Surely Microsoft innovated, in a less-than-spectacular way most of the times with things like direct3d, plug-n-play etc. They added tcp/ip stack to windows and created internet explorer, even brought the internet directly to the desktop initially. While they did use their monopoly to push their products thus eliminating rivals, they did not directly block competition. You could still install netscape and windows is still a very open platform. I 'm not going to defend MS monopoly tactics of course, or make a scientific account of microsoft's pros and cons, but Windows had competitors in the open market, and yet they prevailed, not always because they were the better system, but they were not significantly worse than others.
Compare that to the oft-praised Apple model: directly block competitors from making similar functionality instead of competing with them and lock people into their systems. Does that model not impede technology from advancing?
Adding TCP/IP to Windows was an innovation? Winsock was pretty much lifted from BSD. The API is so compatible that you can target both with a simple include file to #define away some differences (that's a good thing, but it hardly indicates innovation on Microsofts side), largely because Windows can't easily support the equivalence between filehandles and sockets that implementing the BSD socket API 1:1 relies on.
This.
I remember having to play games without any sound because I couldn't figure out how to fit the SCSI Hard Disk driver and CD Rom driver and soundblaster driver into the 640k especially since lots of games flat out refused to work if you loaded himem.sys or QEMM which allowed you to load the drivers into the extended/expanded memory areas.
Adding TCP/IP - thanks, Microsoft, for not embracing and extending that one.
Internet explorer? Gee, thanks Microsoft for abusing your monopolistic powers to knock out Netscape and start embracing and extending the HTML platform?
The best defence you can mount is that they prevailed because they weren't significantly worse than others. That in itself says a lot -- but I'd make the argument they prevailed because they used tactics that make it hard for others to compete (including illegal ones).
It is in the "ISA-compatible hardware that doesn't need jumpers" sense, yes. It has a silly-sounding name, but it's a very specific set of hardware standards.
It was a huge, huge advance, speaking as somebody who got multiple graphics and sound cards working, simultaneously, with jumpers, back on an old pre-PnP 486 tower. I literally had a piece of paper with my IRQ and DMA channels and what cards they were assigned to, along with similar mappings for memory ranges, so that I could move the jumpers to reasonable locations for the next piece of hardware I added.
If you screwed it up, the new peripheral (and the old one it fought with) just stopped working. And then you got to go remap everything by literally moving a little fiddly piece of plastic from pin to pin on a card you had to open your computer to get at.
Plug-and-Play was huge. I originally griped about it too, back when it meant new hardware wouldn't work in my 486 :-P
Those are some rose-colored glasses there. I remember the plug-n-play days and there was a reason that people nicknamed it plug-n-pray. It wasn't until PCI that plugging boards into PCs became sane.
It was far from perfect. But do you remember just how awful it was to go fishing for jumpers in your computer's tower case after they popped off? And you had to keep lots of spare jumpers around because some memory addresses and IRQs required more physical jumpers to specify than shipped with the card. Yes, yes, jumpers were cheap... But also hard to order and impossible to get hold of on short notice unless you had friends with spares.
I occasionally rearranged which IRQs/memory cards used specifically to use fewer jumpers so that I would have enough.
PnP wasn't technically amazing. Yes, PCI was much better. Jumpers sucked.
Plug-and-Play was a shitty alternative when far superior systems existed on other platforms. Instead they opted for a really horribly limited system. Compare Autoconfig on the Amiga, for example, which was older and provided far more advanced capabilities and was well documented if they could've been bothered to at least copy something that worked well instead of coming up with yet another technically inferior solution.
Microsoft did the same with software. It's called "Embrace & Extend". Of the available evils at the time, the Amiga was ridiculously better thought out and implemented.
The PC hardware of 1995 was still quite far behind the Amiga hardware of 1985 (which had 14 bit stereo DMA sound or 4 channel 8 bit DMA sound, 2D hardware acceleration and sprites, and more). The PC operating system of 1995 was starting to equal the Amiga operating system of 1985 (Win95 had separate memory space, which was better than AmigaOS 1.0's shared space -- but otherwise, Win95 was mostly inferior)
I'd wait 1-2 years for cheaper standards. I'd rather pay a little more for getting to the future 10 years in advance.
Your comment about "the future" is interesting, considering how little the Amiga actually resembles what was eventually common or popular by 1995 ;)
For the hardware comparison to be fair, you'd need a 1985 PC, I think.
Let's be generous (to the Amiga) and assume a 1995 PC would probably be a 66MHz 80486DX2 with 8MB RAM, a VLB graphics card, and a Soundblaster of some description. Even at the same clock rate, a 486 of this type would handily beat the Amiga just from having a more efficient CPU and from not having a dirty planar framebuffer layout. But as it is, it is clocked 9x faster.
In practice, for 1995, a Pentium would not be THAT unlikely...
That Soundblaster was capable of 8-bit stereo; if you wanted to do anything non trivial like changing the sample rate, or mix more than one channel, that cost you a significant amount of CPU -- and most probably, you couldn't get more than 8 bit no matter what. The amiga had 14 bits (8 bit digital + 6 bit analog scaler, on each of 2 channels per side) in 1985, with hardware mixing.
You wanted a mouse cursor? that was zero cost in 1985's Amiga, thanks to hardware sprites; it was hell to get right on the PC (and while Windows mostly solved that if your app ran on Windows, WHICH WAS NOT A GIVEN in 1995, many apps still had cursor bugs, even in Windows).
Yes, the 1985 68k was slower than the 1995 486. But 2D hardware acceleration made up for a lot of it -- the Amiga 1985 GUI was more responsive than the 1995 Win95 GUI (that's hard to appreciate on emulators today -- you had to use it to believe it).
And finally, multitasking with 512K ram on the Amiga 1000 on AmigaOS 1.0 actually worked properly - way better than Win 3.11 (which was what you had on the PC until 1995), and mostly on par with Win95, even though the latter had MMU separation to help (which 68k did not in 1985, but 68020/AmigaOS 3 did in 1995, IIRC).
Really, look at Amiga games like Shadow of the Beast (1989), Sword of Sodan (1988), The Great Gianna Sisters (1989), and compare it to PC games of 1995.
The Amiga hardware of 1985 was better than the PC hardware of 1995 in every way except raw CPU power.
"Except raw CPU power" is not much of an "except". If you have raw CPU power, that fixes everything. You don't need a blitter, and you don't need sprites, and you don't need playfields, and you can do your sound mixing on the CPU. Then 256 colours makes up for some of the things you'd use the copper for; the rest, the CPU power will sort out. A byte-per-pixel framebuffer memory layout makes easy what would be a thorough pain on the Amiga - though with a faster CPU, not like it would be a massive problem anyway.
"Except raw CPU power" is why the Amiga had all the custom junk in the first place - it didn't have the raw CPU power to do without!
(As for the mouse cursor, what need is there for a zero-cost one? Just draw the cursor at the end of the frame, when you're done. No problem. I don't remember any mouse cursor problems in games; presumably they all did what I suggest, because it's the obvious, easy, reliable thing to do.)
Some PC games I remember playing in 1995, or before: Doom II, Hi-Octane, Magic Carpet, Syndicate (the Amiga version of this on a stock Amiga was a sad joke), Command and Conquer, Wing Commander (the Amiga version of this was pretty poor, even on an Amiga 1200), Comanche: Maximum Overkill, Microprose Formula 1 Grand Prix, Stunt Island, various of the X-Wing series, and Indianapolis 500.
Looking at the games you cite, I think the PC ones are much better! Better graphics, due to having more colours or resolution. More interesting play, due to having more CPU time to do interesting stuff. Yes, even better sound, because you've got as many channels as there's CPU time for, and more accurate stereo positioning because of it.
OK, so I like the art direction of Shadow of the Beast. That's about it.
> Then 256 colours makes up for some of the things you'd use the copper for; the rest, the CPU power will sort out. A byte-per-pixel framebuffer memory layout makes easy what would be a thorough pain on the Amiga
That's simply not true.
Take e.g. Doom II from your list; In order to have double buffering (required for smooth screen update), it had to use http://en.wikipedia.org/wiki/Mode_X which is .. a planar mode even if it is 1-byte per pixel (in a slightly nicer way than the Amiga's planar mode; but switching planes cost an arm and a leg in I/O communication, so to actually work efficiently with it, it was as bad as the Amiga's planar mode).
> If you have raw CPU power, that fixes everything. You don't need a blitter, and you don't need sprites, and you don't need playfields, and you can do your sound mixing on the CPU.
I remember one (1) reasonable multi-level parallax scrolling game for the PC with reasonable performance, but not its name. (I remember it well, because I spent a few afternoons with a debugger trying to figure out how the f*ck they managed). CPU was NOT enough to not need all of these.
Almost all the games you mentioned have a 3D point of view (those that I'm familiar with, anyway) - for which the Amiga's '85 2D acceleration wasn't helpful, of course. But the 2D games of the PC, especially scrollers, were inferior.
Non of them, as far as I can tell, had 3d sound or multi channel stereo mixing in 1995 -- PC mod players back then were struggling to go below 30% CPU (I know, I wrote one of the fastest ones), and no game could afford them.
I released a sideways scroller for the PC in '94, that achieved constant 60fps (with not a lot of CPU to spare) on the 486DX of the time, and a respectable 30fps on the 386 that were still super common. To get that, I needed to use ModeX, used Adlib for music so I can use DMA sample playing for the effects (also, pure adlib was _still_ common enough to warrant support if you didn't have a soundblaster). It needed an awful lot of tricks to get reasonable performance; The graphics engine was all assembly, and some of the game logic too. And I needed continuously recalibrating the int8 timer to get interrupts at the vertical retrace so that everything worked properly without wasting CPU.
I never got to do an Amiga version, but it would have been 100 times simpler. And would have worked equally well on an 1985 Amiga 1000.
Setting the precedent that web browsers should be free (even for commercial use)?
Adding XHR to javascript?
Making plug & play work (yes, it's a standard now)?
MS prevailed because at almost every critical junction over the last 30 years they've managed to put together the best collection of compromises that maximized utility for the average computer user.
Look at the evolution of the Macintosh throughout the 1990s and 2000s as a comparison. Through almost every iteration it has continued to try to hew to some "higher" standard than the average PC. When PCs were using IDE Macs were using SCSI. Until Macs started using IDE too. When PCs were using tcp/ip over ethernet Macs were using appletalk, until Macs started using tcp/ip too. Firewire/usb. PowerPC/x86. Etc.
There's a reason why the de facto filesystem for flash drives is FAT32 and it's not just because windows is so popular (if so it would just be ntfs), it's because it represents a well balanced compromise and is dead simple to implement.
Windows has been a fairly decent embodiment of "worse is better" for a long time.
Because OpenGL is Not Good, even to this day. It's a brutally fragmented system that's got little going for it except ubiquity; the problems with it are numerous and ugly. (But you can't expect much from something where everything's a frigging GLuint, can you?) It is popular because it's everywhere, not because it's good.
Direct3D was initially much worse than OpenGL, which they saw as a threat because it was cross platform. To this day they are doing something similar with the refusal to adopt WebGL as a standard in their browsers.
Just ask yourself these two questions and you'll see what I mean.
Did Microsoft contribute in advancing certain areas of computing and speed up the adoption of Internet technologies?
Did Microsoft contribute in impeding certain areas of computing and slow down the adoption of Internet technologies?