The type of apps I used to run without delays on a Win98 Pentium 200+MHz w/ 64MB of RAM continues to be an illustration of how bloated today’s apps are.
I don’t mean to sound like a grumpy curmudgeon repeating the same old gloomy refrain, but I believe that some day Wirth’s law will win over Moore’s law.
I’d have previously say it was in embedded at least. Then, the recent surveys on that show they’re cramming all kinds of complexity into devices that ought to be simple. A lot can’t be simple given even basic requirements like WiFi on top of something that otherwise might use 8-bitter with simple, control software. Wirth’s philosophy was never grounded in user psychology or market economics. Seems stuff has to be overcomplicated at least a little bit even in sectors benefiting from simplicity. That’s even happened for “high-assurance” systems to a degree where market demand forced more politics into evaluations increasing complexity.
I’m not saying Wirth’s philosophy is dead. I’m saying it was always partly wrong with even more potential proponents working against it. I always respected it but thought it overly simplistic. The compiler heuristic was especially bad once hardware got faster. I’m still about leaner baselines. I just think a bit more complexity in language and stack is good if it makes every app on top have desirable qualities. Wirth was always willing to complicate everything on top to make compiler/runtime/OS simpler. That’s backwards to me after watching how people actually use those things.
That would be quite a hack. I doubt Electron could even be made to run on Windows 95. Once Windows 98 came out, Win95 was all but forgotten by 99% of the computing world in short order. I would guess that most programs of the pre-Win7 era that are still actually useful have roughly this level of support:
Pre-Win7 would have been Windows Vista. Nearly all programs should have run on Windows XP that were being developed on Vista. Typically you’re going to want to target the current release and at least the last major release. I think you’re correct about 98 and 95 though. Even today with Visual Studio 2017 compiling C++ I can target Windows 7, although I think by default you only get to target Windows 10 and Windows 8.x
It may be worth noting that copyright on Windows 95 is still intact. In all likelihood, this infringes upon Microsoft’s copyright.
ok but why
[Comment removed by author]
I guess it’s a bit of a moot point, since it’s an Electron app, but I’d like to bring up the fact that Win95 could comfortably run w/ 16 MB core.
The type of apps I used to run without delays on a Win98 Pentium 200+MHz w/ 64MB of RAM continues to be an illustration of how bloated today’s apps are.
I don’t mean to sound like a grumpy curmudgeon repeating the same old gloomy refrain, but I believe that some day Wirth’s law will win over Moore’s law.
I’d have previously say it was in embedded at least. Then, the recent surveys on that show they’re cramming all kinds of complexity into devices that ought to be simple. A lot can’t be simple given even basic requirements like WiFi on top of something that otherwise might use 8-bitter with simple, control software. Wirth’s philosophy was never grounded in user psychology or market economics. Seems stuff has to be overcomplicated at least a little bit even in sectors benefiting from simplicity. That’s even happened for “high-assurance” systems to a degree where market demand forced more politics into evaluations increasing complexity.
I’m not saying Wirth’s philosophy is dead. I’m saying it was always partly wrong with even more potential proponents working against it. I always respected it but thought it overly simplistic. The compiler heuristic was especially bad once hardware got faster. I’m still about leaner baselines. I just think a bit more complexity in language and stack is good if it makes every app on top have desirable qualities. Wirth was always willing to complicate everything on top to make compiler/runtime/OS simpler. That’s backwards to me after watching how people actually use those things.
I am reasonably certain my first (and only) Win95 machine was a Pentium 100 MHz with 8 MB of RAM.
Next level: port electron to windows 95, run Slack on this.
Ironically, the maker of the win95-in-electron hack works at… slack. https://github.com/felixrieseberg
That would be quite a hack. I doubt Electron could even be made to run on Windows 95. Once Windows 98 came out, Win95 was all but forgotten by 99% of the computing world in short order. I would guess that most programs of the pre-Win7 era that are still actually useful have roughly this level of support:
Pre-Win7 would have been Windows Vista. Nearly all programs should have run on Windows XP that were being developed on Vista. Typically you’re going to want to target the current release and at least the last major release. I think you’re correct about 98 and 95 though. Even today with Visual Studio 2017 compiling C++ I can target Windows 7, although I think by default you only get to target Windows 10 and Windows 8.x