Slashdot Mirror
If Windows Came to PPC, Would You Switch?
An anonymous reader asks: "This question was posted on Ask Slashdot about a week ago: 'If Mac OS X Came to x86, Would You Switch?' This makes me ask why not have Windows run on PowerPC? Windows/PPC would not necessarily have to run on Apple hardware, or at least not exclusively on it. I'm sure their friends at IBM and Motorola would be happy to provide chips to anyone that wanted to make computers to run this new OS. Microsoft could dust off the code from NT4/PPC, add some code from Virtual PC to get Windows/x86 compatibility, and have it up and running in about the same amount of time it would take Apple to get Mac OS X running on common Intel hardware." An additional question comes to mind, however: If Microsoft made this move, how would Intel react?
Mac OS X is 90% of the reason I have PPC.
If Mac OS X was on x86 I'd have a x86.
Ok, so in 1996, NT4 came out on x86, which was the first step that Microsoft really took into making Windows a real OS.
It ran on PPC, Intel, Alpha and MIPS. That's a lot of architectures. Now, think about it: One of the things about Microsoft is, generally speaking, they have no soul. If they make money selling a product, they'll sell it. Now, that's not to say they won't STOP selling any product that's not making money (*cough*XBOX*cough*) just to drag their competition to the ground, but they also won't turn down cash for ideological reasons.
The fact that when Windows 2000 came out reflects that no one really used NT 4 on anything other than Intel hardware. Now, this might be because the hardware developers never really were 100% behind MS, or it might be because someone that was shelling out cash for an Alpha or a MIPS workstation (but I do remember there being a drop-in MIPS chip that would work in a socket.... 5? Pentium board?) wanted a better OS, or any other reason.
The fact is, you can say that PPC might be a faster processor platform today, with a higher bus speed and better performance per clock, but its close. Very close. I don't think MS would be able to polish a PPC version of Windows as much as they have the Intel version, meaning you might take a relative performance penalty... and there isn't a price advantage in PPC over x86.
So yeah, the previous failure, combined with the pitfalls of a new version listed above make a pretty strong case for "no."
NT4 ran on PPC, up until SP3 (the last install discs with PPC support were SP3 based).
Nobody switched, and that was in the days of the gratuitously unstable System 7.5 and Mac OS 7.6, which tended to crash if you looked at them wrong.
I suspect that BeOS has more users than NT for PPC, at least for Macs. And neither OS ran on G3's or later CPU's.
Now, with OS X and VPC, why the hell would I want to run Windows of all things on a Mac? other way 'round I can see, especially with WINE support or something similar (like Mac-on-Linux) to get Windows software compatibility. But even then, I'd probably stick to PPC, as the hardware is generally better quality and definitely better designed.
"You've got an invalid haircut" -Warren Zevon - Life'll Kill Ya
If Microsoft made this move, how would Intel react?
They would panic, of course ! The whole x86 architecture is ugly as hell, and the IBM PC architecture even more so, so low level programmers would propably open a bottle or two and party if we could ditch our x86's for PPC's :-)
The 8086/8088 (to which even the Pentium 4 tries to be backwards compatible with to some degree) was a hack at Intel to get a 16 bit processor to market fast and was meant to have a very short lifespan. Intel was developing a way better processor then (can't remember its number, could anyone fill it in ?). So they took the Z80 processor and extended it. You see the relation even today in the register namings.
I wasn't aware how much the x86 really sucks until I began programming the Motorolla M68000 in the Sega MegaDrive/Genesis as a hobby a few weeks ago. That processor is about as old as the 8086/8088 but has so many cool and useful features that the x86's doesn't have even today (like the eight address registers and the postincrement/predecrement features which make it trivial to set up eight stacks at once, just to name two features).
And then IBM came along. They wanted to get a "cheap" computer to market fast, and used Intels 8086/8088. And like the processor, the whole IBM PC was meant to have a short lifespan.
Unfortunately the PC became a success, and so its lifespan had to be expanded artificially and backward compability had to be put in. This is true for the Intel processors as well as the whole PC architecture. As time passed by more and more things were added without really fixing the underlying problems.
I think computers could be cheaper and more powerful if we'd had a better mainstream processor and computer architecture, one that was meant to live long and thus was better designed. But this is just a dream, I'm afraid...
The app situation on NT/Alpha is often misrepresented like this -- on the SERVER, there was hardly anything you couldn't get. SQL Server, Exchange, Oracle, Domino, all ran on Alpha.
The big problem with Alpha is that price/performance wasn't *that* overwhelming after the Pentium Pro shipped. Also, there was the inherent risks in running a "Tier 2" platform, even when some uses (like Exchange) really needed the CPU power.
(We had DEC out to demo NT/Alpha for us, and on two seperate occassions their show-n-tell systems failed to boot. So, there probably was a big vendor factor there too.)
Business. Numbers. Money. People. Computer World.
Well, NT on Alpha didn't fail miserably. There were a lot of these boxes out there, and I'm willing to bet there still are. Microsoft developed for Alpha longer then the other ports; they had Exchange for Alpha and some of their other server software packages.
Alpha was quite a bit quicker then x86 in it's day; it was a full 64-bit system from the start and the processors were clocked pretty aggressively. NT's x86 compatibility layer for the Alpha actually worked pretty damned good too- it ran 95% of the software on x86 and once you ran the apps enough, they ran pretty quickly. Alphas also weren't outrageously priced.
They just didn't keep up with the x86 boxes in the end, Digital was on the way out, and the Alpha just faded away.
- It's not the Macs I hate. It's Digg users. -
I was thinking about this for a while, and had an interesting thought. So, MacOS (9 and X) runs on PowerMac hardware only, right? But, you can use Mac-on-Linux on Linux/PPC, any Linux/PPC. Not just Linux/PowerMac. That means that you can "run" MacOS on an RS/6000, the same computer that NT/PPC ran on. It's like VMWare. It's not emulation, it's virtualization.
Here's my question: Would it be possible to run NT/PPC on PowerMac hardware through a MOL like virtualization layer? I don't know how useful this would be, but it might be fun. (Actually, I can think of a couple of uses for it, like recompiling existing Win32/x86 apps to Win32/PPC and running them on the Mac, albeit in an NT virtual machine. It would, however, run at native CPU speed.)
-twb
Okay, get off the OS idealogy wagon for a second, and entirely off which hardware is better. Now start thinking about barrier to entry and business models.
Name the OSes that run on x86. Now name the OSes that run on PPC.
Any low level geek can name three, and lots of computer users these days can name three as well, and even more can name two, even if they have contempt for it, be it for reasons they don't understand.
1) Windows
2) Mac OS
3) Linux
Now linux is intimidating for the average user. Most people won't bother to install it. It runs on both, but the cost to entry is too high for the average user. It costs no money, but way too much time.
Now look at the remaining two. One only runs x86, one only runs PPC. For 90% of the populace, the only choice is windows on x86. Most people don't think they have a choice. I'm dealing with more and more people that have problems with computers and bring them to me to fix. I have a way of making windows a little more secure, but that's only because I know and use features and free software which most people don't even know exist. Most require a complete wipe and reinstall.
Now think about a hardware switch to PPC. Intel dies but Dell and the others adapt over 5-10 years. Windows chugs along.
Then there are people like me continuing to reinstall windows in that time.
"Hey, yanno this is the third time you sent this to me. Maybe you should think about another OS. I got a copy of Mac OS X here if you'd like to try it. In my professional opinion its more secure and will save you money and time." No need to buy any new hardware"
And maybe this action won't kill microsoft over night, but it will erode markets share, and microsoft cannot abide eroding market share of any amount.
"All great wisdom is contained in .signature files"
MacOS X's core is already available as an x86 version. All they'd need to really do, since a very sizeable portion of the Aqua interface system is written in Objective C, would be to account for endianness and call it done. It'd take all of a 6-12 month project, I'd suspect, to put it into an alpha class release stage.
Microsoft, on the other hand, would probably have a nightmare on their hands as I suspect they've not taken any consideration for endianness, 64 bits (No, they still don't have it out in the hands of the public- it's been months now and they knew about amd64, etc. for some time now...)- it's probably all nasty, crufty x86-32 code and using some aborted NT 3.51 code wouldn't help out much...
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
Actually, NT was first written for the (nonexistant) i910 processor
MIPS was the first port, largely because of the lack of delivery of the i910 processor.
-- The universe began. Life started on a billion worlds...
-- Except on one where stupidity was there first.
I remember back in '96-'97 timeframe hearing from a number of Sun vendors about experiments with NT on Ultrasparc, but could never get a demo (and we had Sun workstation vendors falling all over us to give us hardware at the time).
:-)
That's because Sun played Microsoft for a bunch of fools. Microsoft realized that Sparc was *the* platform to support at the time. As a result, they were falling over themselves when Sun offered to sign an exclusive contract to develop NT for the UltraSparc.
After the papers were signed, McNealy laughed as he happily sat on the port of NT and used his newfound legal authority to prevent Microsoft from bringing it. That's why there was a port for MIPS, Alpha, and PPC, but no port for Sparc.
Javascript + Nintendo DSi = DSiCade
"Not quite true. Remember Mac clones? Based on a standard reference design (CHRP? PREP? one of those) they could in theory run either NT4 or MacOS. I don't know if Apple hardware ever ran it, Apple could get away with having not-quite-conformant hardware that would still run MacOS."
This I can vouch for. I worked for a company for a short time that had a dual processor Motorola StarMax running NT4.0 as some kind of server that was semi-public, and they were using the PPC architecture because it severely reduced the utility of the box to anyone who would break into it.
I would imagine that the only Mac that could run it natively would be the PowerPC 4400, which was based on the same architecture as all of the cloned Macintoshes.
Do not look into laser with remaining eye.
Every MSCE or MCT or MCP I have talked to have told me that NT stands for "New Technology"..
Although, looking at this page, I may be slightly mistaken..
in girum imus nocte et consumimur igni
The moniker "New Technology" was retrofitted onto Windows NT long after the product had been shipped. Originally the code was developped on the Intel i860 CPU, which was called N10 (N-Ten). The "NT" derived from this CPU. Windows 2000 says it is being "build on NT technology", which is utter nonsense, if you look at it as "New Technology technology".
The same goes with Windows CE, with CE standing for nothing. Some Microsoft guys just thought it sounded cool.
actually there was a PPC port of NT years ago. It was dropped beacause...
NT was developed on the Intel i960, a RISC processor. Intel never went anywhere with it, tho' the i860 is still used (for example, for RIP in printers). One of the design goals was to be platform independent, hence the HAL. NT shipped on x86, Alpha, PPC and MIPS. There was also a SPARC port that never made it into commercial distribution.
The problem was that MIPS and PPC, at the time, were in the middle as far as performance went. People who wanted to run NT for ordinary desktop workstations bought x86, because it was cheap. People who wanted to run NT for CPU-intensive apps (CAD, FEA, CFD, etc) bought Alphas. There was simply no demand for people who needed a little less power than Alpha at a price higher than x86, so Microsoft stopped selling those editions.
Let me make this very clear: the market decided that it did not want a multiplatform OS.
There's no technical reason that MS couldn't release a version of NT on PPC. You might say that there's a case to do that now that Alpha is history. In fact, I wouldn't be surprised at all if MS continues to do builds of NT on PPC just to maintain the ability to do so (a common practice in large scale projects is to build on another platform that you don't ship on, just to keep the codebase clean). But, the fact is, the price/performance of PPC versus x86 simply means that there'd be no advantage to running NT on PPC, and all the disadvantage of less ISV support.
So in conclusion, people would switch if a) PPC had as big a performance gap over present day x86 as Alpha did over x86 back in the day and b) there was some ISV support for it.
There was indeed a version of NT 4 for the Alpha, but it didn't work very well.
A few years ago, we had a spare DEC Alpha, and decided to run a data warehouse on it. We put Windows NT and Microsoft SQL Server 6.5, an assload of RAM, and two full shelves of fancy 10,000-RPM disks, with a catastrophically-expensive RAID controller.
You'd think that performance would be pretty snappy.
Maybe it was just the talents of the administrators (SQL 6.5 had a lot of stuff that you could tweak), but we could never get the performance of this seven-foot-tall behemoth to particularly exceed that a standalone i386 server.
But it did have this cool picture of cowboys that came up when you booted it. Which we did a LOT.
Never attribute to malice that which can be explained by mere idiocy.