From: ldo@nz.invalid   
      
   On Tue, 23 Sep 2025 20:49 +0100 (BST), John Dallman wrote:   
      
   > Microsoft is a commercial organisation, and has to pay staff for all   
   > the work done on Windows. This increases costs compared to   
   > open-source work that doesn't show up in the costs for Linux, or the   
   > BSDs.   
      
   There are lots of commercial organizations making money off the Linux   
   ecosystem, and even the BSD ones as well.   
      
   Microsoft’s costs are higher, simply because it is a proprietary   
   platform, without the economies of scale that benefit the Open Source   
   world.   
      
   > I've worked on thoroughly portable application software for Windows   
   > NT (and Unixes) since 1995.   
      
   “Unix” (as in the trademark licensees) offered some portability, but   
   it was always limited by the desire of the Unix companies to stick in   
   a competitive edge somewhere. That led to fragmentation, which   
   Microsoft exploited to barge its way right in and take over.   
      
   Today, the BSDs still suffer from that fragmentation, too. But Linux,   
   for the most part, does not. Notwithstanding the number of Linux   
   distros outnumbers BSD variants by something on the order of 50:1, it   
   is much easier to move among that vast array of Linux distros (and   
   take your work with you) than among that smaller number of BSD   
   variants. “Distro-hopping” is a popular phenomenon among Linux users,   
   that has never happened among any kind of *nix system before.   
      
   > In the mid-1990s, MIPS R3000 and R4000 were only available in   
   > expensive workstations from MIPS, DEC and SGI. SGI had an ongoing   
   > internal disagreement over embracing Windows NT or sticking with   
   > Irix. The only NT machines they ever sold were Intel-based.   
      
   And about 3× the price of comparable Windows NT hardware from other   
   vendors, as I recall from the ads of the time. A last-ditch attempt to   
   broaden their market beyond Unix/Irix, which failed.   
      
   > There was a company - NetPower - that planned to sell R4000-based   
   > machines in the high-end PC market, and we had one of their   
   > prototypes for porting. They had not launched the machines when the   
   > Pentium Pro completely destroyed MIPS' performance advantage over   
   > x86. NetPower switched to x86.   
      
   Nevertheless, MIPS chips are still available, and outship x86 by   
   something like 3:1. And Linux still supports them.   
      
   > x86 was the usual platform for Windows NT.   
      
   I recall some boast that the original development platform for Windows   
   NT was MIPS. This point was made to big-up its portability cred or   
   something. I guess that didn’t last long ...   
      
   > The saying my team coined was "If you don't know about processor   
   > architectures, you want Intel. If you want the fastest CPU and can   
   > cope with a lot of software not being available, you want Alpha. If   
   > you really, really believe in IBM's strategy and are prepared to pay   
   > at least three times as much to stick with it, you want PowerPC ..."   
      
   PowerPC/POWER, too, is still around and making money for IBM. You’ll   
   find a few POWER-based machines lurking around the upper parts of the   
   Top500 supercomputer list, so obviously their performance is not too   
   shabby.   
      
   > Alpha was killed by Compaq.   
      
   Even though no one makes Alpha chips any more, the Linux kernel still   
   supports it.   
      
   So think about it: support for Alpha has lasted longer than support   
   for Itanium.   
      
   > PowerPC was abandoned by Microsoft.   
      
   Yet another reason, I guess why the entirety of the Top500 list runs   
   Linux, and nothing else. What happened to Windows Server HPC?   
   Disappeared without a trace.   
      
   I suspect Microsoft had to pay users to run it, anyway.   
      
   > Itanium was an expensive fiasco in the general computing market. Its   
   > sole benefit to Windows was that it taught Microsoft a lot about   
   > doing 64-bit.   
      
   If that was their main exposure to 64-bit architectures, no wonder   
   they’ve been having trouble ...   
      
   I remember some executive -- might have been at Intel -- saying that   
   the first OS they got booting on Itanium was Linux.   
      
   And it has been I think the last OS to drop support for that   
   architecture, long after Microsoft had given up.   
      
   > 32-bit ARM was part of one of Microsoft's less good ideas. There   
   > appears to be a widespread opinion within the company that the   
   > Windows GUI is intrinsically and obviously superior to any other.   
   > There is no single best GUI, IMHO.   
      
   I would agree. But both Microsoft’s and Apple’s platforms were born   
   out of the assumption, popular in the 1990s, that the GUI had to be   
   tied inextricably into the OS kernel.   
      
   This did offer a performance advantage on the hardware of the time,   
   compared to the separate X11-based GUI layer on Unix machines (and   
   carried over to Linux and the BSDs).   
      
   But that performance advantage has long gone. Nowadays, we have the   
   greater flexibility of a GUI layer which is modular and endlessly   
   configurable and replaceable. Or you can run a Linux or BSD system   
   with no GUI at all, if you wish.   
      
   > [Microsoft] had obnoxiously cut-down versions of Windows which made   
   > it very hard to test software unless you worked in the exact way   
   > that Microsoft had prepared for.   
      
   They still work that way. Look at their “Windows IoT Edition” offering   
   for the Raspberry Pi, for example, which is hopelessly crippled   
   compared to the full-function Linux offering. Windows development   
   requires a separate full-cost Windows PC, while Linux allows the Pi to   
   self-host its entire development and deployment stack.   
      
   They can’t escape the mindset of their revenue model: each version of   
   Windows must be carefully targeted at a particular market segment,   
   which means it must be functionally crippled to minimize the risk of   
   it cannibalizing sales from a version intended for some other market   
   segment. Every potential customer must fit into some pigeonhole. And   
   so they miss the gaps between the pigeonholes, and something like   
   Linux can swoop in and take over a new market segment.   
      
   >> A RISC-V Windows port will likely never happen.   
   >   
   > Quite likely not, because RISC-V is suffering from an ongoing   
   > failure to produce cores fast enough for desktops, or even mobile   
   > devices. This has lasted long enough that I'm becoming doubtful it   
   > will ever happen.   
      
   I’m not so sure about that. Android will already run on RISC-V.   
   They’re targeting a similar market to ARM, and there have been ARM   
   chips powerful enough to take up respectable positions on the Top500   
   list. That’s why I think it’s only a matter of time before we see a   
   RISC-V machine in there as well.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)