Okay, so technically no CPUs were harmed in the making of this Linux kernel. And at least according to Linus Torvalds, this was a good thing, not a bad thing. Your mileage may vary, of course, but it comes down to this: 80386 CPU support has been completely dropped from the latest Linux kernel.

With all the fanfare of this Git commit title, “Merge branch ‘x86-nuke386-for-linus”, the 386 backward compatibility was finally dropped from Linux. Ye Olde Torvalds states in the comment, “This tree removes ancient-386-CPUs support and thus zaps quite a bit of complexity,” further adding, “… which complexity has plagued us with extra work whenever we wanted to change SMP primitives, for years.” And he signs off with a heartfelt, “I’m not sentimental. Good riddance.”

In all seriousness though, the 386 support was only there for legacy embedded systems. While the 386 may have been all the rage when it was introduced in 1985, it wasn’t long before it was replaced by the 486 in PCs. But it stuck around all these years in various embedded systems, from early BlackBerries to aerospace microcontrollers because sometimes that just happens. But even that couldn’t save it forever, and Intel finally gave up manufacturing 80386 CPUs for embedded systems in 2007. But it took until 2012 for Linux to finally let go of the past.

Of course if you actually still use an embedded 386 microcontroller then you’re probably grinding your teeth right now. But let’s face it, maybe you really should have upgraded at some point, eh? Three decades is an awfully long time to continue support for any computer, and five years notice is enough time for even the most dysfunctional government agency to have found and tested an upgrade. So if you still need 386 support in Linux, you only have yourself to blame for being SoL at this point. It’s not like the writing wasn’t in big bold neon lights on that particular wall. Maybe you’ve heard of ARM?

The paradigm shift has begun. HTML5 is replacing Adobe Flash on mobile devices. And it’s not looking so great for Microsoft’s Silverlight either. (Well, if you ever thought it looked okay for Silverlight on mobile in the first place.)

There are a number of reasons for this, but the main long and short of it comes down to these factors:

1) ARM isn’t just ARM

Unlike x86 and such, each ARM core is somewhat different, and as a result Adobe got tired of trying to support them all with its Flash Player. The constant code tweaks were just more costly than the benefit of Flash on mobile devices.

2) i allow you not

Companies of certain mobile devices (Take a wild guess which one, Apple!) have rather objected to technologies like Adobe Flash on their phones and tablets and locked them out. Whereas HTML5 is for everyone.

3) Power to the people

Studies into usage of Adobe Flash on mobile devices have also shown it to be a power hog compared to HTML5. And since no one likes a dead battery…

So it’s official, of sorts. Adobe has no plans to release an update to their Flash Player beyond version 11.1 for Android and BlackBerry devise. Though while Flash for the PC will likely remain, one has to wonder how long it actually will once everyone starts catering to mobile devices by switching to HTML5 there.

Meanwhile, Microsoft hasn’t actually said that the same will become true of their Silverlight software. (Does anyone even use Silverlight in the first place?) That however hasn’t stopped people from drawing the same logical conclusion, that Silverlight’s days are equally numbered. Of course that the only mobile version of Silverlight is for the Microsoft Windows Phone 7, not exactly a market-shattering alternative to Apple’s iOS and Google’s Android, it’s entirely possible that Microsoft is too oblivious of the concept of multiplatform to understand it need not bother supporting Silverlight at all. In fact, Windows 8 on ARM doesn’t support Silverlight at all. Metro (the name for mobile-centric portion of the schizophrenic Windows 8 ) doesn’t do Silverlight. The PC-side of Windows 8 does. But the writing is very much on the wall for mobile Silverlight.

So frankly, for web development, if you’re not looking at porting to HTML5, you probably should be, because it won’t be long before you really don’t have a choice. Unless you think all of those people are going to put down their phones and lug around a laptop instead.

Hey, those of you interested on running Windows 8 on ARM, guess what!  The cake ARM is a lie.

But as if that we’re bad enough, we now have some more official word on just exactly how ARM will be supported. No, there will be no lightened kernel to run Windows with. Microsoft is not taking a page from Linux in any way. If you install Windows 8 on an ARM processor, you don’t get Windows. You get Phone. That’s it. That’s all that you get.

What does that mean?

That means that you can only install “apps” on Windows 8 ARM, and so far sounding like only through a Microsoft app store no less. Called, imaginatively, the “Windows Store”. There will be no running of full Windows applications, at all, ever, on Windows 8 ARM. (Though I don’t expect it will take long for people to “unlock” their Windows 8 ARM to install their own apps manually at the very least.)

At least as far as I can tell by reading between the lines. Because, frankly, Microsoft is not being very frank and clear with us on this matter. In theory it may be possible that Windows 8 ARM will actually be a full version of Windows, and not just Phone. In theory. In practice however this is of very little value if you cannot install an application onto it because of being locked in to the Windows Store. Further, even if any old application could be recompiled for ARM, there are bound to be bugs and kinks to work out. And that’s even if someone bothers to try, which most software companies will not!

How do we know these things?

Because Windows has been here once before.

Only back then it wasn’t ARM, it was the DEC Alpha. Microsoft made an NT4 compilation of it. And it almost even worked. Except when it didn’t. Because almost no third parties bothered to support it in any way. And in fact even Microsoft kind of didn’t, as their compatibility system, basically an emulator-on-demand as far as I could tell, failed often. Very often. And worse, on things as common as most major installation packaging systems. So just trying to install non-Alpha software often crashed. You couldn’t even get to actually trying to run the software. That was Microsoft’s idea of support. And third parties. It was an absolute failure!

And now Microsoft is doing it all over again with ARM.

So even if Microsoft at some point tries to claim that your x86-compiled applications can be installed and run on Windows 8 ARM, the cake is a lie.  It will never happen if third party software producers don’t likewise port all of their software over as well, which they won’t!

What you are effectively getting if you install Windows 8 ARM on your tablet is … Windows Phone 7+.

Not Windows.

If a software vendor completely ports their application for WinRT then that’ll run on Windows 8 ARM. But most won’t, because that’d be a lot more work than just recompiling for ARM, and most won’t even bother recompiling, let alone porting.

That’s the only truth.

Except that it’s actually worse than that! Because while you should be able to run Windows Phone 7 apps on Windows 8 ARM / Phone (Is there anyone who will be running Windows 8 on a PC that will actually want to run Windows Phone 7 apps?) you will not (if I’m reading things correctly) be able to run Phone (yes, Metro, we went over this in the first Windows 8 rant) apps on your Windows Phone 7 smartphones. They may be similar platforms, but not identical. Windows 8 apps are built on Windows Runtime (WinRT), where as Windows Phone 7 apps are Silverlight and XNA. Windows Phone 8 is backward compatible with Windows Phone 7, but not vice-versa. It is not a forward-compatible design. Sorry all of you Windows Phone app developers. (All two of you.)

It even sounds like Microsoft is taking this a segmentation further, and that Windows 8 will only be for tablets. Phones will be stuck on Windows Phone 7. So it’s not just a matter of backward compatibility, but also about … modem form factor? I mean seriously, what’s the difference between a 3g-capapble 4 inch tablet and a smartphone? Not bloody much!  And far less yet between a 4g tablet!

But so anyone who ever envisioned of running Windows on ARM, sorry, you’re SOL. The tabletard has commanded that “Thou shalt not be productive on ARM.” The most you can do on ARM is still run apps. No soup applications for you!

You can think of Windows 8 ARM then being just another tablet OS like iOS and Android. It might say Windows 8, but it’s not Windows.

And actually, I’m kind of okay with that part. Honestly, what I want is a full-blown Windows smartphone/netbook combo device. You know, a PCphone. A Smarter phone. ARM may be The Next Big Thing, but I’m actually okay with x86 being the processor to make Smarterphones happen. It’s not like Intel and AMD both aren’t trying to make low-power mobile chips for just this purpose. (Why nVidia or even -gack!- Via doesn’t do so though is beyond me.)

Still, it’s very disappointing that Microsoft didn’t take this opportunity to do ARM right by any sense of the imagination. To have actually run a full-blown Windows OS on ARM would have been nice to see. Of course with what that could do to the server world, maybe it was done this way for a reason.

I was recently catching up on reading about the Intel Atom architecture, or more specifically, about Moorestown and Medfield, and I have to say that Nokia’s ditching of Intel and Intel’s Moblin (married to Nokia’s Maemo through the new joint venture of MeeGo) for Microsoft’s Windows Phone 7 almost starts to actually make sense. And this from someone who loves developing software in Qt!

And it all comes down to one simple reason: Intel doesn’t know what they’re doing.

I know. It’s hard to believe. But really, it’s true!

Why do I say that? Because even with Medfield, Intel is still missing the PCI bus in their System on a Chip (SoC). Why is this important, you ask? Because operating systems like Windows still require a PCI bus being present to work. Which means that even though Medfield is bringing the Intel Atom CPU (along with graphics, sound, camera, memory, etc.) to the smartphone / tablet world, it has completely missed the bus (Har har!) of being able to run a full-out real OS like Windows.

Now it remains somewhat unclear to me if Linux can be made to run without the PCI bus. Being Linux, you would think it could. In fact I’m pretty darn sure it can be done.  Honestly, I’m not sure why Windows absolutely needs a PCI bus, frankly.  Maybe just some bad assumptions and Windows is too heavy weight to change?

But then why did Ubuntu Mobile development come to a halt? Why did Nokia ditch working with Intel on MeeGo and switch to befriending Microsoft of all things? Anecdotally, it suggests that Linux, in fact may not be so happy without the PCI bus either. Or that there’s some kind of difficulty involved. Frankly, I’m not sure even I believe this.  It sounds like FUD.  Especially as Linux has been ported to all sorts of small devices with micro kernels.  But one does have to wonder why this keeps happening to mobile-centric versions of Linux, an OS known for being ported to everything and anything.  It is an odd mystery what is up with Linux there.

And will Microsoft address this PCI bus dependency in Windows 8 if they’re already trying to port it to ARM for tablet use?

But so if that’s the reason these Linux variations keep getting abandoned (and that’s a mighty big if, I do admit), then Nokia’s freakout might actually start to almost make some sense. (Beyond simple Microsoftian funding, the almighty dollar.)  What would be the point of bringing x86 to the smartphone if you couldn’t actually run your OS on it so that you can continue to use the same applications that you do at work or at home on your computer? If you’re eternally stuck with just apps instead of real applications, then why bother to deviate from ARM in the first place? Granted, it’s still a true-enough thought process for Windows alone. Maybe Nokia just wasn’t convinced that Linux is any good on a smartphone? (Though after buying Trolltech, I’d find that really hard to believe.)

This could also, by the way, be the whole reason why Intel developed Moblin. Because if their choice to avoid the PCI bus when bringing everything else from an x86 PC into a smartphone prevented them from running an already available OS, then they would pretty much have to create their own operating system to run this deficient x86-almost-PC. That or wait for Google to port Android from ARM to x86-sans-PCI. Or wait even longer for Microsoft to do something.

Or, try to cram all of the separate chips for a full-blown Atom-based PC into a tiny smartphone package, a task not easily done!  Hence the need for a simple (and small!) SoC instead of separate chips.

Okay, so I get that from a power-saving perspective, the PCI bus isn’t ideal. And so it’d eat into a smartphone’s battery life. I get that. To an extent. I mean it’s not like they’re missing from laptops, netbooks, and UMPCs (like the Viliv S5 that can fit in your pocket) which are also run on batteries. Surely someone could think of some nifty power-saving way to put PCI bus support into a smartphone SoC like Medfield … if they had their head on straight.

But, alas, no. That’s what Intel’s Oak Trail is for: Atom in netbooks and tablets. That can run Windows.

Not smartphones though.  Because it’s not one SoC, it’s many chips.

Is anyone else confused?

What in the world is Intel thinking?

How do I even put this in a way that makes sense…

Okay. So Intel’s focus here is on crushing ARM. Intel is tired of missing out on this hugely growing market. Smartphones are the new sugar rush and everyone wants their candy. Intel can’t afford to neglect this market any longer. (Or any longer than they already have done by twiddling their thumbs while testing the waters instead of just boldly going forward and setting trends.)

ARM, a competing processor architecture, is frankly just nowhere near as mature as x86. It may be low power, but it’s also lesser in functionality. This is why your smartphones only run apps instead of full-blown applications. (Well, the main reason anyway. Apple’s strict near-fascist software development regime is another matter entirely and really only applies to their products.) ARM might be great for phones, but not for computers. Not yet at least. Maybe not even ever. Only time will tell there.  Though certainly many would try to debate the point, from both sides.  Honestly, I don’t care either way so long as at the end of the day I have full-blown OS and application software and a compiler to make my own software with.

But so Intel wants into the market that ARM is massively consuming. What do they do? Do they use all of the great versatility and advanced features of their x86 Atom line to show how you can make a smartphone truly smart?

No.

Instead Intel works hard on making a super-low-power x86 Atom-based solution for smartphones so that their answer to ARM has the same (or better) battery life.

That’s it.

That’s all Intel is doing. They have effectively dumbed down the PC to make it as useless as a smartphone. Instead of upping the smartphone to the capabilities of a PC.

So here’s a wild analogy for you: Say you’re a hunter. Say you’re hunting wild boar, a powerful and respected creature in its own right. It may be smaller than man, but it’s fast, agile, and has big sharp tusks. Do you, the hunter, get on your hands and knees and chase around the wild boar tooth-to-tusk? Of course not! You’ll get shredded! That’d be downright daft to fight a wild boar on its own terms. No, you, the hunter, get out your spear, or your bow, or your gun, and you take the wild boar down with your superior technology.

Intel, get off your darned hands and knees, stop trying to out-bite the boar, and pick up a gun already! Or the ARM boar is just going to continue to gore you in the face!

Microsoft would like to remind us that Windows Vista Service Pack 1 support has ended, and if you’d like support on Vista, you really should upgrade (for free) from SP1 to SP2. It’s easy. Just run Windows Update.

Or, better yet, upgrade to Windows 7. (Microsoft’s suggestion, not mine. See?)

As self-serving as Microsoft’s suggestion of upgrading from Windows Vista to Windows 7 may be, I find myself in the rare position of actually agreeing with them on that. Yes, it costs money, but let’s face it, Windows Vista is a joke. It’s Windows ME2. Do you really want to be stuck on it?

Well, your choice. Either way, Microsoft’s point is that Windows Vista SP1 is officially dead to them.

No surprises there, as Microsoft probably wishes everyone would just forget that Windows Vista ever existed. Heck, I wouldn’t be surprised at all if support of Windows Vista ended with Service Pack 2 and Windows XP actually outlived it!

Speaking of, if you’re one of those people holding on to dear life to Windows XP (and I even fall into that category to some extent) the Windows XP support will end… One day… Maybe… If it isn’t extended again. I guess over in MS World, now that Windows 7 has finally given people something stable and usable to upgrade to (since Windows Vista failed miserably on that front) we’re all supposed to upgrade to Windows 7. (Except for those of us who can’t, of course.)

Here’s the really odd thing though. Microsoft claims that Windows XP will indeed die. It has less than 1000 days less of its extended support. Officially, on the 8th of April, 2014, Microsoft will no longer provide security patches, hotfixes, etc. for any version of Windows XP. I get that. It’s something of a bummer for those of us using it on low-powered low-memory laptops and netbooks where Windows 7 is a less-than-convincing “upgrade”, leaving us … well, SoL. Not to mention those of us just keeping old PCs alive! It’d be one thing if Windows 7 could actually run as well as Windows XP on an old or budget system, but it’s another thing entirely when it can’t, and in even more cases where Windows 7 can’t be installed on the box at all!

But things take an even stranger turn, because along come the rumors and innuendo of Windows 8. Microsoft exec Tami Reller just told folks at Microsoft’s Worldwide Partner Conference 2011 that if a PC could run Windows 7, it could also run Windows 8. It’s an assurance to PC makers (and consumers) that anyone who is running Windows 7 now will be able to upgrade easily to Windows 8. You won’t have to worry about memory or processor speed at all. (Which could be a Microsoft first since it left the concept of Windows as a DOS shell behind!)

However, that raises an odd question in my mind: What about the army of folks still on Windows XP? Can they upgrade to Windows 8?

Why do I ask that?

Why do I think that they even could if they can’t upgrade to Windows 7?

Simple!

Windows 8, supposedly, will not just be available for your PC, but also for your ARM-based tablet! And let’s face it, Android tablets are not running Windows 7 for a reason. (Well, a couple of them.) It’s not just a matter of ARM vs. x86, but also of resources. Linux and Linux-based OSes such as Meego, sure. Windows 7? Nuh-uh!

And Intel is still pushing hard for Atom to replace ARM in Android tablet manufacturer’s minds. Will it? I doubt it. It’s like swatting a fly with a … hardcover book. But a Windows 7 tablet, that’d be spiffy … if it could run faster than molasses in January. Which it can’t. (I would know, as I have an Atom-based tablet. I technically can install, but if I thought it was already slow running Windows XP…)

But if Windows 8 could run well on low resources like on a tablet?

Because let’s face it, Windows is a resource hog compared to Android or iOS as an operating system goes. We’re even seeing some grumbling from WebOS tablet users that the OS itself is sluggish. The performance of the OS on a dinky little tablet processor makes a world of difference. One that Windows 7 just can’t even try to compete. So if Microsoft wants people to actually bother putting Windows 8 on tablets, it’s going to have to trim up that kernel. Will Windows 8 be split into two worlds: PC running kernel-heavy like Windows always is and tablet running kernel-lite like Windows CE or Windows PHONE 7? Or will Windows 8 actually be lean by design and work for anyone and everyone equally well with the same kernel for all?

If the latter then it would be a Microsoft first! Which makes me very doubtful.

But it would be an extremely welcome change, especially as a lightweight kernel from Microsoft that still ran PC software might even just give all of those aging computers running Windows XP an actual upgrade path! If Microsoft were interested in listening to what customers want (instead of just telling us what we need) Windows 8 could possibly be the savior of Windows XP users who would like to upgrade, if only Microsoft would make an OS that they could actually upgrade to.

I wouldn’t hold my breath on it though. Because that just isn’t the Microsoft way of doing things. Even if it would make sense. And sales. Lots and lots of sales.