActiveX can't be used anyway.

At the same time, provide a mechanism like IO Slaves for applications to install controls... a mechanism that can not be requested by a document.

Modify Windows Explorer and Software Update to use this application-controlled mechanism to install components into the HTML control.

Create an IE shell that installs an "ActiveX IO Slave" to restore the existing behaviour. This shell will display windows with some visual indication that they are untrustable and dangerous. Users who acually require this functionality during the transition can run the "Insecure IE" shell.

In the next major release of Windows, remove that component.

Conservative?

[julesh]

"The MSRC folks are, understandably, very conservative and would rather err on the side of people deploying updates rather than trying to downgrade bug severity"

Err, right. So if they're so conservative, how come they'll rate a remote code execution bug as "moderate" if the code is run in a restricted context (see, e.g. http://www.microsoft.com/technet/security/Bulletin /MS06-013.mspx - particularly the DHTML bug)?

It's Simple - This is the "new" Microsoft spin

[Master of Transhuman]

Vista is going nowhere, so now they trot out some bozo to say that Vista security problems won't be as bad as XP's.

Then they also had some Microsoft bozo post on his blog that he was going to compare vulnerabilities - actually, not even vulnerabilities but FIXES - between OS's - using the same discredited methodologies they've been using since forever. Naturally Windows came out ahead. He even tried to head off criticism by admitting he was a Microsoft bozo. Naturally, that didn't work.

In other words, Microsoft is trying to spin Vista's failure to be a "Windows security cureall" - especially since OneCare has been a PR nightmare by failing antivirus checks and then deleting users Outlook email files.

It's just another pathetic Microsoft pack of lies.

Remember, folks: ANYBODY authorized by Microsoft to talk to the public is a LIAR.

Microsoft does NOT sell software. It sells LIES.