When you buy the software at the store, you get one of those little cards like you do now that has a product key on it. Right now, you just put that code in and it enables the software and off you go.
With the new plan, you'll still get the product key. The difference is that when you put that into the software on installation, it'll contact MS with the key and some hash of your hardware. Then they'll keep all that info in a big fatty database somewhere. When you want to reinstall, you'll put in the same product key you bought with the software. Then Office'll go register, they'll bring up the earlier registration info from the database, and check your new hardware key against the old one. If they're the same, you go on your merry way. If they're "within tolerance" (only slightly different, for some definition of "slightly"), they'll give you a new key and increment the number of keys they've given you. If you're out of tolerance, you have to call MS and they'll make a call about whether you get to install again.
But they *are* working on all that good luvin'. To increase parallelism, the folks working on Alpha and Intel's Foster are doing SMT. Intel's working on Infiniband to increase bus speeds, while AMD's got LDT which will either replace or aid the Infiniband thing (depending on who you ask). DDR's basically here, and with luck we'll see a QDR standard soon (Quad Data Rate - like Double, but Quad ).
The thing is, each of the groups needs to keep on working, and getting a single-atom transistor in the lab is a far cry from getting one in your grandma's new computer from CompUSA. You've gotta lay the research groundwork, then maybe get some standards, and only then start worrying about actually producing something to sell. All of the things in the first paragraph are pretty much pie-in-the-sky at this point, but so are single atom transistors.
Yes, I know this is still a troll, but currently with things like GNOME, most of the innovation is in the programming APIs and code implementations - the actual user interface is neglected, programmers are just happy to leave it looking like existing interfaces because they're not UI experts, and they at least want the user to be instantly familiar, even if they do just steal layouts (such as M$ does heavily, eg Start button vs Apple menu). Personally, if I were to come up an innovative compression method, the user would not care. All he would care is that my program had the same user interface as zip, otherwise he'd say "it compresses much better, but it's a bitch to use!"
This comment points out one of the big areas where corporations like MS, and more notably Apple, have a significant and obvious advantage over Open Source: They _design_ software.
Most Open Source projects are run by programmers, and for programmers. As a result, you get programs which are very satisfying to programmers. But pretty obviously, the vast majority of the population is not made up of programmers. As the poster points out, most Open Source improvements consist in better algorithms, making the program faster, smaller, and better by a variety of programming metrics. If Open Source wants to succeed in getting onto the desktops of grandma's across the world, it needs to get some design - artists, interaction designers, user-interface designers, and the like.
The problem is that most programmers left to their own devices will not listen to a designer. They do what's right for the program, for the machine, and tend to ignore the poor sap who's actually got to use the thing. The advantage that corporations have is that they have been building in the process to get design involved (think Apple), which makes for great design and very loyal customers, who love the product. It's not something that is unique to corporations, and it could get into Open Source movements if they would acknowledge that it's important. But with programmers doing the design for other programmers, it's just not happening. And when my grandma goes to buy a computer, it doesn't matter how amazingly the TCP/IP stack optimizes for X, or how the VM does this kind of awesome pre-paging, or whatever. If it's ugly, or more imporantly, if it's tough to use and makes her feel stupid, she's not going to use it. In her mind, saving $90 is worth a lot less than having to feel like an idiot every time she has to call somebody because some esoteric configuration file got corrupted, or she has has to compile some new software she just got.
So if Open Source wants to really take a shot at the desktop, it needs to see the importance not just of good programming, but also of good design.
Slashdot Mirror
With the new plan, you'll still get the product key. The difference is that when you put that into the software on installation, it'll contact MS with the key and some hash of your hardware. Then they'll keep all that info in a big fatty database somewhere. When you want to reinstall, you'll put in the same product key you bought with the software. Then Office'll go register, they'll bring up the earlier registration info from the database, and check your new hardware key against the old one. If they're the same, you go on your merry way. If they're "within tolerance" (only slightly different, for some definition of "slightly"), they'll give you a new key and increment the number of keys they've given you. If you're out of tolerance, you have to call MS and they'll make a call about whether you get to install again.
The thing is, each of the groups needs to keep on working, and getting a single-atom transistor in the lab is a far cry from getting one in your grandma's new computer from CompUSA. You've gotta lay the research groundwork, then maybe get some standards, and only then start worrying about actually producing something to sell. All of the things in the first paragraph are pretty much pie-in-the-sky at this point, but so are single atom transistors.
This comment points out one of the big areas where corporations like MS, and more notably Apple, have a significant and obvious advantage over Open Source: They _design_ software.
Most Open Source projects are run by programmers, and for programmers. As a result, you get programs which are very satisfying to programmers. But pretty obviously, the vast majority of the population is not made up of programmers. As the poster points out, most Open Source improvements consist in better algorithms, making the program faster, smaller, and better by a variety of programming metrics. If Open Source wants to succeed in getting onto the desktops of grandma's across the world, it needs to get some design - artists, interaction designers, user-interface designers, and the like.
The problem is that most programmers left to their own devices will not listen to a designer. They do what's right for the program, for the machine, and tend to ignore the poor sap who's actually got to use the thing. The advantage that corporations have is that they have been building in the process to get design involved (think Apple), which makes for great design and very loyal customers, who love the product. It's not something that is unique to corporations, and it could get into Open Source movements if they would acknowledge that it's important. But with programmers doing the design for other programmers, it's just not happening. And when my grandma goes to buy a computer, it doesn't matter how amazingly the TCP/IP stack optimizes for X, or how the VM does this kind of awesome pre-paging, or whatever. If it's ugly, or more imporantly, if it's tough to use and makes her feel stupid, she's not going to use it. In her mind, saving $90 is worth a lot less than having to feel like an idiot every time she has to call somebody because some esoteric configuration file got corrupted, or she has has to compile some new software she just got.
So if Open Source wants to really take a shot at the desktop, it needs to see the importance not just of good programming, but also of good design.