ITunes 8 a Real Killer App; Taking Down Vista

CWmike writes "Apple 's latest version of iTunes crashes Windows Vista when an iPod or iPhone is connected to the PC, scores of users have reported on Apple's support forum. Plug in and Vista crashes and shows the 'blue screen of death.' The errors began showing up immediately after updating iTunes to Version 8.0, which Apple released Tuesday as part of its iPod refresh. 'I just installed iTunes 8 over my iTunes 7 on Vista [and] now whenever I plug in my iPod, I get a blue screen death. Three times so far. Even if it is plugged in on boot, I get a blue screen," said a user identified as 'sambeckett' on the support forum about 90 minutes after Apple CEO Steve Jobs wrapped up the iPod launch."

Comments

Good Marketing

[TheSpoom]

Expect Apple to blame Vista.

Re:Good Marketing

[Binder]

Regardless of what the itunes application might be doing it should not cause the OS to crash. The application to crash yes... but not the OS.
One of the main jobs of the OS is to protect processes from badly behaving neighbors.

This is definitely a bug in Vista.

Re:Good Marketing

[fo0bar]

Yeah, because an user running an app causing the entire OS to crash for all users is totally acceptable. Unsigned drivers making the OS crash may not be Vista's fault, but this definitely is.

Yeah, because iTunes totally doesn't use a custom proprietary USB protocol to talk to iPods that would never require special drivers.

Re:Good Marketing

[LWATCDR]

Here let me explain how computers work to you.

Problem: Application crashes the OS.
Blame: OS. Modern operating systems should not crash because of an application.

Problem: Device driver crashes OS.
Blame: In a monolithic OS the driver is at fault. So it is the drivers fault. If the driver is approved by the OS manufacturer then you can also blame the QA department of the OS manufacture.

Problem: Application crashes.
Blame: Application but maybe the OS if the Application works on a different version of the OS. No sane programmer uses undocumented interfaces any more. It is too risky and computers are fast enough. So the cause is broken compatibility.

So in this case I would have to say that it is Microsoft's fault the OS crashes. It is Apples fault for not doing enough testing.

Re:Good Marketing

[Hyppy]

The special drivers are still signed by Microsoft. If they weren't it would be quite obvious due to the many "Arte you SURE?" messages.

Re:Good Marketing

[Anonymous Coward]

No it's not. If you are going to deploy an application and you are a well funded commercial entity it is your burden to test it on whatever platform you plan on supporting. While I am not letting Vista off the hook for this flaw you cannot say Apple is 100% in the clear here. Either they didn't test it, which is incompetence, or they didn't care.

But I am sure if the next version of Microsoft Office somehow crashed OSX, the conspiracy nuts would be in here complaining about how Microsoft is trying to tarnish Apple's good name.

Re:Good Marketing

[Shadow Wrought]

But if I build a house that isn't up to code, and it passes inspection, it's the inspector that looses his job when the house collapses.

True, but it is you who is buried under the pile.

Re:Good Marketing

[UnknowingFool]

Either they didn't test it, which is incompetence, or they didn't care.

So there exists only 2 possibilities here? I offer another. They did test it, but they couldn't test every single hardware/software permutation that exists which is not realistic. The fact that XP users don't seem to be affected points to problem unique only to Vista. I'm not saying it is not Apple's fault. It may be but until we have better understanding of the issue realize that this isn't the first application to crash Vista, and it probably won't be the last.

Re:Good Marketing

[Moridineas]

If Windows didn't allow such extensive use of making every bit of software installing useless drivers and daemons

I'm not sure how to parse that...what do you mean?

With Windows apps, it's more than frequent, it's the norm.

What other windows apps do you run that require drivers to be installed?

Re:Good Marketing

[teridon]

iTunes is "GEAR powered"; i.e. it installs the GEAR CD/DVD burner ASPI drivers. See:
http://www.gearsoftware.com/support/drivers.cfm
http://www.gearsoftware.com/wiki/index.php?title=GEAR_Powered_Products#iTunes_for_Windows

Wouldn't those be kernel-space drivers?

Re:Good Marketing

[moderatorrater]

I have never had a blue screen of death in Vista ever. Period. This is after a year of using it with many different programs and many different hardware configurations. It's much more stable than it's predecessors.

I have never had an Apple program be stable on the Windows platform. iTunes is bloated at best, absolutely unusable if you catch it at the wrong time of the moon cycle, and just generally not that good. Safari crashes constantly for me whenever I've tried to use it. I've had iPods not work at all when trying to use them with windows.

My money is on Apple being the fucked up one this time. Someone should verify that it requires administrator rights to install, and that will settle the deal. I'm guessing that they have DRM measures implemented in v8 that weren't in v7 and that's what's causing the problem.

Re:Good Marketing

[spun]

Eh? Bad analogy. Replace 'china shop' with 'cement barrier specifically designed to stop dump trucks.' That is one of the main things operating systems do, stop processes from interfering with each other. Understand? It is one of the fundamental reasons operating systems exist at all. To say it is not the fault of the operating system is to misunderstand the purpose of an operating system. Everything else (scheduling, memory and resource allocation) could be done cooperatively by each running program.

Re:Good Marketing

[Sandbags]

BSOD in Vista is either hardw3are level resource conflicts, or an actual memory fault.

I'm actually betting, based on the I/O is see in IT8 while it's building the thumbnail images, the massive background effort to create genius database info, and the high level of memory and disk I/O present in it's basic use that did not occur under IT7, that these machines were ALREADY FAULTY, but simply were not utilized enough to trigger these memory or hard disk faults.

The DRM is exactly the same btw. No changes. Besides, DRM is a user space application, and can not cause BSOD. It;s impossible for that to be the cause...

If you have not had vista BSOD, then all that means is your hardware is exceptionally well built and defect free, and that none of your components have resource conflicts with any others. My guess is your PC is OEM manufactured, likely by Dell, and is on the lower end of the spectrum (under $800 base system, that maybe you added a nice video card and some extra RAM to)

Vista may not BSOD on you, but I bet you have frequent application crashes... I don't typically go more than a few days without an application bombing out, my desktop refreshing from an explorer crash, my printer loosing connection, or an app just hanging and needing to be killed by task manager.

Sure, memory leaks may be a thing of the past, and generally when an app bombs, the machine stays up... My Mac has had those features for 6 years! When an app does bomb, I typically see in the logs that it;s a core driver or service at fault, and not even a file installed by the application.

Re:Good Marketing

[Sandbags]

Hey bud, I'm a software tester for a disaster recovery company. Let me put this past you:

- 4 identical hardware machines
- 4 exact copies of an ISO of Vista EE
- all 4 machines side by side, I make the same click on the same screen on each one.

After installation is completed with exactly the same settings on all 4 machines, we install our corporate AV program, then allow the machines to download and install updates. Each of the 4 machines has exactly the same list of updates added in the same order.

All 4 machines are benchmarked prior to and after installation, using 2 different tools from boot CDs. All 4 machines benchmark the same (within less than a second on a multi-million CPU cycle run, and withing 5 seconds on an multi-hour RAM test. No machines exhibit RAM, disk, or CPU issues and are regularly burned in and tested both before and after installations.

of the 4 machines, 1 has a C: with 3.9GB used space. 1 has 4.7GB used, 1 has 5.2 GB used, and 1 has 8.2!

One machine boots Vista in less than 22 seconds, one takes longer than 2 minutes. others are inbetween. Hard drive I/O pattern tests are run and all 4 drives exhibit nearly identical characteristics and I/O patterns for randomized I/O read/write testing.

One machine has 860MB or RAM free out of 2GB, one has less thqan 200MB free out of 2GB. Others fall between.

When we blank the drives and repeat the tests, machines exhibit completely different results. Sometimes the slowest one to boot in one test wiull be the fastest one to boot in another. Swapping components from system to system was no impact on the performance of a machine.

We started doing this test a few years ago when I was setting up several exchange 2003 servers to be used in a classroom and noticed simalar wild divergence between system performance, install size, and more, even before service packs or patches were added. We have repeasted this test with every version of 2000, XP, Vista, 2003 servre, 2008 server, as well as older OS. Anything NT or previous, and installer seems to be very consistent. Anything 2000 or later, very inconsistent. The newer it is, the worse the installer inconsistancies.

We have also done this test by repeating installations on the same single machine over and over, and the installation size on that machine is just as inconsistent.

So, what's the deal? Why does the same installer batch file, which is basically a top down program that collects data based on pre-defined rules, and processes installation order based on documented, databased information, produce such inconsistancy?

As a result, we no longer test product on a single machine, but we bought 4 each of 5 different machines, and installed the OS seperately on each one, then burned images of it. We test each application agains all 20 OS images with the same OS on each one, then swap images for each OS service pack and each OS supported. This can mean teasting a single application against over 70 Windows versions, each on 20 machines. This is a test base run of 1400 installs. After this, we release internally, and install to roughly 100 machines. We purposely buy only 2 or 3 of each model from a manufacturer to deploy in the company, so we have dozens of different machines, and allow each user to basically maintain their own box the way they like, creating what in other companies would be an IT nightmare, but in this case it;s a benefit (and each user is an admin on some level, and completely capable of maintaining their own machine). We'll find a dozen bugs we didn't find running through the 1400 image test run. After release to the field, when roughly a thousand customers get a hold of it, we'll find more bugs...

You CAN NOT test for every hardware reviosion on every machine made, and every OS that could be on it. If you're asking Apple to do that,. you're expecting them to have access to over 10 million differing system images to test on. This is IMPOSSIBLE.

The fault is ENTIRELY microsofts. If the soft

Surprising

[ohxten]

That this wasn't caught in the testing stages?

MS or Apple

[JohnVanVliet]

well is it Apples fault OR is it MS Vista that has the problem . Apple would have debugged it on a vista box , But it is posable that a vista update killed it .By mistake or on purpose ,your guess is as good as mine.

The new "better" driver model sucks.

[Sj0]

You know, it seems like there are an awful lot of problems with drivers under Vista. Certainly far more problems than I've seen on Vista.

The thing that bothers me about that is the change in driver architecture was billed as a way to make Vista faster and more stable. Why, then, is it that most of the drivers for Vista are less stable and slower than the same hardware running in XP?

Re:The new "better" driver model sucks.

[truthsearch]

For every major Windows release vendors spend a lot of time and money on new sets of drivers. My guess is that to save time and money some of them didn't rewrite their drivers from scratch for Vista, but instead migrated as much code as possible. That would let certain problems slip through the cracks, such as the kernel level security issues we've heard about.

But still...

[Tmack]

A driver should not cause the OS to crash. Your printer should be able to load its driver in a manner such that if it catches fire the kernel stays alive and can tell you so.

Tm

Re:But still...

[Anonymous Coward]

Maybe for userland drivers like printers under Vista using the latest driver model, okay, but if you mean that no driver should ever be able to crash the OS, you clearly don't understand how drivers work.

Re:But still...

[Jugalator]

Oh and yeah, sound cards must also now (in Vista) run more in user space.

That actually caused a ton of people to complain on Microsoft as it could no longer as easily do Creative EAX. Damn if you do, damn if you don't.

Re:But still...

[cbreaker]

But, Apple does install drivers, and those drivers CAN crash the operating system.

It's no different from any other popular operating system. If you have a bad OSX driver - boom. Grey box. If you have a bad Linux driver - boom. Kernel panic.

The only utter nonsense is that Apple can't write a driver that doesn't crash the operating system. There's tens of thousands of drivers out there, and most of them run great. Apple is big enough to do proper testing. They didn't QA properly, obviously.

Re:But still...

[SanityInAnarchy]

One of the big reasons no one had any interest in minix is the incredible performance hit the design entails.

At what point?

I think the main reason no one had an interest in Minix was the cost, and the restrictive licensing. Linus admitted that he never expected Linux to be much more than a stopgap until GNU/HURD was released -- except that HURD took too long to get any kind of release out the door, so Linux already had adoption at that point.

The best argument at the time was: You could spend the money and buy Minix, and install it. And then install the source code, and download a number of patches needed to get something approximating a modern OS, recompile, and reinstall.

Or you could spend that same money on a faster computer (a 386), and get Linux for free. Linux could do everything Minix could, and it already ran in native 32-bit mode (which Minix needed patches for).

In fact, that's one of several other ways Minix took performance hits -- ways that I'd call bad design. The filesystem, for example -- Minix has a non-threaded filesystem; Linux had a threaded filesystem.

Who uses iTunes on Windows?

[gelfling]

Geesh the last three versions were bloated to the heavens. I think it takes about a minute to start now. A least few recent versions flat out crashed or did not update correctly. And is there any piece of application software that takes more time to update than iTunes? I don't think so.

Both Vista Users Are Horrified!

[Philip K Dickhead]

In other news 1000 Ubuntu users running under WINE without whining.

I Blame DRM

[kitgerrits]

If iPods were simply accessible as a USB mass storage device, I don't think there would have been a problem.
From what I can see, Apple uses a proprietary device-type, so they can talk to it using an encrypted connection.
All that, simply to keep you from copying files you supposedly don't have the right to copy.

Re:I Blame DRM

[99BottlesOfBeerInMyF]

If iPods were simply accessible as a USB mass storage device, I don't think there would have been a problem. From what I can see, Apple uses a proprietary device-type, so they can talk to it using an encrypted connection.

Apple uses a proprietary device type so they can hash the files so reading them uses less battery than it would if it was arranged like a storage drive. I've never heard of any "encryption" for the USB connection and numerous other programs have figured out how to read the hash tables without any problems (e.g. Amarok, Banshee, Floola, gtkpod, MediaMonkey, Rhythmbox, SharePod, Songbird,Winamp, YamiPod). If Apple is trying to stop other programs from interfacing with iPods they're doing a lousy job of it.

All that, simply to keep you from copying files you supposedly don't have the right to copy.

I don't think you know what you're talking about.

Re:I Blame DRM

[leamanc]

Sorry, there is no DRM involved in mounting the iPod. It basically is a USB mass storage device. All Apple's drivers do is detect when a USB mass storage device matching an iPod's filesystem is plugged in, and launch iTunes. Seriously, that's it.

The way that Apple keeps you from copying files is by hiding the directories that contain the music files. The files are then scattered amongst a bunch of obscurely-named directories to make it a little more difficult to find them, after you figure out how to show the hidden directories. (On a Mac or Linux, it's as simple as "ls -a" in a terminal.) An iPod database file is how the iPod and iTunes keep track of what files are on the device, and where to find them. Dozens of other applications (including Linux music players like Rhythmbox or Amarok) have figured out how to read the database. There's also apps that read the database and let you copy files directly, like the Mac app Senuti.

The only DRM involved is in files purchased from the iTunes store. You can access and copy these files, but you just can't play them unless your computer is authorized for the account that purchased them.

Re:I Blame DRM

[99BottlesOfBeerInMyF]

Rubbish. The hashing would not even save the most trivial amount of power, and may actually cost a minute amount more, since the Song Name now absolutely has to come from inside the file instead of from the filesystem, meaning every "directory listing" on an ipod has to open up every file, parse the id3 tags and cache the results for display instead of simply showing the filename.

Funny I just read an IEEE article on forensic analysis of iPods that disagrees with you. [Pod Forensics: Forensically Sound Examination of an Apple iPod, Jill Slay; Andrew Przibilla, System Sciences, 2007. HICSS 2007]

This is simply a lame way of satisfying the RIAA that Ipods are not used for piracy.

I have no doubt that Apple did make certain things harder to do using iTunes at the request of the RIAA, but I don't think that is the reason for their hashing, which is so easily bypassed.

The 3G nanos have built in encryption and have yet to be hacked by any of those programs...

Are you talking about the checksum it took the Amarok people all of two days to work around?

...and the Ipod Touch doesn't even offer disk mode anymore, most likely to keep the hackers out.

The iPod touch is using the same firmware as the iPhone so they keep it locked down the same way. All the other iPods work fine with third party music jukebox software I've used.

Here is a test:

[geekoid]

Try plugginh an HP printer into the same usb Channel as your iPhone.
This will cause the crash.
In the scenerio I presented to you, whose fault would you think was the crash if you ahdn
t read this story?
Probably HPs.

Just an example of how overly complex windows driver architecture is.
This is why I feel we should go back to the applications installing everything it needs under a directory it creates.
Less mess, easy trouble shooting easy uninstall, not files scattered all over your system.

But

[Gr8Apes]

Maybe Apple did QA it properly.

THink of the bad press for MS this heaps ontop of Vista. What are people going to remember? iTunes crashes Vista? Or Vista crashes when you plugin an iPod?

Re:But

[Joe U]

I'll second the notion that iTunes for Windows is a steaming heap of crap.

You're being mean to crap everywhere.

iTunes on Windows is on par with Realplayer, complete and total shit. Treat it like you would a virus, kill it. Avoid installing it.

If you have to install it then use VMWare. VMWare traps the itunes shitstorm quite nicely.

Re:I had no idea that it was so successful...

[ozphx]

Wrong metric.

Apple is the largest supplier of DRM media via the iTunes store.

Microsoft is one of many vendors who has been strongarmed into supporting playback of DRM files. You think they want to spend money developing DRM shit, or snorting blow off hookers?

Content owners are pushing DRM the hardest. They get the most blame. Then the content providers that agree to push this bullshit onto their customers.