Ballmer Talks XP Extension, Walking Away from Yahoo!

BTW: i told him to look on the Windows Marketplace for "Certified for Windows Vista" scanners, and he ended up looking at some of HP's page scanners. the price of new ones is the same as the price he was paying for used Fujitsu page scanners from some used business tech supply store when he bought them 2 years ago (before he was my client). the current models all feature WHQL-signed drivers for Windows Vista x64 too. he hasn't made a purchase decision yet though, since he's an accountant, and is busy with tax time right now....

XP

"64-bit Vista requires signing, but any company can obtain a signing certificate for a small fee. Garbage drivers can easily be created and signed."

not through the proper channels they can't. i'm talking about following the "correct" path to getting drivers signed and certified - through WinQual. the fact of the matter is that many hardware companies (even big ones) just refuse to sign/certify their drivers. one such company i found out about recently was Fujitsu, specifically for their scanners. a client of mine had purchased several (before i was their "IT guy"), and also recommended them to his clients. while working on his IT setup, he was looking at taking advantage of more memory with 64-bit versions of Windows, but he couldn't due to Fujitsu's ignorant attitude about driver development. they make 64-bit drivers for only a handful of models, but all of them are unsigned (neither are the 32-bit drivers btw), so of course, Windows Vista x64 will have you acknowledge the risk of instabilities at startup. i talked to one of the tech agents from Fujitsu, and he said "Fujitsu has no plans on certifying their products with Microsoft or digitally-signing drivers for Windows in the future, for any products" (!!!). well la-dee-dah. i know one company that i won't be recommending, or selling in the future then, simple as that. this is total foolery as we move into the 64-bit mainstream. Panasonic is another company that acts the same way, also for scanners/copiers/printers. both companies are gonna getta hurt real bad when they realize they can't just port aging Windows 2000 driver code over again in the future.

if you want to see what happens to companies that can't properly support their products with decent drivers, just look at Creative, and look at how their marketshare has been completely decimated by companies that CAN write drivers, albeit for inferior hardware (specifically, audio codecs for Azalia HD controllers).

XP

Christopher, Waethorn, tayme, thanks for a return of "Windows IT PRO". I didn't fully understand why I would have to replace my scanner, webcam etc before I could change from Vista 32bit to Vista 64 bit, and why I couldn't just upgrade Vista rather than re-install. Now that you guys have given me the facts, I can make an informed decision about how to migrate my system to Vista 64 bit. Great work !!

"64-bit versions of Windows Vista now REQUIRE driver signing"

64-bit Vista requires signing, but any company can obtain a signing certificate for a small fee. Garbage drivers can easily be created and signed.

Signing prevents is 3rd parties from hacking drivers and releasing them as legitimate. For example, certain hardware vendors have had their drivers hacked by enthusiasts and re-released on the web.

You seem to be getting signing confused with WHQL certification, which is unrelated.

"I mean, from an app standpoint, how hard is it to port an app from IA64 to x64 (or vice-versa) if it uses cross-platform API's?"

Most apps are portable depending on their complexity. I couldn't necessarily speak to IE, etc, because I'm not sure in what ways it could interact with various x64 APIs.

Creating a 32-bit app and going forward to 64 should be simple, going the other way requires prudence, and trying to create something that fully leverages everything is more difficult. There is API compatibility, but each of the 64-bit kernels expose methods that simply don't exist on x86. Plus RISC/CISC chips fundamentally excel in different tasks, so one set of variable declarations could be zippy on ia64, but slow on x64 (the app might compile, but not be something you'd be proud to pin your name on).

Another problem is actually less than obvious. Computers running x64 have statistically identical hardware configurations to x86 machines. As a result, once you have access to the extended resources of x64, you need to be careful not to assume an abundance of capability that isn't physically present. In some respects, x86 development is simpler because it forces you to deal with an address space that probably equals physical resources (at present). x64 may once again force coders to learn tighter memory management.

"driver support is generally better on the 32 bit platform (a *much* larger base of legacy hardware)"

let's not forget that it's not just legacy hardware, but legacy drivers. 64-bit versions of Windows Vista now REQUIRE driver signing, so it is less compatible with the ****-poor driver development of some hardware companies. it's bad for those companies, but better for the end-user in the long run. those hardware companies that don't change, and keep recycling their 32-bit drivers over and over need to adapt to a 64-bit world, or else fall by the wayside.

"A good portion of the kernel is still in assembler specific to that platform."

i completely understand that bit, but i was talking more about the other 64-bit components, such as apps like IE 64-bit and 64-bit API's. the main thing i'm wondering is how much app support is cross platform....i mean, from an app standpoint, how hard is it to port an app from IA64 to x64 (or vice-versa) if it uses cross-platform API's? what are the complications/disadvantages/comprimises/consequences of making it cross-platform? how much of the 64-bit hardware is "virtualized" (term used loosely) behind API's?

XP