I don't know where you saw Win3.1 at that price. The typical retail price that I saw, back in the day, was about $35-$50 tops. I sure like where I shopped better than where you shopped.:)
That was the Windows 3.1 recommended retail price. Your $40ish copies were probably OEM versions.
Now, the OS is priced at 50% to 100% of the price of the average consumer-grade PC. Something's wrong with this pricing, inflation and Moore's Law and all. Do you really feel that the OS is HALF the value of your machine?
Except "nobody" pays full retail price for Windows. The vast, vast majority "pay" however much their OEM has negotiated when they get a new PC. Of the few people who actually buy Windows, most of them buy upgrade versions or "retail OEM" versions.
In short, the price comparison you present is silly, because it's utterly unrealistic.
Further, OSes today have significantly functional improvements over those sold back in the early '90s. This is in contrast to the hardware, which offers next to no functional improvements (networking is the only one that springs to mind - maybe sound and 3D accleration, depending on how you're counting), and is basically just a lot faster and bigger than it used to be.
Also, historically, M$ stock did well (roughly doubling in value every year) up until XP and activation and price confusion. Since then, it's been flat. One can't help but consider a correlation.
Yeah, I'm sure that was completely due to Windows XP. The whole tech market meltdown thing wouldn't have had any impact at all.
Drivers are not hardware. Above the hardware was pointed out for blame.
No, both hardware and software were blamed.
The driver model is defined by Microsoft and allows mapping of memory into the space that renders PAE useless which means the driver model is the problem. Furthurmore I am very suprised and disappointed that the problem did not go away with Vista, a slightly different driver model and driver certification.
Microsoft can define the driver model, but they can't make people follow it properly (well, they can, but look at the shitstorm and conspiracy theories when they try).
Arguing that Microsoft is not responsible for defining core components of their operating systems is a little odd.
I'm not. I'm arguing they're not responsible when hardware vendors write shitty drivers, or don't bother to test them with machines using PAE (although given the scarcity of machines running in PAE mode, it is somewhat excusable for vendors targeting their products at the consumer market).
Personally I hope we'll be seeing more 64 bit software so we can abandon their 32 bit platforms and just forget about it - for the moment now we're stuck using disk for virtual memory for some applications due to this flaw.
I'm not sure why you think 64 bit removes the need to use disk for virtual memory. Many of my Linux machines are using gigabytes of swap, even though they are running in x64 mode and have 32G of real RAM in them.
People howled a lot less when Windows cost $35 than when it costs $350.
When did Windows cost $35 ? Windows 3.1 had a retail price of $150.
Which, in today's dollars, is about $240. Add on the cost of DOS 5.0 ($100) adjusted for inflation ($160) and you get to $400.
So, in real terms, Windows Vista "Ultimate" costs as much today as DOS+Windows 3.1 did in 1992. Taking into account a) the massive increase in actual functionality and b) cheaper versions of Vista, it's quite clear that, far from being "expensive", Windows is - comparatively - "cheap".
At $400, considering that the cost of other consumer-level software has NOT gone up by a factor of 10 [...]
By no rational measure whatsoever, has the price of Windows gone up by a factor of 10. Windows Vista "Home Basic" - a more reasonable modern comparison to DOS+Windows 3.11, has a retail price of $200. In other words, in real terms Windows has halved in price since 1992.
You can do whatever you want with GPL'd code, so long as you contribute your changes back into the "commons".
No, you must *GPL* your code. That's not the "commons".
BSD license is ok as long as you're ok with writing code and someone else taking that code, profiting from it without giving you *anything* in return. You don't even have the right to run your own code without paying them.
Yes, you do. You might not have the "right" to run your code plus their modifications, but your code remains under the BSDL.
I believe MS took a BSD-licensed IP stack, modified it, and are now (potentially) selling Windows back to the people who wrote the stack without any compensation. This would not be possible with a GPL'd implementation. MS would have either had to incur the expense of writing their own, or (better?!) return their changes back into the commons for others to use as well.
It's been a long, long time since Windows had a BSD-derived IP stack.
Incidentally, the Linux community did exactly what you describe. They took the BSD licensed code and GPLed it - or, conceptually, the same thing you're so annoyed at Microsoft doing.
Obviously, to each their own, but I personally wouldn't want to share with people who don't share back. I've no problem with sharing with people who will share back, and that is the definition of the GPL.
Why should someone whose software is 90% their own work and 10% GPLed code (and only in shared libraries, at that) "share back" all their hard work ? Where's the fairness in that ?
The discussion has come up many times before and I know RMS has stated that there is no better word than free and it only takes a second to explain that you mean free as in freedom.
Of course, it takes a lot longer to explain you really mean "GNU/freedom".
I'm still mostly taken aback that an iMac would do a better job of running Vista with 1GB of RAM than a PC did with 2, even though they had the same processor and whatnot.
They don't have the same processor. Not only is the Mac's processor ca. 20% faster by clock, it's also faster clock-for-clock. It's impossible to compare the video cards from the information given, but I'd be willing to bet there's more inequality there as well.
More interesting is that benchmarks whose results should be basically identical (minus statistical noise) aren't - eg: the Integer and Floating Point tests. The results here say a lot more about the usefulness of Geekbench as a cross-platform benchmark, than they do about either the hardware or the OSes.
Really, because pretty much every company I've ever worked at has profited from innovations we added to OSS code (often GPL). What you're missing is the business model of developers as a service.
No, I'm not missing it in the slightest. I'm *making* the point (which everyone going "GPL, rah, rah, rah" just quietly ignores) that you can't make money with GPLed software UNLESS you're going to tie your software to some other (non-GPLed, or equivalent) product or service (ie: "developers as a service").
The GPL is not the only way to make money creating and selling software. It is not the most profitable way all the time. To pretend it cannot be done, however, is to ignore a significant portion of today's software market.
Who is making money selling only GPLed software ? Not GPLed software tied to support contracts. Not GPLed software tied to hardware. Not dual-licensed software. Not "work for hire" that just happens to be GPLed. Just plain old GPLed software.
Yeah, if you're locked into that business model and in that particular situation, then you screwed yourself choosing to use GPL code in the first place. Why not write the other 1% yourself instead of agreeing to swap your 99% of the code for that 1% work from others?
I'm glad we agree that the GPL is not the best or only solution to software licensing. Baby steps. No we just have to convince all the other GPL zealots.
That said, I've never known anyone stupid enough to put themselves in your hypothetical situation.
Indeed, that's because they're so morally bankrupt they write closed-source code instead.
From this, we may safely draw the conclusion that Microsoft has done a lot of research, with a lot of lawyers, and they've determined that the GPL represents the biggest threat to their revenue model. And what's bad for Microsoft is generally good for everyone else. So if you're going to develop FOSS, the GPL is the obvious safe choice.
The GPL represents a threat to anyone who wants to sell software. Microsoft are hardly alone in this.
The GPL may have more restrictions than BSD, but the GPL itself doesn't take anything away [...]
Practically speaking it does, due to the ludicrously wide definition of a "derivative work".
The most obvious example of this is linking to shared libraries. If those shared libraries are GPLed - sorry, you're going to have to GPL your code as well[0]. Contrast this to the average "closed source" shared library, that requires nothing of the sort.
[0] Or maybe you won't. It hasn't actually been tested in court yet as far as I know - but it's certainly not a risk any sane person would take.
The GPL adds restrictions that protect software freedom and innovation.
The GPL makes it essentially impossible to profit from your innovations without tying your code to some other non-GPLed products (be they something as simple as a support contract or something more concrete like a piece of hardware).
*That* is why some people don't think it's the best thing since sliced bread. They want to make money selling software. Even if you have some awesome idea for a piece of software that every man and his dog will want to use, and do 99% of the heavy lifting yourself, that 1% worth of boring, mundane crap you used GPL code for essentially brings the potential price for your software crashing down to roughly $0.
Joe sixpack is not concerned what instruction set is in his DVD player or Microwave Oven.
Joe sixpack isn't trying to use his DVD player or microwave oven to browse the web, pull pictures off his digital camera or write letters to mum, either.
The UMPC is not a PC. It is a quick boot, pocket sized device that can view web pages and in a pinch edit documents. So end users don't expect to run PC Software on it.
The UMPC is a _laptop_. As such, people will expect it to be able to do everything the laptop they own today does. Which is, for all intents and purposes, run x86 code. The average customer is not the geek who wants multiple computers for toys.
No normal person wants multiple computers if they can avoid it. Multiple computers are a pain in the arse. Just like multiple mobile phones, music players, and the like.
The benchmark for good enough, is something that feels fairly snappy and is able to video cam, view web pages, play audio and full screen video with at least a resolution of 800x640. As long as people can stand the keyboard, touch pad and the quality/size of the display, it will sell.
I am curious where you're getting your idea of "good enough" from. Because by your measure "good enough" PCs were around back in about 1998, yet strangely people have been snapping up better machines since then in earnest.
The lower end. Which is what users are expecting as a computing experience. Is not moving upward any longer.
Of course it is. Has been for years. And in the next 6-12 months its going to move up to meet Vista's needs (which isn't much further anyway). What doesn't move downward often is price points - and this new US$400ish price point isn't going downwards any time soon.
Here's the crux of the issue you (and many others) seem to be having the most trouble with: hardware prices at the bottom end of the price range do not scale down in the same way they scale up (and similarly at the top end). Just because, in 6-9 months, you'll be able to buy a dual core, 1.6Ghz, 2G RAM UMPC for $400 does *NOT* mean that for $200 you'll be able to buy a machine with half the power (ie: today's Eee PC and equivalents). You might be able to get a machine half as powerful for $300 - $350, if you're lucky enough that anyone sells them - at which point buying the more powerful and more useful machine is an easy choice.
What people will not buy, is a sub $400.00 device where the Microsoft Tax is such a huge proportion of the cost it is not worth it. If there was a DVD player for $50.00 and one for $450 and the only difference is the one for $450 comes with a license for Windows Vista. Guess which one is going to sell?
The "Microsoft Tax" is a non-argument. Always has been. I don't know why anyone bothers to even bring it up. Microsoft can and will sell Vista to OEMs for whatever price is necessary. If they have to "create" a special "UMPC OEM" license at $10/copy, they will.
I am expecting that by August 2009 that "good enough" hardware will be in the $250-$275 price range with 1024x768 resolution (1280x800 when docked). This way if she loses it or it gets broke, I can buy her another one. I want a cheap throw away appliance.
Today's UMPCs aren't "good enough". They're not "good enough" because they're not capable of being the typical person's only computer.
Your average UMPC in August 2009 is going to be $350-$400, have a 9-10" screen @ 1280x720, dual cores @ 1.2-1.6Ghz, 2G (with 4G a cheap and obvious upgrade) of RAM and be sold running Vista. This is the point at which their popularity is really going to explode because this is the point at which it will be quite feasible for a UMPC to be the average person's only computer.
This trend is going to continue to drive prices down. I don't think x86 compatibility will matter. Linux fills this space just fine. For those that want a UMPC that does more. Heck just pay an extra $25. All of the Linux apps
The $400ish price point is going to be around for a good year or two. The amount of power available at this price will increase, but the price will not decrease (at least not significantly).
This has been the consistent pattern with computer hardware for the last couple of decades. I'm not sure why anyone thinks it's suddenly going to change.
Further, a "reduced spec" machine will be, like it is today, little more than a toy. With the US (and a fair chunk of the world) heading into a recession, the market for toys is going to dry up pretty quickly.
As it turns out. That would also mean there will be a $250.00 or $300.00 machine with about the same specs as now.
What do you base this on ? Because it's certainly not the typical pattern that computing hardware pricing follows. Power at a given price point increases far, far faster - and more frequently - than price points decrease.
The thing everyone seems to be missing is this. Currently these machines don't quite have enough "oomph" to be the one machine the average person needs. They're toys for geeks.
However, with the power they'll have in 6-12 months, they will quite feasibly be the only computer the average person needs. The big disadvantages - screen and keyboard - will be nullified by a cheap docking station that people can just snap their "UMPC" into and out of in a second. It's going to be like an ipod - get home and drop it into the "dock", rip it out as you run out the door.
And most of them are going to be sold running Vista.
Microsoft can't close that door. If someone can knock the price down and improve performance even more by moving to something that is NOT x86 compatible. Microsoft is about to feel a swift kick to their family jewels.
No chance. A non-x86 machine isn't going to run the software most people want to run.
If she could see him, why would she need to use the software to take a snapshot? Don't apple computers have Prnt Scrn or an equivalent button?
If the camera display on-screen is done via an overlay in the video card, then a "capture screenshot" won't work (you'll just get a black box where the camera output would be).
No idea if that's how OS X handles it, just sayin'...
Those of us who were around then will remember that Microsoft mounted a sustained attack on Java, and deliberately crippled it's multi-platform capabilities.
And by "deliberately crippled", you mean "released the best JVM", right ?
Linux is a family of operating systems. The kernel isn't bloated, and there are Linux operating systems that require 8MB of RAM or less.
Unbelievably bloated. DOS 3.3 would run on a machine with half a megabyte of RAM.
There are no bloat-free versions of Vista. If Vista was modular and I could remove components, I would have no problem with the bloat. Alas..
Sure you would. You'd just be able to remove it. Which exposes the colossal arrogance of your position - it's only "bloat" because _you_ don't think it's a worthwhile feature. In the context of systems with equivalents features - say, OS X - Vista isn't bloated. You need similar hardware to get similar results.
They're pretty damned close -- even though the iMac had a faster processor, the PC had 3x the RAM.
The iMac has a 20% clock speed advantage. In and of itself, that's not trivial. To say nothing of it having a CPU that noticably faster per clock because of the larger L2 cache (6M vs 2M) and higher bus (1333Mhz vs 800Mhz). Finally, with an odd amount of RAM like 3G, I'd be willing to bet the Gateway also has unmatched DIMMs, meaning it's not using interleaved memory, further hurting performance.
Regardless of OS, one would expect the iMac to be ~20-30% faster than the PC (which is basically what the Vista benchmarks demonstrate).
I am also somewhat sceptical of Geekbench's relevance as a cross-platform benchmark. For example, one would expect the solely-CPU benchmarks (Integer and Floating Point) to be essentially identical on both OS X and Vista. Instead we see OS X is significantly slower in integer and significantly faster in floating-point. Similarly for the "graphics"
benchmark (indicating possible driver discrepancies).
It would be interesting to compare Vista and XP numbers, however, I suspect these results say more about Geekbench and its usefulness as a cross-platform benchmark, than they do about how Vista and OS X compare.
(These are issues that call into question any conclusions that are apparent from just a cursory look at the parameters. I'm sure there are others that would appear on deeper analysis. For example, did they give Vista's initial disk index the chance to finish before running benchmarks ?)
I'm sorry but how can anyone seriously argue that "our program is bloated, but it's OK because relatively decent hardware can handle it!" with a straight face.
Because the only logical conclusion to the opposing argument is an abacus.
If you have to argue that your bloat is acceptable, then it's too bloated.
People who can still do their work on a DOS 3.3 machine would argue Linux is incredibly bloated. Would you agree with them ?
I've taken multiple OS courses, and none of them discussed Windows directly. I never did Windows development. I think you would find that many CS students simply don't care about Windows, and you 'expect [me] to be well aware of' it.
I'm incredulous any university-level OS "course" wouldn't cover Windows NT, probably the next most important OS in the modern world outside of UNIX (and derivatives).
Any "CS student" who "simply doesn't care about Windows" should be ashamed to call themselves such. Mindless bias like that has no place in academia.
These machines will never run Vista well. Let's keep that important knowledge in front of people. Intel expects to move 10 million Atom platforms in the first wave, and none will have Vista.
What ? A dual core machine with 2G of RAM will run Vista fine. *Most* of them will be sold running Vista.
Slashdot Mirror
Come on Apple, put out a mid-tower priced around $800-$1,000 and they will sell like crazy.
That's the problem. A machine like that would slaughter higher margin Mac Pro sales.
I don't know where you saw Win3.1 at that price. The typical retail price that I saw, back in the day, was about $35-$50 tops. I sure like where I shopped better than where you shopped. :)
That was the Windows 3.1 recommended retail price. Your $40ish copies were probably OEM versions.
Now, the OS is priced at 50% to 100% of the price of the average consumer-grade PC. Something's wrong with this pricing, inflation and Moore's Law and all. Do you really feel that the OS is HALF the value of your machine?
Except "nobody" pays full retail price for Windows. The vast, vast majority "pay" however much their OEM has negotiated when they get a new PC. Of the few people who actually buy Windows, most of them buy upgrade versions or "retail OEM" versions.
In short, the price comparison you present is silly, because it's utterly unrealistic.
Further, OSes today have significantly functional improvements over those sold back in the early '90s. This is in contrast to the hardware, which offers next to no functional improvements (networking is the only one that springs to mind - maybe sound and 3D accleration, depending on how you're counting), and is basically just a lot faster and bigger than it used to be.
Also, historically, M$ stock did well (roughly doubling in value every year) up until XP and activation and price confusion. Since then, it's been flat. One can't help but consider a correlation.
Yeah, I'm sure that was completely due to Windows XP. The whole tech market meltdown thing wouldn't have had any impact at all.
Drivers are not hardware. Above the hardware was pointed out for blame.
No, both hardware and software were blamed.
The driver model is defined by Microsoft and allows mapping of memory into the space that renders PAE useless which means the driver model is the problem. Furthurmore I am very suprised and disappointed that the problem did not go away with Vista, a slightly different driver model and driver certification.
Microsoft can define the driver model, but they can't make people follow it properly (well, they can, but look at the shitstorm and conspiracy theories when they try).
Arguing that Microsoft is not responsible for defining core components of their operating systems is a little odd.
I'm not. I'm arguing they're not responsible when hardware vendors write shitty drivers, or don't bother to test them with machines using PAE (although given the scarcity of machines running in PAE mode, it is somewhat excusable for vendors targeting their products at the consumer market).
Personally I hope we'll be seeing more 64 bit software so we can abandon their 32 bit platforms and just forget about it - for the moment now we're stuck using disk for virtual memory for some applications due to this flaw.
I'm not sure why you think 64 bit removes the need to use disk for virtual memory. Many of my Linux machines are using gigabytes of swap, even though they are running in x64 mode and have 32G of real RAM in them.
People howled a lot less when Windows cost $35 than when it costs $350.
When did Windows cost $35 ? Windows 3.1 had a retail price of $150.
Which, in today's dollars, is about $240. Add on the cost of DOS 5.0 ($100) adjusted for inflation ($160) and you get to $400.
So, in real terms, Windows Vista "Ultimate" costs as much today as DOS+Windows 3.1 did in 1992. Taking into account a) the massive increase in actual functionality and b) cheaper versions of Vista, it's quite clear that, far from being "expensive", Windows is - comparatively - "cheap".
At $400, considering that the cost of other consumer-level software has NOT gone up by a factor of 10 [...]
By no rational measure whatsoever, has the price of Windows gone up by a factor of 10. Windows Vista "Home Basic" - a more reasonable modern comparison to DOS+Windows 3.11, has a retail price of $200. In other words, in real terms Windows has halved in price since 1992.
You can do whatever you want with GPL'd code, so long as you contribute your changes back into the "commons".
No, you must *GPL* your code. That's not the "commons".
BSD license is ok as long as you're ok with writing code and someone else taking that code, profiting from it without giving you *anything* in return. You don't even have the right to run your own code without paying them.
Yes, you do. You might not have the "right" to run your code plus their modifications, but your code remains under the BSDL.
I believe MS took a BSD-licensed IP stack, modified it, and are now (potentially) selling Windows back to the people who wrote the stack without any compensation. This would not be possible with a GPL'd implementation. MS would have either had to incur the expense of writing their own, or (better?!) return their changes back into the commons for others to use as well.
It's been a long, long time since Windows had a BSD-derived IP stack.
Incidentally, the Linux community did exactly what you describe. They took the BSD licensed code and GPLed it - or, conceptually, the same thing you're so annoyed at Microsoft doing.
Obviously, to each their own, but I personally wouldn't want to share with people who don't share back. I've no problem with sharing with people who will share back, and that is the definition of the GPL.
Why should someone whose software is 90% their own work and 10% GPLed code (and only in shared libraries, at that) "share back" all their hard work ? Where's the fairness in that ?
The discussion has come up many times before and I know RMS has stated that there is no better word than free and it only takes a second to explain that you mean free as in freedom.
Of course, it takes a lot longer to explain you really mean "GNU/freedom".
I'm still mostly taken aback that an iMac would do a better job of running Vista with 1GB of RAM than a PC did with 2, even though they had the same processor and whatnot.
They don't have the same processor. Not only is the Mac's processor ca. 20% faster by clock, it's also faster clock-for-clock. It's impossible to compare the video cards from the information given, but I'd be willing to bet there's more inequality there as well.
More interesting is that benchmarks whose results should be basically identical (minus statistical noise) aren't - eg: the Integer and Floating Point tests. The results here say a lot more about the usefulness of Geekbench as a cross-platform benchmark, than they do about either the hardware or the OSes.
If you're going to say "not all are threatened" or "they could have made more money without GPL", [...]
None of them sell software, they sell software tied to services.
Really, because pretty much every company I've ever worked at has profited from innovations we added to OSS code (often GPL). What you're missing is the business model of developers as a service.
No, I'm not missing it in the slightest. I'm *making* the point (which everyone going "GPL, rah, rah, rah" just quietly ignores) that you can't make money with GPLed software UNLESS you're going to tie your software to some other (non-GPLed, or equivalent) product or service (ie: "developers as a service").
The GPL is not the only way to make money creating and selling software. It is not the most profitable way all the time. To pretend it cannot be done, however, is to ignore a significant portion of today's software market.
Who is making money selling only GPLed software ? Not GPLed software tied to support contracts. Not GPLed software tied to hardware. Not dual-licensed software. Not "work for hire" that just happens to be GPLed. Just plain old GPLed software.
Yeah, if you're locked into that business model and in that particular situation, then you screwed yourself choosing to use GPL code in the first place. Why not write the other 1% yourself instead of agreeing to swap your 99% of the code for that 1% work from others?
I'm glad we agree that the GPL is not the best or only solution to software licensing. Baby steps. No we just have to convince all the other GPL zealots.
That said, I've never known anyone stupid enough to put themselves in your hypothetical situation.
Indeed, that's because they're so morally bankrupt they write closed-source code instead.
From this, we may safely draw the conclusion that Microsoft has done a lot of research, with a lot of lawyers, and they've determined that the GPL represents the biggest threat to their revenue model. And what's bad for Microsoft is generally good for everyone else. So if you're going to develop FOSS, the GPL is the obvious safe choice.
The GPL represents a threat to anyone who wants to sell software. Microsoft are hardly alone in this.
The GPL may have more restrictions than BSD, but the GPL itself doesn't take anything away [...]
Practically speaking it does, due to the ludicrously wide definition of a "derivative work".
The most obvious example of this is linking to shared libraries. If those shared libraries are GPLed - sorry, you're going to have to GPL your code as well[0]. Contrast this to the average "closed source" shared library, that requires nothing of the sort.
[0] Or maybe you won't. It hasn't actually been tested in court yet as far as I know - but it's certainly not a risk any sane person would take.
The GPL adds restrictions that protect software freedom and innovation.
The GPL makes it essentially impossible to profit from your innovations without tying your code to some other non-GPLed products (be they something as simple as a support contract or something more concrete like a piece of hardware).
*That* is why some people don't think it's the best thing since sliced bread. They want to make money selling software. Even if you have some awesome idea for a piece of software that every man and his dog will want to use, and do 99% of the heavy lifting yourself, that 1% worth of boring, mundane crap you used GPL code for essentially brings the potential price for your software crashing down to roughly $0.
I'm with FSF about this one. The "open source" term made it all less clear what this whole movement is all about.
Rrrright. But using a definition for "free" that no normal person would think of with makes it crystal clear...
Joe sixpack is not concerned what instruction set is in his DVD player or Microwave Oven.
Joe sixpack isn't trying to use his DVD player or microwave oven to browse the web, pull pictures off his digital camera or write letters to mum, either.
The UMPC is not a PC. It is a quick boot, pocket sized device that can view web pages and in a pinch edit documents. So end users don't expect to run PC Software on it.
The UMPC is a _laptop_. As such, people will expect it to be able to do everything the laptop they own today does. Which is, for all intents and purposes, run x86 code. The average customer is not the geek who wants multiple computers for toys.
No normal person wants multiple computers if they can avoid it. Multiple computers are a pain in the arse. Just like multiple mobile phones, music players, and the like.
The benchmark for good enough, is something that feels fairly snappy and is able to video cam, view web pages, play audio and full screen video with at least a resolution of 800x640. As long as people can stand the keyboard, touch pad and the quality/size of the display, it will sell.
I am curious where you're getting your idea of "good enough" from. Because by your measure "good enough" PCs were around back in about 1998, yet strangely people have been snapping up better machines since then in earnest.
The lower end. Which is what users are expecting as a computing experience. Is not moving upward any longer.
Of course it is. Has been for years. And in the next 6-12 months its going to move up to meet Vista's needs (which isn't much further anyway). What doesn't move downward often is price points - and this new US$400ish price point isn't going downwards any time soon.
Here's the crux of the issue you (and many others) seem to be having the most trouble with: hardware prices at the bottom end of the price range do not scale down in the same way they scale up (and similarly at the top end). Just because, in 6-9 months, you'll be able to buy a dual core, 1.6Ghz, 2G RAM UMPC for $400 does *NOT* mean that for $200 you'll be able to buy a machine with half the power (ie: today's Eee PC and equivalents). You might be able to get a machine half as powerful for $300 - $350, if you're lucky enough that anyone sells them - at which point buying the more powerful and more useful machine is an easy choice.
What people will not buy, is a sub $400.00 device where the Microsoft Tax is such a huge proportion of the cost it is not worth it. If there was a DVD player for $50.00 and one for $450 and the only difference is the one for $450 comes with a license for Windows Vista. Guess which one is going to sell?
The "Microsoft Tax" is a non-argument. Always has been. I don't know why anyone bothers to even bring it up. Microsoft can and will sell Vista to OEMs for whatever price is necessary. If they have to "create" a special "UMPC OEM" license at $10/copy, they will.
I am expecting that by August 2009 that "good enough" hardware will be in the $250-$275 price range with 1024x768 resolution (1280x800 when docked). This way if she loses it or it gets broke, I can buy her another one. I want a cheap throw away appliance.
Today's UMPCs aren't "good enough". They're not "good enough" because they're not capable of being the typical person's only computer.
Your average UMPC in August 2009 is going to be $350-$400, have a 9-10" screen @ 1280x720, dual cores @ 1.2-1.6Ghz, 2G (with 4G a cheap and obvious upgrade) of RAM and be sold running Vista. This is the point at which their popularity is really going to explode because this is the point at which it will be quite feasible for a UMPC to be the average person's only computer.
This trend is going to continue to drive prices down. I don't think x86 compatibility will matter. Linux fills this space just fine. For those that want a UMPC that does more. Heck just pay an extra $25. All of the Linux apps
so in your opinion it will look like this then?
No.
The $400ish price point is going to be around for a good year or two. The amount of power available at this price will increase, but the price will not decrease (at least not significantly).
This has been the consistent pattern with computer hardware for the last couple of decades. I'm not sure why anyone thinks it's suddenly going to change.
Further, a "reduced spec" machine will be, like it is today, little more than a toy. With the US (and a fair chunk of the world) heading into a recession, the market for toys is going to dry up pretty quickly.
As it turns out. That would also mean there will be a $250.00 or $300.00 machine with about the same specs as now.
What do you base this on ? Because it's certainly not the typical pattern that computing hardware pricing follows. Power at a given price point increases far, far faster - and more frequently - than price points decrease.
The thing everyone seems to be missing is this. Currently these machines don't quite have enough "oomph" to be the one machine the average person needs. They're toys for geeks.
However, with the power they'll have in 6-12 months, they will quite feasibly be the only computer the average person needs. The big disadvantages - screen and keyboard - will be nullified by a cheap docking station that people can just snap their "UMPC" into and out of in a second. It's going to be like an ipod - get home and drop it into the "dock", rip it out as you run out the door.
And most of them are going to be sold running Vista.
Microsoft can't close that door. If someone can knock the price down and improve performance even more by moving to something that is NOT x86 compatible. Microsoft is about to feel a swift kick to their family jewels.
No chance. A non-x86 machine isn't going to run the software most people want to run.
If she could see him, why would she need to use the software to take a snapshot? Don't apple computers have Prnt Scrn or an equivalent button?
If the camera display on-screen is done via an overlay in the video card, then a "capture screenshot" won't work (you'll just get a black box where the camera output would be).
No idea if that's how OS X handles it, just sayin'...
Part of that was because peoples Java experience came from Microsofts own crappy version, as another user have already pointed out.
Say what ? Microsoft's JVM shat all over Sun's until they weren't allowed to distribute it any more.
Those of us who were around then will remember that Microsoft mounted a sustained attack on Java, and deliberately crippled it's multi-platform capabilities.
And by "deliberately crippled", you mean "released the best JVM", right ?
Linux is a family of operating systems. The kernel isn't bloated, and there are Linux operating systems that require 8MB of RAM or less.
Unbelievably bloated. DOS 3.3 would run on a machine with half a megabyte of RAM.
There are no bloat-free versions of Vista. If Vista was modular and I could remove components, I would have no problem with the bloat. Alas..
Sure you would. You'd just be able to remove it. Which exposes the colossal arrogance of your position - it's only "bloat" because _you_ don't think it's a worthwhile feature. In the context of systems with equivalents features - say, OS X - Vista isn't bloated. You need similar hardware to get similar results.
You cannot easily study closed operating systems.
Sure you can. There's more than enough documentation to study the design of Windows NT in an "OS course".
Open operating systems are meant to be studied. Since Windows is a moving target, it's even more difficult to come up with accurate information on it.
Windows is no more a "moving target" than Linux is. Less, if anything - its internals have changed far less over the last ~15 years than Linux's has.
The Windows design concerns nobody but Microsoft.
The Windows design should concern anyone who purports to have an interest in Operating Systems.
They're pretty damned close -- even though the iMac had a faster processor, the PC had 3x the RAM.
The iMac has a 20% clock speed advantage. In and of itself, that's not trivial. To say nothing of it having a CPU that noticably faster per clock because of the larger L2 cache (6M vs 2M) and higher bus (1333Mhz vs 800Mhz). Finally, with an odd amount of RAM like 3G, I'd be willing to bet the Gateway also has unmatched DIMMs, meaning it's not using interleaved memory, further hurting performance.
Regardless of OS, one would expect the iMac to be ~20-30% faster than the PC (which is basically what the Vista benchmarks demonstrate).
I am also somewhat sceptical of Geekbench's relevance as a cross-platform benchmark. For example, one would expect the solely-CPU benchmarks (Integer and Floating Point) to be essentially identical on both OS X and Vista. Instead we see OS X is significantly slower in integer and significantly faster in floating-point. Similarly for the "graphics" benchmark (indicating possible driver discrepancies).
It would be interesting to compare Vista and XP numbers, however, I suspect these results say more about Geekbench and its usefulness as a cross-platform benchmark, than they do about how Vista and OS X compare.
(These are issues that call into question any conclusions that are apparent from just a cursory look at the parameters. I'm sure there are others that would appear on deeper analysis. For example, did they give Vista's initial disk index the chance to finish before running benchmarks ?)
I'm sorry but how can anyone seriously argue that "our program is bloated, but it's OK because relatively decent hardware can handle it!" with a straight face.
Because the only logical conclusion to the opposing argument is an abacus.
If you have to argue that your bloat is acceptable, then it's too bloated.
People who can still do their work on a DOS 3.3 machine would argue Linux is incredibly bloated. Would you agree with them ?
I've taken multiple OS courses, and none of them discussed Windows directly. I never did Windows development. I think you would find that many CS students simply don't care about Windows, and you 'expect [me] to be well aware of' it.
I'm incredulous any university-level OS "course" wouldn't cover Windows NT, probably the next most important OS in the modern world outside of UNIX (and derivatives).
Any "CS student" who "simply doesn't care about Windows" should be ashamed to call themselves such. Mindless bias like that has no place in academia.
These machines will never run Vista well. Let's keep that important knowledge in front of people. Intel expects to move 10 million Atom platforms in the first wave, and none will have Vista.
What ? A dual core machine with 2G of RAM will run Vista fine. *Most* of them will be sold running Vista.