Interesting provocative article, I bet it will be praised on some Microsoft sponsored conference.
Wine and Proton are not tributes to Win32's portability. They are symptoms of a desktop market that Microsoft locked hard enough that the rest of us had to reverse engineer our way out. Market damage, not collaboration.
The ecosystem was not won on technical merit. OEM per-processor licensing, embrace-extend-extinguish against Java and the web, document format lock-in, and a long pattern of obstructing standardization attempts that would constrain Windows (PWI in 1994, ECMA-234 in 1995, OpenDocument later) while pushing their own through when it extended reach.
No CS curriculum holds up Win32 as exemplary API design. No system copied it. A successful API earns adoption. Win32 enforced it.
Early Java was horrid for everybody except the architecture astronauts who could cram ten GoF design patterns into a hello world program. It only got traction because a different wannabe monopolist, Sun Microsystems, spent heavily to get it pushed into CS curriculums. Fortunately, the one-two punch of Linux and Intel killed Sun or we might all be cursing them today instead of Microsoft.
Sun was actually a decent company once upon a time. Their problem was that Microsoft excluded them from the market and starved them for money for so long that their hardware stopped being competitive, so that by the time Java made it so you could run some software on it, nobody wanted their hardware regardless.
It was only after they went bankrupt and got bought by Oracle that things like OpenSolaris getting killed off and Java lawsuits started happening.
> Wine and Proton are not tributes to Win32's portability.
I emphatically disagree. It is a hilarious and catastrophic failure of Linux userspace that the best API for running games on Linux is Win32. This has absolutely nothing to do with what Microsoft locking down the desktop market. It has 100% literally everything to do with Linux userspace being a clusterfuck of terrible design.
Linux adopted Win32 because it actually worked. They didn't have to. They could have simply invented a better API that didn't suck. But that's quite hard.
So yeah. I emphatically, but respectfully, disagree with your entire thesis.
All true, but simultaneously, if you look at it not in the “how did we get here” but “what’s out there today” light, it’s an option that can run on a lot of platforms. Not by some particular merit but because history happened this way - but that’s not a problem with the technology itself.
The article defines "success" in the Windows context as being "available everywhere". It does not address how it got to that point.
And sure, you might not like Microsoft, and you may not like how it became successful (using the above definition) but the fact that it is available everywhere is not in dispute.
Of course most successful things have murky pasts. We don't necessarily agree with how it got there, but there it is. That is, at least in the technical sense irrelevant. You may prefer LP's or CD's, but streaming is now the successful way to get your music.
That doesn't mean it's the only way though, and of course you are free to not use Windows programs, or play games via Steam etc. That is your choice.
If you listen to CS professors you'd believe that Haskell would be very successful despite having limited if none real world usage. The reality is that catering to your users is much more important than having idealistic system.
In my experience even with games that have native Linux support running through proton seems to have less issues.
There's something deeply satisfying about compiling a Win32 desktop application and knowing that single binary will run unmodified on essentially every Windows machine from XP onward, and look nearly identical doing it. High-DPI is the one real caveat, but even that's manageable with a manifest. I'm not sure any other platform-native API comes close to that combination of reach and consistency. Running them on other operating systems is really just a bonus.
They exist but they’re in the minority. Compare this with the Linux world, where userspace compatibility between one major distro release and the next, 12 months apart is very much a roll of the dice.
As an added bonus, since win32 is so old. AI is quite strong at generating code for it, I have even had success with qwen 3.5 1.5B parameters with win32 and opengl 1.1
Sometimes I think that, if Java ditched the idea of JVM and followed what modern languages did today (write once, compile on every major platform), things would be different.
The trouble there is that Sun was right about that.
Doing it that way works great for open source where anyone can recompile the software for a new target, but for proprietary software they would have given you a Windows blob and that's about it.
Meanwhile the problems with Java were mostly not the JVM. Its current problem is, of course, Oracle.
Hasn't it taken gargantuan multi decade efforts of hundreds of developers, multiple open source projects and the backing of major corporations to make it work at all, let alone well, on Linux?
Hasn't it taken gargantuan multi decade efforts of hundreds of developers, multiple open source projects and the backing of major corporations to make [WINE] work at all, let alone well, on Linux?
Arguably there has been equal or larger effort invested/wasted in cross-platform and cross-distro frameworks/APIs/packaging and yet the result still doesn't work. Partly that's due to duplication of effort; there's (mostly) one WINE competing with Qt/Gtk/whatever times Snap/Flatpak/AppImage/whatever.
Because hundreds of developers, multiple open source projects and the backing of major corporations made it happen, not because Microsoft wanted it but in spite of it.
In this case the route to success was via marketing (isn't it always so?), via market share, via application dominance (attracting developers to develop for the platform), and via insane levels of backward's compatibility. It was successful not because of the code itself (end users don't care 2 figs about the elegance of the code) but because they optimized for the end user experience.
Linux optimized for the experienced, technically adept user, who wanted to fiddle, customize and could write programs. Apple optimized for the "now", ignoring the past and regularly made existing programs obsolete and unrunnable.
I wrote windows programs in 1995. They still run today. They have run on all versions of Windows since then, without even a recompile. Everything I have [1] just keeps running. And it turns out, that's something users really want.
I get that we're all technical folk here. I get that we strive for technical excellence and elegance. I get that we operate in the "now", ignoring hardware and software from the past. But the market is different, and wants different things. If you want a successful business you need to understand the market, not just your own aesthetics. Microsoft understands that, and that's why the market (especially the business market) relies on them.
[1] - Except games. Copy protection on some of my games means they don't run anymore - but to be fair those were hacks designed specifically to prevent the game running in the first place.
Considering you can use 500 kilobyte distro of tcc that can self compile to build win32 apps and those versions will very easily run on wine. I think it is a valid point and a strategy I have used for sometime. Since windows can't decide what you should use for developing Windows apps
Technically the Internet runs on "the Internet Protocol Suite" but people just say "TCP/IP" for short. That doesn't mean to exclude UDP, SCTP, or whatever.
He means the Internet runs on the Internet Protocol Suite and not the "superior" OSI Protocol Suite (that people under age 50 have never heard of because it failed).
Yes, which the article claims is just a theory now, irrelevant for the real world. More crucially though, http3 doesn't use TCP because it is built on top of UDP.
(2024)
Interesting provocative article, I bet it will be praised on some Microsoft sponsored conference.
Wine and Proton are not tributes to Win32's portability. They are symptoms of a desktop market that Microsoft locked hard enough that the rest of us had to reverse engineer our way out. Market damage, not collaboration.
The ecosystem was not won on technical merit. OEM per-processor licensing, embrace-extend-extinguish against Java and the web, document format lock-in, and a long pattern of obstructing standardization attempts that would constrain Windows (PWI in 1994, ECMA-234 in 1995, OpenDocument later) while pushing their own through when it extended reach.
No CS curriculum holds up Win32 as exemplary API design. No system copied it. A successful API earns adoption. Win32 enforced it.
> "...embrace-extend-extinguish against Java..."
Early Java was horrid for everybody except the architecture astronauts who could cram ten GoF design patterns into a hello world program. It only got traction because a different wannabe monopolist, Sun Microsystems, spent heavily to get it pushed into CS curriculums. Fortunately, the one-two punch of Linux and Intel killed Sun or we might all be cursing them today instead of Microsoft.
Sun was actually a decent company once upon a time. Their problem was that Microsoft excluded them from the market and starved them for money for so long that their hardware stopped being competitive, so that by the time Java made it so you could run some software on it, nobody wanted their hardware regardless.
It was only after they went bankrupt and got bought by Oracle that things like OpenSolaris getting killed off and Java lawsuits started happening.
> Wine and Proton are not tributes to Win32's portability.
I emphatically disagree. It is a hilarious and catastrophic failure of Linux userspace that the best API for running games on Linux is Win32. This has absolutely nothing to do with what Microsoft locking down the desktop market. It has 100% literally everything to do with Linux userspace being a clusterfuck of terrible design.
Linux adopted Win32 because it actually worked. They didn't have to. They could have simply invented a better API that didn't suck. But that's quite hard.
So yeah. I emphatically, but respectfully, disagree with your entire thesis.
> Linux adopted Win32 because it actually worked.
Completely untrue - Linux "adopted" Win32 because the majority of video games are written for Windows (and thus Win32).
They could not have invented a better API because the entire reason Proton exists is because developers don't build native Linux games.
All true, but simultaneously, if you look at it not in the “how did we get here” but “what’s out there today” light, it’s an option that can run on a lot of platforms. Not by some particular merit but because history happened this way - but that’s not a problem with the technology itself.
I concur.
The article defines "success" in the Windows context as being "available everywhere". It does not address how it got to that point.
And sure, you might not like Microsoft, and you may not like how it became successful (using the above definition) but the fact that it is available everywhere is not in dispute.
Of course most successful things have murky pasts. We don't necessarily agree with how it got there, but there it is. That is, at least in the technical sense irrelevant. You may prefer LP's or CD's, but streaming is now the successful way to get your music.
That doesn't mean it's the only way though, and of course you are free to not use Windows programs, or play games via Steam etc. That is your choice.
If you listen to CS professors you'd believe that Haskell would be very successful despite having limited if none real world usage. The reality is that catering to your users is much more important than having idealistic system.
In my experience even with games that have native Linux support running through proton seems to have less issues.
There's something deeply satisfying about compiling a Win32 desktop application and knowing that single binary will run unmodified on essentially every Windows machine from XP onward, and look nearly identical doing it. High-DPI is the one real caveat, but even that's manageable with a manifest. I'm not sure any other platform-native API comes close to that combination of reach and consistency. Running them on other operating systems is really just a bonus.
I am not sure if this holds, I have witnessed apps that dont work in newer versions of Windows.
They exist but they’re in the minority. Compare this with the Linux world, where userspace compatibility between one major distro release and the next, 12 months apart is very much a roll of the dice.
As an added bonus, since win32 is so old. AI is quite strong at generating code for it, I have even had success with qwen 3.5 1.5B parameters with win32 and opengl 1.1
Sometimes I think that, if Java ditched the idea of JVM and followed what modern languages did today (write once, compile on every major platform), things would be different.
What's even sadder is that was built and Sun fought against it because they were worried that devs would only compile for Win32.
The trouble there is that Sun was right about that.
Doing it that way works great for open source where anyone can recompile the software for a new target, but for proprietary software they would have given you a Windows blob and that's about it.
Meanwhile the problems with Java were mostly not the JVM. Its current problem is, of course, Oracle.
Hasn't it taken gargantuan multi decade efforts of hundreds of developers, multiple open source projects and the backing of major corporations to make it work at all, let alone well, on Linux?
How is that successful cross platform?
Hasn't it taken gargantuan multi decade efforts of hundreds of developers, multiple open source projects and the backing of major corporations to make [WINE] work at all, let alone well, on Linux?
Arguably there has been equal or larger effort invested/wasted in cross-platform and cross-distro frameworks/APIs/packaging and yet the result still doesn't work. Partly that's due to duplication of effort; there's (mostly) one WINE competing with Qt/Gtk/whatever times Snap/Flatpak/AppImage/whatever.
>> How is that successful cross platform?
Because hundreds of developers, multiple open source projects and the backing of major corporations made it happen, not because Microsoft wanted it but in spite of it.
In this case the route to success was via marketing (isn't it always so?), via market share, via application dominance (attracting developers to develop for the platform), and via insane levels of backward's compatibility. It was successful not because of the code itself (end users don't care 2 figs about the elegance of the code) but because they optimized for the end user experience.
Linux optimized for the experienced, technically adept user, who wanted to fiddle, customize and could write programs. Apple optimized for the "now", ignoring the past and regularly made existing programs obsolete and unrunnable.
I wrote windows programs in 1995. They still run today. They have run on all versions of Windows since then, without even a recompile. Everything I have [1] just keeps running. And it turns out, that's something users really want.
I get that we're all technical folk here. I get that we strive for technical excellence and elegance. I get that we operate in the "now", ignoring hardware and software from the past. But the market is different, and wants different things. If you want a successful business you need to understand the market, not just your own aesthetics. Microsoft understands that, and that's why the market (especially the business market) relies on them.
[1] - Except games. Copy protection on some of my games means they don't run anymore - but to be fair those were hacks designed specifically to prevent the game running in the first place.
Cuz there's good software written against it. And it doesn't need to be recompiled.
Will I ever target it? No, I'd rather you rip my bones and eyes out. But it's unarguably successful.
Considering you can use 500 kilobyte distro of tcc that can self compile to build win32 apps and those versions will very easily run on wine. I think it is a valid point and a strategy I have used for sometime. Since windows can't decide what you should use for developing Windows apps
> In the real world, the internet is TCP/IP
I guess he missed http3, which now makes up 35% of web traffic.
Technically the Internet runs on "the Internet Protocol Suite" but people just say "TCP/IP" for short. That doesn't mean to exclude UDP, SCTP, or whatever.
So what exactly does he mean then?
He means the Internet runs on the Internet Protocol Suite and not the "superior" OSI Protocol Suite (that people under age 50 have never heard of because it failed).
http3 operates on a different OSI layer
Yes, which the article claims is just a theory now, irrelevant for the real world. More crucially though, http3 doesn't use TCP because it is built on top of UDP.
It's like how people say "pass me the butter" despite there's just a tub of margarine on the table.
It's not that the train wasn't successful.
kinda