it seems MS has even managed to persuade many Slashdotters that they are going to use open formats. Poor fools.
Yeah, in less than five minutes on this thread I've seen the terms "rich API" and "framework", two of the biggest Microsoft-parrot terms of the last few years. BTW, I think it's safe to say that the word "framework" is on its way to be the official buzzword of 2003.
...And don't forget Windows 2008, the Newest, most Open Windows ever! With TurboKernel (*BSD) at its core, you can be assured of full interoperability with Unix code worldwide...
Hey, they've copied Apple so far, I pretty much expect Windows to be Unix-based before the decade is up...
Can Beowulf replace traditional parallel apps?
on
Ask Donald Becker
·
· Score: 1
Since MPI pretty much forces one to pass data through little tunnels between processes, couldn't this be a widespread method for parallel processing even on single-CPU systems or on SMP systems? It seems like the biggest problem for the typical developer doing multithreaded apps is when they start handling memory that another thread is holding but forgot to lock. Since each "thread" would be a separate program/process using MPI, this would be a lot harder to do. Also, on big iron multi-processor systems running nodes as virtual machines, using MPI would give side benefits more easily, such as being able to add/remove nodes on the fly. Of course, the MPI overhead may be too much of a drain to make it worthwhile. What do you think?
I don't know about your area, but where I live there are several thousand dollars of property taxes per year. Secondly, houses don't maintain themselves, so unless you got a comprehensive new home warranty, add in the invariable several thousand in unexpected furance fans, shingles, eavesdroughs, etc
Well, if it's a relatively new home, most of this would be covered under the manufacturers' warranty for each appliance/hardware, and you can get a home warranty for very little ($400/year?) if it's older than 10 years old.
Thirdly, you compare a $200K house with a $1500 apartment: In areas where apartments are $1500, houses are dramatically more than $200K (at least where I live).
Well, let's compare apples to apples (which is what I was trying to do). In my area (Southeast USA) a new $200K home sits on about.25 -.3 acres, is two floors, and about 2000 square feet with a two car garage. Sure, I can rent a nice 2 bedroom apt for about $800/month or a 3 bedroom for about $1100-1200/mth, but those are usually about 700 and 1100 square feet each, respectively. Worst of all, you can hear your neighbors peeing in the bathroom, listening to the stereo in the living room, and making whoopee in the bedroom. That ain't living, and they'll have to drag me kicking and screaming back to that world...:-) To rent a comparable house would cost at least $1500; and probably more like $1900/mth, assuming I even found one.
I'm not anti-home ownership, but a lot of people really create an illusion around home ownership.
True, it's not for everyone at every point in their lives. Obviously a 22 year old single guy who's sowing his wild oats doesn't need the burden of a house, but the hard numbers say that everyone at some point should work toward a station in life where a home is feasible.
if you bought a house in the last 2 years, you're going to look worse than this guy after the bubble bursts in the housing market.
Depends on your region, and even then, housing bubbles don't really exist unless you bought a $10 million dollar mansion with tacky customized features that nobody wants, much less can afford. If an average house in SF costs $400K, it's not going to drop by very much even if a bubble bursts. Maybe down as far as $375K briefly, but in the end you still have the sticks and bricks of the house and the scarce land it's on, so the value will tend to remain. Stocks don't have a fixed scarcity, so they can evaporate. In short, if you stay around the middle of the road, you won't lose on a house even if you sell within a few years.
houses arent that great of investments, and unless you are sure you are going to be in it for 5-10 years, you will get screwed.
Let's say I buy a house for $200K in 2000 and sell it for the same $200K in 3 years. Assuming 6% in realtor costs ($12K) plus 2% misc costs ($4k), I've gotten "screwed" by 16K in closing costs which they will deduct from the selling price, so the bank will only get $184K, of which I've only paid down to $190K, so I owe them 6K (total, $22K). For a 30 year note at 6.75% plus some taxes and insurance my payments would be around $1600/mth. For the first three years, about $1000/mth of this would be interest, which can be deducted from fed taxes at probably the 27% bracket = $270/mth * 36mths = $9720. So, our $22K loss is now $22000 - 9720 = $12280. This averages a net outlay of $341 over the 36 months to live in a home of your own. If that is getting screwed, sign me up...:-) If you want to nitpick, we'll add in the buying closing costs of about 2% ($4K), which adds $111/mth to the "rent".
Now, to compare, let's rent a modest house or large apartment, which would cost around $1500/mth * 36mths = $54,000. None of this is deductable, and you get nothing back when you move out.
True, renting is nice when you are unsure of where you'll be next year and don't want to risk holding two house notes at a time, but according to a book I read (Home Buying for Dummies, I think. Don't laugh, it's a really good book!) homeowners on average have 20-40 times the net worth of lifetime renters.
Sounds like your real beef is with.Net. Don't forget,.Net != web services. There are many of us developing web services with no intent to use.Net in any way.
I wasn't picking on.Net alone; I'm referring more to its target audience; those who have never coded plain old HTTP GET/POST calls to see just how easy it is to pass a set of parameters to a remote function (www.myurl.com/services/currentTemperature) and get return data back, whether it be XML, HTML, formatted ASCII, or raw binary. These people have been convinced that you *need* a huge runtime library and VM to accomplish web services, when you could do it with a couple of perl scripts or C code.
And unfortunately, speech recognition and AI (which might be the same) are probably the most important for making computers truly useful for the ordinary end users that don't have to time to learn complex interfaces.
Well, that's the problem, isn't it? I agree with this when the interface is overly complex or cumbersome, but if a person doesn't want to learn the concepts of a technology, job, or process, then a talking computer won't help them anymore than a teacher or helpful co-worker. Better to just make the computer a little bit smarter and leave this non-learning user out of the loop altogether, IMO. For the classic consumer uses of computers like kiosks and teller machines, touchscreens work just a well, keep information more private, and cost less.
This notion of the talking computer making everything possible reminds me of the notion of self-programming computers. The problem is this: most users can't describe their software needs to human analysts and developers. Why should they be able to describe ("program") the software to a machine any better? Or, to put it another way, who cares if the Enterprise mainframe can increase your Tachyon emmission field by 75% if you have know idea what a Tachyon field is, or how increasing it 75% can save you from the enemy ship (the computer said it was a "Klingon". What's that?).
...is those pesky users and their fickle minds. For instance, who would have thought that most people actually don't *want* video phones or flying cars or talking computers? Or at least, they don't want them enough to drive the technical development of these things, since standard phones, autos, and Windows seem to do the job well enough.
In other words, just because a technology looks like it's the "right" way to progress next, doesn't mean the market will allow it to move along.
I think we'll see this with Web Services (noted in the artcle as the current Next Big Thing). At it's core it's simply a formalization of how CGI developers have been working for years, yet most people and developers still prefer to use a generic web browser to diseminate most information, vs. using a custom client and a web service. Why? Because developers don't want to support another client program, and users don't want to download another one when they can just enter www.weather.com/my-zip-code to get the current weather forecast. I don't think it's been the lack of a formal parameter/return value standard that has held this idea back.
Don't get me wrong, I think Web Services are a nice tool, but unfortunately I see it as a problem looking for a solution. For most end-users it will mostly be a poor substitute for a URL (wait until your co-worker comes in to show you his spiffy new.Net web services demo! It will show you the current news and weather! OVER THE INTERNET! Oh, just install this 300MB library+runtime first. Ok, now install my 30MB client app. Oh, yeah, that didn't refresh properly, did it? Exit out and restart. Dang. [this is better than a browser how?]), and for most in-house developers it will be just another call to use instead of dlopen() to open shared routine. And until the Net becomes totally ubiquitous and telecom-reliable, I don't see many shrink-wrap developers linking in lots of remote Web Services on the fly, when most of that functionality can be placed locally during the install.
Seems like they could make a decent living by selling an enhanced, secured version and then have the "last version" free as in beer & speech to help spread the product. Similar to how Aladdin has done with Ghostscript.
Re:Console key(s) instead of root password?
on
AOL's new Linux PC
·
· Score: 1
Couldn't Lindows (or other local, single-user-oriented distros) just assign a special console key to count for root access? For instance, we allow a reboot with CTRL-ALT-DEL because we only honor that key sequence from the local hardware and not from some network-borne process. Ditto for CTRL-ALT-BACKSPACE to kill X. Why can't we have a CTRL-ALT-F1 or similar to allow root access when requested? You click to install a package or configure the system, and a window pops up with a picture of the keyboard and says to hit these keys to continue. Or, have the magic keys only enabled on demand, and you could make the magic key *any* key, so long as it came from the local console.
I haven't thought this through enough to figure out the details, but it seems like it could work.
Now, if we can get BusinessWeek or the Wall Street Journal to start saying things like "...businesses should begin investigating Linux to remain cost competitive", the C*O's in America would start herding over to Linux (regardless of its merits and/or limitations, but that's another story).
Indeed, saying that the Konquerer team fixed it in 90 minutes makes them sound very irresponsible, not proactive...
I agree with you about small changes creating bugs elsewhere in the system, but the OSS philosophy is "Release Early, Release Often", and that methodology has held up pretty well over the past several years.
Right now, lots of pretty sophisticated users and developers are testing Konqeror from the CVS "alpha". Behind the closed doors of Microsoft, who is testing the one-line fix for IE? a) Whatever developers/QA staff happen to be between projects, and b) the hotshots that are now being pulled (grudgingly) off their current project to work on this bug.
Battery life has to improve drastically (currently about 2 hours in developer's model)
I have the 5000D model, too. Turn the backlight down to 50% and the battery will last a lot longer; of course, you have to get used to the lower light conditions.
Supposedly, this is only a problem with the 5000D. The 5500 fully charges the battery, whereas the 5000D only gets to about 50%. I read this on one of the zaurus sites... don't ask me for the URL, tho... it's long gone.:)
...I never seen an atheist wear a piece of clothing to proclaim to the world their religion...
Counterpoint: ever see the little "Darwin" car emblems, spoofing the christian fish? Not only is it proudly espousing the (non)beliefs of the driver, but doing so with a stinging parody of the latter.
My home box (Linux) appears to be faster than my Windows box at work because (a) I can run a lightweight windowmanager (Fvwm2) and (b) the virtual memory system actually works; i.e. I'm not swapping when I have 200MB free, unlike in NT/2k, which both appear to constantly swap when plenty of memory is available. One can can also get speed improvements by compiling their own kernel.
I also recall some benchmarks (warning!) about a year ago that showed some surprising speed differences between Linux and Windows when doing CPU-intensive work. I think that article was in LJ, as well.
However, if they are learning the "Business" side of the systems, they will be sorely lacking in how to run MS based Windows applications, which apparently is what 90% of the computers in the world run.
I would argue that if you teach someone the principles of word processing, spreadsheets, and presentations using almost any office suite, they will be right at home in any other office suite. I know, because I have used many different office suites over my career (never just one exclusively) and I find that I am often giving Word/Excel advice to those who have just learned Office by rote.
Given that some states are beginning to lay off teachers from lack of funds, it is foolish to waste dollars on MS Office when OpenOffice or StarOffice will do just fine. Plus, the kids can take their work home without having to convert docs or shell out hundreds to buy Office at home.
Woah, now if Microsoft did something like that, Mac users would be screaming bloody murder!
You are correct, and if Apple controlled 90%+ of multiple markets we would scream at them as well. There has been and will be different standards of conduct between smaller companies and large monopolies.
Personally, I think it is unwise to cancel the Windows version. They should offer a low-end Windows version and offer the "premium" features only on the Unix versions to pull more users to their platform.
So, does this mean that after 2003 they are planning on killing the product for Irix and Linux...
Hopefully they will continue to support both, since OS-X is also a Unix variant. Even if Linux represents a competitive threat to them, it's better to have Linux succeed (and thus have lots of Linux renderfarms) and then Apple can work on getting people to migrate to OSX from Linux in the years to follow. It's a whole lot easer to migrate from Linux -> OSX than Windows -> OSX.
Story about building a PC using an old shoe box, chicken wire, and old Mac SE parts: $25M
Story about same machine running Linux: $50M
Story (with link) to same machine running a web server and powered by potato energy: $150M
Obligatory flaming of JonKatz techno rant: $300M
Story Proving/Disproving Evolution, Story about paid Microsoft benchmarks, Story with anything about genetics: 8x the annual GDP of small European country.
[Having no karma and learning to love it: Priceless]
A lot of people have said similar things about Microsoft's initial failures, only to be proven very wrong after a year or two once Microsoft has won the war.
Partly agree, but unlike software, which has nearly zero-cost of replication per unit (after development), with X-Box this is the first endeavor in which they are actually losing lots of money with each unit. I wonder how well that will sit with management, who are used to "taking losses" at only a few bucks per unit with software CDs.
Which system comes built-in with next to unlimited storage for saved games (hard drive)? Xbox
I think that H/D is going to prove to be a pain for MS in the long run. Think about it: the PS2 is essentially a much faster PS1, but as far as manufacturing complexity, it's about the same (some chips, I/O ports, CD/DVD drive). That means that in a couple of years Sony can drop the PS2 down to $129 PS1 levels and not completely lose their shirts. The extra H/D will make that price point much harder for the XBox, especially considering they're probably losing about $75-100 USD per unit as it is.
If you were to replace "Playstation 2" with "Windows", and replace "XBox" with "Linux", this thread would have 1200 comments all stating how just because Windows sells more doesn't mean it's better. But, when MS is the underdog,/. tastes blood in the water and moves in for the kill. Let's be objective!
Considering that it's human nature to see a bully get a black eye once in a while, I'd say these threads are already fairly objective today.
I, for one, will win whether XBox succeeds or fails, even though I won't buy one. If it succeeds, then the sub-$350 PC system will come along that more quickly, and everybody will win when that happens, except maybe the Windows tax. If it fails, then that's several more billion USD Microsoft doesn't have in its coffers to throw around.
Slashdot Mirror
it seems MS has even managed to persuade many Slashdotters that they are going to use open formats. Poor fools.
Yeah, in less than five minutes on this thread I've seen the terms "rich API" and "framework", two of the biggest Microsoft-parrot terms of the last few years. BTW, I think it's safe to say that the word "framework" is on its way to be the official buzzword of 2003.
...And don't forget Windows 2008, the Newest, most Open Windows ever! With TurboKernel (*BSD) at its core, you can be assured of full interoperability with Unix code worldwide...
Hey, they've copied Apple so far, I pretty much expect Windows to be Unix-based before the decade is up...
Since MPI pretty much forces one to pass data through little tunnels between processes, couldn't this be a widespread method for parallel processing even on single-CPU systems or on SMP systems? It seems like the biggest problem for the typical developer doing multithreaded apps is when they start handling memory that another thread is holding but forgot to lock. Since each "thread" would be a separate program/process using MPI, this would be a lot harder to do. Also, on big iron multi-processor systems running nodes as virtual machines, using MPI would give side benefits more easily, such as being able to add/remove nodes on the fly. Of course, the MPI overhead may be too much of a drain to make it worthwhile. What do you think?
I don't know about your area, but where I live there are several thousand dollars of property taxes per year. Secondly, houses don't maintain themselves, so unless you got a comprehensive new home warranty, add in the invariable several thousand in unexpected furance fans, shingles, eavesdroughs, etc
.25 - .3 acres, is two floors, and about 2000 square feet with a two car garage. Sure, I can rent a nice 2 bedroom apt for about $800/month or a 3 bedroom for about $1100-1200/mth, but those are usually about 700 and 1100 square feet each, respectively. Worst of all, you can hear your neighbors peeing in the bathroom, listening to the stereo in the living room, and making whoopee in the bedroom. That ain't living, and they'll have to drag me kicking and screaming back to that world... :-) To rent a comparable house would cost at least $1500; and probably more like $1900/mth, assuming I even found one.
Well, if it's a relatively new home, most of this would be covered under the manufacturers' warranty for each appliance/hardware, and you can get a home warranty for very little ($400/year?) if it's older than 10 years old.
Thirdly, you compare a $200K house with a $1500 apartment: In areas where apartments are $1500, houses are dramatically more than $200K (at least where I live).
Well, let's compare apples to apples (which is what I was trying to do). In my area (Southeast USA) a new $200K home sits on about
I'm not anti-home ownership, but a lot of people really create an illusion around home ownership.
True, it's not for everyone at every point in their lives. Obviously a 22 year old single guy who's sowing his wild oats doesn't need the burden of a house, but the hard numbers say that everyone at some point should work toward a station in life where a home is feasible.
if you bought a house in the last 2 years, you're going to look worse than this guy after the bubble bursts in the housing market.
:-) If you want to nitpick, we'll add in the buying closing costs of about 2% ($4K), which adds $111 /mth to the "rent".
Depends on your region, and even then, housing bubbles don't really exist unless you bought a $10 million dollar mansion with tacky customized features that nobody wants, much less can afford. If an average house in SF costs $400K, it's not going to drop by very much even if a bubble bursts. Maybe down as far as $375K briefly, but in the end you still have the sticks and bricks of the house and the scarce land it's on, so the value will tend to remain. Stocks don't have a fixed scarcity, so they can evaporate. In short, if you stay around the middle of the road, you won't lose on a house even if you sell within a few years.
houses arent that great of investments, and unless you are sure you are going to be in it for 5-10 years, you will get screwed.
Let's say I buy a house for $200K in 2000 and sell it for the same $200K in 3 years. Assuming 6% in realtor costs ($12K) plus 2% misc costs ($4k), I've gotten "screwed" by 16K in closing costs which they will deduct from the selling price, so the bank will only get $184K, of which I've only paid down to $190K, so I owe them 6K (total, $22K). For a 30 year note at 6.75% plus some taxes and insurance my payments would be around $1600/mth. For the first three years, about $1000/mth of this would be interest, which can be deducted from fed taxes at probably the 27% bracket = $270/mth * 36mths = $9720. So, our $22K loss is now $22000 - 9720 = $12280. This averages a net outlay of $341 over the 36 months to live in a home of your own. If that is getting screwed, sign me up...
Now, to compare, let's rent a modest house or large apartment, which would cost around $1500/mth * 36mths = $54,000. None of this is deductable, and you get nothing back when you move out.
True, renting is nice when you are unsure of where you'll be next year and don't want to risk holding two house notes at a time, but according to a book I read (Home Buying for Dummies, I think. Don't laugh, it's a really good book!) homeowners on average have 20-40 times the net worth of lifetime renters.
Maybe some of the editors should write their submissions in MS Word or one of its equivalents before posting it to the masses.
And then it would be ?major?
Sounds like your real beef is with .Net. Don't forget, .Net != web services. There are many of us developing web services with no intent to use .Net in any way.
.Net alone; I'm referring more to its target audience; those who have never coded plain old HTTP GET/POST calls to see just how easy it is to pass a set of parameters to a remote function (www.myurl.com/services/currentTemperature) and get return data back, whether it be XML, HTML, formatted ASCII, or raw binary. These people have been convinced that you *need* a huge runtime library and VM to accomplish web services, when you could do it with a couple of perl scripts or C code.
I wasn't picking on
And unfortunately, speech recognition and AI (which might be the same) are probably the most important for making computers truly useful for the ordinary end users that don't have to time to learn complex interfaces.
Well, that's the problem, isn't it? I agree with this when the interface is overly complex or cumbersome, but if a person doesn't want to learn the concepts of a technology, job, or process, then a talking computer won't help them anymore than a teacher or helpful co-worker. Better to just make the computer a little bit smarter and leave this non-learning user out of the loop altogether, IMO. For the classic consumer uses of computers like kiosks and teller machines, touchscreens work just a well, keep information more private, and cost less.
This notion of the talking computer making everything possible reminds me of the notion of self-programming computers. The problem is this: most users can't describe their software needs to human analysts and developers. Why should they be able to describe ("program") the software to a machine any better? Or, to put it another way, who cares if the Enterprise mainframe can increase your Tachyon emmission field by 75% if you have know idea what a Tachyon field is, or how increasing it 75% can save you from the enemy ship (the computer said it was a "Klingon". What's that?).
...is those pesky users and their fickle minds. For instance, who would have thought that most people actually don't *want* video phones or flying cars or talking computers? Or at least, they don't want them enough to drive the technical development of these things, since standard phones, autos, and Windows seem to do the job well enough.
.Net web services demo! It will show you the current news and weather! OVER THE INTERNET! Oh, just install this 300MB library+runtime first. Ok, now install my 30MB client app. Oh, yeah, that didn't refresh properly, did it? Exit out and restart. Dang. [this is better than a browser how?]), and for most in-house developers it will be just another call to use instead of dlopen() to open shared routine. And until the Net becomes totally ubiquitous and telecom-reliable, I don't see many shrink-wrap developers linking in lots of remote Web Services on the fly, when most of that functionality can be placed locally during the install.
In other words, just because a technology looks like it's the "right" way to progress next, doesn't mean the market will allow it to move along.
I think we'll see this with Web Services (noted in the artcle as the current Next Big Thing). At it's core it's simply a formalization of how CGI developers have been working for years, yet most people and developers still prefer to use a generic web browser to diseminate most information, vs. using a custom client and a web service. Why? Because developers don't want to support another client program, and users don't want to download another one when they can just enter www.weather.com/my-zip-code to get the current weather forecast. I don't think it's been the lack of a formal parameter/return value standard that has held this idea back.
Don't get me wrong, I think Web Services are a nice tool, but unfortunately I see it as a problem looking for a solution. For most end-users it will mostly be a poor substitute for a URL (wait until your co-worker comes in to show you his spiffy new
Seems like they could make a decent living by selling an enhanced, secured version and then have the "last version" free as in beer & speech to help spread the product. Similar to how Aladdin has done with Ghostscript.
Couldn't Lindows (or other local, single-user-oriented distros) just assign a special console key to count for root access? For instance, we allow a reboot with CTRL-ALT-DEL because we only honor that key sequence from the local hardware and not from some network-borne process. Ditto for CTRL-ALT-BACKSPACE to kill X. Why can't we have a CTRL-ALT-F1 or similar to allow root access when requested? You click to install a package or configure the system, and a window pops up with a picture of the keyboard and says to hit these keys to continue. Or, have the magic keys only enabled on demand, and you could make the magic key *any* key, so long as it came from the local console.
I haven't thought this through enough to figure out the details, but it seems like it could work.
Now, if we can get BusinessWeek or the Wall Street Journal to start saying things like "...businesses should begin investigating Linux to remain cost competitive", the C*O's in America would start herding over to Linux (regardless of its merits and/or limitations, but that's another story).
Indeed, saying that the Konquerer team fixed it in 90 minutes makes them sound very irresponsible, not proactive...
I agree with you about small changes creating bugs elsewhere in the system, but the OSS philosophy is "Release Early, Release Often", and that methodology has held up pretty well over the past several years.
Right now, lots of pretty sophisticated users and developers are testing Konqeror from the CVS "alpha". Behind the closed doors of Microsoft, who is testing the one-line fix for IE? a) Whatever developers/QA staff happen to be between projects, and b) the hotshots that are now being pulled (grudgingly) off their current project to work on this bug.
Battery life has to improve drastically (currently about 2 hours in developer's model)
:)
I have the 5000D model, too. Turn the backlight down to 50% and the battery will last a lot longer; of course, you have to get used to the lower light conditions.
Supposedly, this is only a problem with the 5000D. The 5500 fully charges the battery, whereas the 5000D only gets to about 50%. I read this on one of the zaurus sites... don't ask me for the URL, tho... it's long gone.
...I never seen an atheist wear a piece of clothing to proclaim to the world their religion...
Counterpoint: ever see the little "Darwin" car emblems, spoofing the christian fish? Not only is it proudly espousing the (non)beliefs of the driver, but doing so with a stinging parody of the latter.
My home box (Linux) appears to be faster than my Windows box at work because (a) I can run a lightweight windowmanager (Fvwm2) and (b) the virtual memory system actually works; i.e. I'm not swapping when I have 200MB free, unlike in NT/2k, which both appear to constantly swap when plenty of memory is available. One can can also get speed improvements by compiling their own kernel.
I also recall some benchmarks (warning!) about a year ago that showed some surprising speed differences between Linux and Windows when doing CPU-intensive work. I think that article was in LJ, as well.
Anyone know if Doom III will be driven by OpenGL? I didn't see mention of it in the article(s).
However, if they are learning the "Business" side of the systems, they will be sorely lacking in how to run MS based Windows applications, which apparently is what 90% of the computers in the world run.
I would argue that if you teach someone the principles of word processing, spreadsheets, and presentations using almost any office suite, they will be right at home in any other office suite. I know, because I have used many different office suites over my career (never just one exclusively) and I find that I am often giving Word/Excel advice to those who have just learned Office by rote.
Given that some states are beginning to lay off teachers from lack of funds, it is foolish to waste dollars on MS Office when OpenOffice or StarOffice will do just fine. Plus, the kids can take their work home without having to convert docs or shell out hundreds to buy Office at home.
Woah, now if Microsoft did something like that, Mac users would be screaming bloody murder!
You are correct, and if Apple controlled 90%+ of multiple markets we would scream at them as well. There has been and will be different standards of conduct between smaller companies and large monopolies.
Personally, I think it is unwise to cancel the Windows version. They should offer a low-end Windows version and offer the "premium" features only on the Unix versions to pull more users to their platform.
So, does this mean that after 2003 they are planning on killing the product for Irix and Linux...
Hopefully they will continue to support both, since OS-X is also a Unix variant. Even if Linux represents a competitive threat to them, it's better to have Linux succeed (and thus have lots of Linux renderfarms) and then Apple can work on getting people to migrate to OSX from Linux in the years to follow. It's a whole lot easer to migrate from Linux -> OSX than Windows -> OSX.
- Story about building a PC using an old shoe box, chicken wire, and old Mac SE parts: $25M
- Story about same machine running Linux: $50M
- Story (with link) to same machine running a web server and powered by potato energy: $150M
- Obligatory flaming of JonKatz techno rant: $300M
- Story Proving/Disproving Evolution, Story about paid Microsoft benchmarks, Story with anything about genetics: 8x the annual GDP of small European country.
[Having no karma and learning to love it: Priceless]A lot of people have said similar things about Microsoft's initial failures, only to be proven very wrong after a year or two once Microsoft has won the war.
Partly agree, but unlike software, which has nearly zero-cost of replication per unit (after development), with X-Box this is the first endeavor in which they are actually losing lots of money with each unit. I wonder how well that will sit with management, who are used to "taking losses" at only a few bucks per unit with software CDs.
Which system comes built-in with next to unlimited storage for saved games (hard drive)? Xbox
I think that H/D is going to prove to be a pain for MS in the long run. Think about it: the PS2 is essentially a much faster PS1, but as far as manufacturing complexity, it's about the same (some chips, I/O ports, CD/DVD drive). That means that in a couple of years Sony can drop the PS2 down to $129 PS1 levels and not completely lose their shirts. The extra H/D will make that price point much harder for the XBox, especially considering they're probably losing about $75-100 USD per unit as it is.
If you were to replace "Playstation 2" with "Windows", and replace "XBox" with "Linux", this thread would have 1200 comments all stating how just because Windows sells more doesn't mean it's better. But, when MS is the underdog, /. tastes blood in the water and moves in for the kill. Let's be objective!
Considering that it's human nature to see a bully get a black eye once in a while, I'd say these threads are already fairly objective today.
I, for one, will win whether XBox succeeds or fails, even though I won't buy one. If it succeeds, then the sub-$350 PC system will come along that more quickly, and everybody will win when that happens, except maybe the Windows tax. If it fails, then that's several more billion USD Microsoft doesn't have in its coffers to throw around.