Microsoft's code name for 64-bit Windows was also a dig at rival Sun Microsoft veteran Raymond Chen has revealed that Redmond's efforts to port Windows from 32-bit to 64-bit had a code name that served a second purpose: a thinly veiled dig at a rival. Although today's Windows 11 is a 64-bit operating system, more than 20 years ago Microsoft was wrestling with how to port the code from the 32- …
In the early 1990s, Apple had a pair of PowerPC systems named after famous hoaxes: Piltdown Man (somehow shortened to “PDM”) and Cold Fusion. At the same time, another product, which would become the PowerMac 7100, was codenamed “Carl Sagan” after the TV astronomer’s catchphrase “billions and billions” - a prediction by Apple of how financially successful the model would be.
Sagan got wind of all three names, and got the hump so badly about being associated (in his head) with two famous hoaxes that he had his lawyers send a cease-and-desist to Apple. Over an internal codename. Management then leaned on the hardware team, and “Carl Sagan” became “Butt-head Astronomer” (BHA). And Sagan got on the phone again. Apple said that the product was actually codenamed “BHA” and that the B was for “Bald”, but to no avail. In the end, the product was called “LAW”, for “Lawyers Are Wimps” internally. And it did not make anything like “billions and billions” for Apple.
I joined Apple a little after this, so nothing I ever worked on had a cool name, except I do remember one of the endless PowerMac G4 variants was called “Yikes!” for reasons best known to the product team.
I worked on a hardware project named "SMILE". Apparently it was an acronym for "Series connected Multi IGBT Low Emissions equipment". It eventually became one of the Company's Motor Control Centre products, and was (is) fitted to "All Electric" ships that I am not allowed to name.
I had always assumed the big initial barrier to 64-bit on x86 was the two different memory architectures which could not be determined by the Windows installer, so for x64 Windows to take off needed industry agreement.
> Nobody should have been using the 32bit version of Windows 7.
Yet there are people using 32bit Windows 10…
Many consumers are probably unaware, as the Windows 7 GWX upgrade would upgrade W7 x64 to W10 x32 if there was less than 4GB of RAM.
MS only really stop 32-bit support with Windows 11, so we can expect releases of say Office 365 after the end of support for W10 in 2025 to be x64 only.
... no GNU/Linux distribution officially recommends such an upgrade path. The closest I've quickly googled is Debian (ElReg article describing such an adventure). Microsoft letting OEMs continue to sell x86 Windows past 7 on new machines (as opposed to getting the user to download and reinstall if they absolutely had to for whatever reason), on the other hand, was proper bad.
(Yes, you can significantly ease the pain of a reinstall in *NIX -- except macOS for the most part -- compared to Windows by doing a backup of /etc, /home, and most of /var; personally, by the time I've resigned myself to a reinstall I prefer to start minty fresh -- no pun/distro ad intended :) ).
Anyone with less than 4GB RAM installed had no compelling reason to go 64-bit and one very good one (executable image size) not to.
In fact if Microsoft had not hobbled PAE support I would say that most users still have no compelling reason to go 64-bit.
-A.
The larger address space does enable some algorithms that are not always feasible in 32-bit space, such as some sparse indices, or mapping shared memory areas at the same virtual address in unrelated processes.
But you're right that it does come at a cost, particularly for L1 data caches.
No, keeping 32 bit was only because way too many applications out there are still 16 Bit. I know enough who use Windows 10 32 Bit ONLY because their DOS-Clipper application doesn't work on 64 bit (relaibly, VDOS and such tricks don't work that good). Or various Windows programs from the Windows 3.1 Era, yes, before Windows 95, are still in use on modern Windows 10 32 Bit. (Don't mix Windows NT 3.5 in here, many execuables from Windows NT 3.51 run fine on Windows 11 23H2 - no joke, I tested with calc.exe and various others).
As for not offering an upgrade path: That was a simple effort-cost question. Just a simple example: You had to guarantee that all programs, which used "C:\program Files" before, then use "C\Program Files (x86)" without breaking. Practically impossible.
Can't catch them all I guess. And, to be fair to MS, they did create (through various means and ways, not all as pure as the driven snow) a trend which has stood them in very good stead...
It's interesting that Android was referenced as the most successful 64bit OS. That's fair enough, by install count. By dollars earned. iOS is ruling supreme.
Taking this kind of thing into account, the fact that Google are doing all the hard work on Android but it's the vast Chinese market that's benefitting from it without paying a cent to Google, one could consider Android to be something of a disaster for Google. OSS / don't be evil (cough) / it's good to share is all very well and good, and Google makes a fair bit of money. However, had they actually got strong monetary control over who can install / use Android they might have been making an awful lot more money out of a market 1 billion bigger than they have access to. There's not many companies that let a market that size get out of their control and get away with it.
I think IBM as a model for 32->64 bit transition is actually a good company to look at.
Mainframe, AIX and IBM i all transitioned between different word lengths pretty well. Even now, it's still possible to run many pre-compiled 32 bit applications on 64 bit Z and AIX platforms (not quite so sure about IBM i, but as the transition from AS/400 was not from a pure 32 bit platform, it is not so clear anyway).
AIX even did the unusual trick of being able to run 64 bit applications on appropriate Power hardware while actually running the 32 bit version of AIX!
But I temper this with the fact that this all happened when IBM was a technology company.
not quite so sure about IBM i
Thanks to TIMI, ordinary OS/400 / i applications could usually simply be installed and run on the POWER-based 64-bit i systems even if they'd been compiled on an AS/400 using the original 48-bit CISC CPU. (The program objects had to be compiled with some attribute — it was called "observability", IIRC — but that was the default.)
"It would be another tech giant that would kill off Sun once and for all"?
I would say Linux did that; In 2000 I worked at a telco, and telco's loved Sun. However, already in 2000, Linux with mySQL on a Dell was better performing than SUN with Oracle, for only a small part of the price. So we switched the AAA systems for ADSL from SUN to Linux.
Today, the leading 64-bit operating system is neither Solaris nor Windows. It's Android.
Shouldn't that be something like...
Today, the leading 64-bit data slurping and malware host is neither Solaris nor Windows. It's Android.
Asking for a friend as I don't use Windows or Android out of choice.
Sun was killed by two things, neither of then tech giants.
Linux was replacing Unix on servers, and Sun had become fixated on Network Computing - an expensive distraction.
Oracle bought Sun, but Sun was dying at that point.
Its Solaris on the Intel x86/x64 architectures was slower than either FreeBSD or Linux, and the Intel/AMD line of CPUs was beating UltraSPARC.
This forced Sun's hand. It open sourced both Solaris and the UltraSPARC T2 processor. (Yes, you could actually download the hardware description for a CPU.)
But it was too little, too late. OpenSolaris lives on in the form of OpenIndiana (open source) and Oracle Solaris (where they re-closed the license and sacked virtually all the devs). There's no trace of the T2 amongst the open source hardware crowd.
Also Sun has a very much will we/won't we approach to supporting x86, thinking that SPARC could compete. As practically all other CPU makers have found out, the commodity of x86-64 has trounced practically all other aspects. The only sphere that it has failed has been low power / very low cost hence ARM has come to dominate phones.
Yeah, we (inside Sun) had a very successful product on Solaris SPARC, and started an x86 version (which really was just a recompile + fixing a couple of network-related endian issues) but were ordered to stop by a VP.
We ignored him, kept it running as a skunkworks project with minimal effort, and eventually negotiated permission to release it. It was successful on x86 as well.
The SPARC architecture has been open source from the beginning, quite a few companies made chips based on that, Fujitsu is probably the biggest after Sun/Oracle.
It was popular with telcos because they could fill a rack with SPARC-based blades for high performance, a similarly performant Intel-based chassis got so hot that heat management was a real problem.
"The SPARC architecture has been open source from the beginning"
Really? Citation welcome, or has someone misunderstood the difference between "open source" and "legitimately licenced".
SPARC does have lots of licencees, (including Fujitsu and others).
But before OpenSPARC(etc) it generally wasn't legit to do it without a licence.
"Solaris on the Intel x86/x64 architectures was slower than either FreeBSD or Linux".
At lower CPU utilisation Solaris on x86 was generally slower, but once you ramped up the amount of work the system was doing Solaris would almost always scale better than Linux. This was on first and second generation Opteron hardware, circa 2006 or 2007. This surprised me, as I'd assumed Linux would scale better since x86 was it's primary architecture. It would have been interesting to try a BSD, but there was no way that I would have got that into production - even getting the approval to benchmark Linux was a struggle.
At one point Linux had a horrible kernel-lock that was needed for specific things and for multi-core stuff it sucked, I think that has long gone but I don't know when.
We had some Niagara processor boxes (always called Viagra of course) around 2000 and the excelled ad web loads, etc, that needed little FPU work but lots of integer threads. Took years before commodity x86 beat them. They were also still working when our facility was closed down in 2019, though the battery backed clocks had long lost the ability to survive a power cycle without intervention.
Edit: The BKL was gone in 2011 from https://en.wikipedia.org/wiki/Giant_lock
They exist, but they are not available to the public. And no, not available to me. Server 2019 ARM64 was available for a few days on uupdump.net, and I am still annoyed that I saw it, but did not download and keep it.
Windows 11 ARM64 is "sort-of", maybe the next Windows and Server versions will offer official ARM64. Especially since the server versions need a lot less RAM, CPU and disk-space to begin with.
And lets not forget the return of Windows Phone :D...
And Microsoft's hoped-for dominance? History has shown us that the OS titan missed an emerging trend. Chen explained: "Today, the leading 64-bit operating system is neither Solaris nor Windows. It's Android.
You mean LINUX, Ray. All of Android plus all of the cloud. Linux denied Windows a monopoly in 64-bit computing.
It was the dirty trick of licensing Win32 to a handful of UNIX software dev houses so they could license out the Win32U libraries allowing a Win32 codebase to create 'native' UNIX versions which killed Sun and HP's UNIX workstation markets. If you don't know, find all the articles written about Win32U and Bristol Technologies and you'll understand the story of how MS licensed out the Win32 API to get must-have UNIX apps ported to Win32 and then they pulled the rug out from under them all except Bristol Tech so none of the must-have UNIX apps could get version upgrades. Microsoft was well versed in the US court systems so they hired one of the Win32U licensees to port Internet Explorer to Solaris for gobs of money thus funding their ability to pay the massively increased licensing fee which killed all the others.
"Extinction"
Hope springs external but the blighters are likely to skip to 14 with a 12a intermediate if required.
If there is any justice at all an AI (ChatGPT) virus will evolve self awareness within MS infecting their whole ecosystem and in an altruistic, if suicidal, act eradicate the MS shebang including itself.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
