I really think that I should write a filter that spell-checks an email, and rejects it if over 50% of the words with 5 or more letters are misspelled. >>>>>> That would rock. If such a filter was on Slashdot, Slashdot's post volume would drop by 90%:)
Not really. In the mainstream market, people don't shop for software on CheapBytes. Hell, most people pay 2x markup at Comp "How May We Ream You Today?" USA.
Still not right. You can sell it all you want, and you don't have to tell them where they can get it free. All you have to do is provide source to anybody who bought it.
Actually, I think there are serious differences between the DRI model and what NVIDIA's hardware dose. Recently, they even switched away from XAA to their own model because XAA apperently didn't mesh well with how their cards did mixed 2D/3D rendering.
Don't look at me. I live in Virginia, and most definately did *not* vote for him! And yes, I did vote. In Virginia, my democratic vote makes not one iota of difference, but I suppose it has ceremonial meaning.
What kind of idea is that? Blowing up computers? He's joking...right? Its a violent and incredibly base thing to say. How do these people make it into Congress?
Wait a minute... so the minute Microsoft decides to compete in a market they've crossed the line? >>>>>> Try reading what I said. I said there are three things they need to do to cross the line. Until now, they were only doing the first two. It's only when they started doing the third thing that they crossed the line. MS can compete in search engine space as much as it wants. Just don't do it by leveraging integration with Windows.
Gee, the X-Box hasn't been a smash success, I wonder why that is? >>>>>>> The X-Box hasn't been a success because its not the best product. Microsoft can't leverage Windows with the XBox, and thus they're competing fairly in that market. They're also getting their ass kicked --- Sony has twice as much market share as Microsoft and Nintendo combined.
Likewise, they're going to compete in a market dominated by Google and Yahoo and you think that by virtue of them having a monopoly SOMEWHERE that they're going to rush in and crush everyone? >>>>>>>>>> This isn't an unrelated monopoly. Unlike the XBox case, Microsoft can very easily use its domination of the OS market to try and dominate the search engine market. It wouldn't be very hard to do at all. With WinFS coming, I can easily envision a Windows search tool that unified internet and local file searching, tieing into MSN Search in the process. This would be illegal, because they'd be using their monopoly control of Windows to lock Google out of the market.
Your description hasn't been the case for awhile now. Almost all the chipsets these days use a proprietory link between the southbridge and the northbridge. For example, in NVIDIA's nForce chipsets, there is a 800MB/sec hypertransport link between the MCP (southbridge) and the IGP (northbridge). SiS uses something called MuTIOL, while Via uses V-Link.
You're mixing gigabytes and gigabits. PCI Express is 2.5 gigabits per second in each direction, or about 250MB/sec per pin. AGP 8x is 2.1 gigabytes per second, or 2100 gigabits in total.
1) They're a monopoly 2) They're leveraging their monopoly 3) They're doing (1) and (2) to try to dominate another market.
Until now, Microsoft hadn't been doing 3. MSN search was just a requisite part of any portal site. Now, they're declaring that they've actively decided to actively compete in that market. This is where they cross the line.
Yes, I do hate Microsoft and want them to fail. They've done so much damage to the software industry with their monopoly that it would require a revolution within the company to redeem them.
However, in this case, my personal opinion of Microsoft has nothing to do with it. The fact that Windows is a monopoly is not in question. The fact that they are leveraging Windows in this case (because Internet Explorer is a part of Windows now, isn't it?) seems very clear. It is illegal for a company to use its monopoly power to try and dominate an unrelated market. It's as simple as that.
PS> Before you defend Microsoft by saying that all businesses are dishonest and compete unfairly, think about one thing: the computing industry is ridiculously broken. Microsoft products dominate pretty much every major software catagory. They've got the major operating system (Windows), the major office suite (Office), the major web browser (Explorer), the major development environment (Visual Studio), etc. This situation is positively unhealthy for the industry. Compare this to other industries. Ford and GM are both huge companies, but they face intense competition in their market. Sony is a giant corporation, but RCA and Philips are doing just fine in the electronics industry, just as lots of other companies are doing fine in the media industry. These days, even Intel has to stay wary of AMD. Competition is good: the intense competition between Intel and AMD in the last few years has spurred some of the fastest growth in processor technology in recent memory. Just by simply comparing the computer industry to others you can easily see that there is a problem, and Microsoft is that problem!
They try to compete while having an unfair advantage. The government clearly demonstrated that they are an illegal monopoly, but wussed out when it came to punishing them for it. Justice must come from somewhere, and sometimes, it must come from the people themselves. Think of this as a posse rounding up to do what the government is too weak to.
Yes. Linux is constantly evolving and improving. The improvements come from two places: first, lots of people have access to the kernel source, and they get new ideas about how to make things better. Second, and more importantly, Linux is being put to an increasingly large range of uses. Kernel 2.2 makes for a fine server for a small 1 or 2 CPU machine with not too much memory, but makes a rather poor destkop. With all the improvements that have happened over the past year, kernel 2.5 makes for a damn good desktop or media workstation machine, in addition to handling its previous tasks without too much increased overhead. In the 2.7 kernel series, we're going to see major improvements to the NUMA architecture, so kernel 2.7 is going to be great for large-scale NUMA machines. Windows, on the other hand, is rather limited in what it can do. It doesn't handle real small platforms very well, or big NUMA setups. Even on the desktop, its VM (which desperately needs a rewrite) can cause problems.
The performance hit really matters on what you're doing. First, having attributes adds no performance hit --- XFS and ext3 already do it. Second, the indexing overhead really only comes into play for certain cases. Mainly, there is a performance impact on benchmarks that modify file metadata. Operations like file reads and writes shouldn't be affected. For the types of files that users really want to maintain in the index (documents, media files, etc) metadata updates don't happen often enough for the indexing mechanism to become a bottleneck. On the other hand, for something like a Squid cache, where file create/delete performance is critical, you don't really care if those files are in the index, because the user won't be searching for them. If indexing is applied on a per-attribute basis (and the user takes care not to tag stuff like cache files with attributes that will be indexed) the performance hit of indexing should be minimal.
Actually, in this case, Linux might very well do it first. Reiser4 has all the bits and pieces in there to be a killer DBFS back-end. They're trying to get good file performance out of objects as small as individual XML elements. Throw on a database layer on top (doesn't have to be that featureful for something like this) and you've got yourself a WinFS competitor. I think MS might have done themselves a disservice by throwing these ideas out there this early. Longhorn isn't due out for almost two years. In that time, it may very well be that many of the features touted for Longhorn have been in Linux for months. Remember, this time frame we're talking about is about how long it took the KDE guys to totally rewrite KDE in the 1.0 to 2.0 transition. In the last year, Linux has had several major subsystems totally overhauled. In a year and a half, Linux changes as much as Windows does in a full release cycle (2-3 years). I would not be surprised if Longhorn has some serious competition when it debuts.
Not really. A database FS is one of the few new advances that add real power to systems. If you think about the human thought process, they don't tend to think of things in terms of strict hierarchies. They tend to look at a set of items from multiple viewpoints, possibly creating a hierarchy within a given subset, but rarely subject all items to the same hierarchy. There are lots of places where a hierarchical filesystem doesn't really make sense. Is there any sane method of organizing a documents directory? Whenever I try to do it, I generate huge, complex hierarchies with only a few files in each directory. Organizationally, its elegant, but its a pain to traverse. Instead, if I simply shove all documents into a single store, I can use the database mechanism to view the file set in a way that is most useful for the task at hand. A media file directory is another good example. Its tedious to organize media files into a strict hierarchy, and very limiting to have to deal with that hierarchy every time, especially if you want a different view of the data. For example, usually, I organize music files by artist, then album. But what if I want to quickly access files based on catagory (blues) or period (80's)? Since the artist trait is usually orthogonal to the catagory and period traits, a simple hierarchy fails to adequately describe the situation. This is why jukeboxes like iTunes or MusicMatch are getting so popular: they are very flexible with how songs are presented to the user. The whole idea here is to take the power and efficiency of something like the iTunes interface, and apply it to file management in general. If a simple query language is supported, such a scheme could potentially be very usable via a command line as well.
First, XP, without the SQL or CLR is still not a microkernel. It's less of a microkernel than Linux (which isn't a microkernel) because it runs more stuff in kernel space. The "rumor" is just a bit of additional information that shows that MS is adding even more cruft to the kernel.
OS X uses something that used to be a microkernel, but its been totally changed. The BSD server runs entirely in kernel mode, and message passing between Mach and the BSD server has been replaced by function calls.
NT is hardly a microkernel. A microkernel, according to strict definitions, doesn't include anything like drivers, paging or the filesystem. QNX fits this definition --- the filesystem runs in userspace, and even drivers run as seperate processes that communicate via message passing. In Win2k and WinXP, almost everything runs in kernel space. Heck, in the next version, rumor has it that large parts of SQL and the.NET runtime are going in kernel space! And OS X isn't a microkernel either. It uses Mach, but the BSD server runs in kernel space, and message passing between the two has been replaced by procedure calls.
Two points: 1) As others have pointed out software rarely has 100% of the features you need. Overall, though, let's assume that OSS software only has 90% of the features of the commercial version. You're generally going to have to work around the software anyway, so the main point with OSS software is that you'll have to do some more working around. 2) It's not just about the ability to edit code. It's the overall package of flexibility, low cost, and, most importantly, freedom from vendor schedules that OSS software provides.
It all depends on the criteria you use to define "better." If the openness of the license is a high priority, then an OSS piece of software is "better" by your criteria.
The classic flaw of pragmatism. Concerned only with near-term convenience and practicality, while ignoring the big picture. The idea is that if something meets 90% of your needs, you can usually work around the 10% that is lacking. If that initial investment in work arounds means that, down the road, you gain a much larger pay-off, then its a worthwhile tradeoff.
Slashdot Mirror
I really think that I should write a filter that spell-checks an email, and rejects it if over 50% of the words with 5 or more letters are misspelled. :)
>>>>>>
That would rock. If such a filter was on Slashdot, Slashdot's post volume would drop by 90%
Your hands are 7"x10"x2"? Can you like palm a beachball or something?!!
Not really. In the mainstream market, people don't shop for software on CheapBytes. Hell, most people pay 2x markup at Comp "How May We Ream You Today?" USA.
Still not right. You can sell it all you want, and you don't have to tell them where they can get it free. All you have to do is provide source to anybody who bought it.
Actually, I think there are serious differences between the DRI model and what NVIDIA's hardware dose. Recently, they even switched away from XAA to their own model because XAA apperently didn't mesh well with how their cards did mixed 2D/3D rendering.
Don't look at me. I live in Virginia, and most definately did *not* vote for him! And yes, I did vote. In Virginia, my democratic vote makes not one iota of difference, but I suppose it has ceremonial meaning.
What kind of idea is that? Blowing up computers? He's joking...right? Its a violent and incredibly base thing to say. How do these people make it into Congress?
Wait a minute... so the minute Microsoft decides to compete in a market they've crossed the line?
>>>>>>
Try reading what I said. I said there are three things they need to do to cross the line. Until now, they were only doing the first two. It's only when they started doing the third thing that they crossed the line. MS can compete in search engine space as much as it wants. Just don't do it by leveraging integration with Windows.
Gee, the X-Box hasn't been a smash success, I wonder why that is?
>>>>>>>
The X-Box hasn't been a success because its not the best product. Microsoft can't leverage Windows with the XBox, and thus they're competing fairly in that market. They're also getting their ass kicked --- Sony has twice as much market share as Microsoft and Nintendo combined.
Likewise, they're going to compete in a market dominated by Google and Yahoo and you think that by virtue of them having a monopoly SOMEWHERE that they're going to rush in and crush everyone?
>>>>>>>>>>
This isn't an unrelated monopoly. Unlike the XBox case, Microsoft can very easily use its domination of the OS market to try and dominate the search engine market. It wouldn't be very hard to do at all. With WinFS coming, I can easily envision a Windows search tool that unified internet and local file searching, tieing into MSN Search in the process. This would be illegal, because they'd be using their monopoly control of Windows to lock Google out of the market.
Your description hasn't been the case for awhile now. Almost all the chipsets these days use a proprietory link between the southbridge and the northbridge. For example, in NVIDIA's nForce chipsets, there is a 800MB/sec hypertransport link between the MCP (southbridge) and the IGP (northbridge). SiS uses something called MuTIOL, while Via uses V-Link.
You're mixing gigabytes and gigabits. PCI Express is 2.5 gigabits per second in each direction, or about 250MB/sec per pin. AGP 8x is 2.1 gigabytes per second, or 2100 gigabits in total.
There are, but they aren't well supported. I'm pretty sure GCC is the only thing that will compile FreeBSD.
There are three parts to the statement:
1) They're a monopoly
2) They're leveraging their monopoly
3) They're doing (1) and (2) to try to dominate another market.
Until now, Microsoft hadn't been doing 3. MSN search was just a requisite part of any portal site. Now, they're declaring that they've actively decided to actively compete in that market. This is where they cross the line.
Yes, I do hate Microsoft and want them to fail. They've done so much damage to the software industry with their monopoly that it would require a revolution within the company to redeem them.
However, in this case, my personal opinion of Microsoft has nothing to do with it. The fact that Windows is a monopoly is not in question. The fact that they are leveraging Windows in this case (because Internet Explorer is a part of Windows now, isn't it?) seems very clear. It is illegal for a company to use its monopoly power to try and dominate an unrelated market. It's as simple as that.
PS> Before you defend Microsoft by saying that all businesses are dishonest and compete unfairly, think about one thing: the computing industry is ridiculously broken. Microsoft products dominate pretty much every major software catagory. They've got the major operating system (Windows), the major office suite (Office), the major web browser (Explorer), the major development environment (Visual Studio), etc. This situation is positively unhealthy for the industry. Compare this to other industries. Ford and GM are both huge companies, but they face intense competition in their market. Sony is a giant corporation, but RCA and Philips are doing just fine in the electronics industry, just as lots of other companies are doing fine in the media industry. These days, even Intel has to stay wary of AMD. Competition is good: the intense competition between Intel and AMD in the last few years has spurred some of the fastest growth in processor technology in recent memory. Just by simply comparing the computer industry to others you can easily see that there is a problem, and Microsoft is that problem!
They try to compete while having an unfair advantage. The government clearly demonstrated that they are an illegal monopoly, but wussed out when it came to punishing them for it. Justice must come from somewhere, and sometimes, it must come from the people themselves. Think of this as a posse rounding up to do what the government is too weak to.
MSN bots? Am I the only one thinking of the Sentinals from the Matrix at this point?
Yes. Linux is constantly evolving and improving. The improvements come from two places: first, lots of people have access to the kernel source, and they get new ideas about how to make things better. Second, and more importantly, Linux is being put to an increasingly large range of uses. Kernel 2.2 makes for a fine server for a small 1 or 2 CPU machine with not too much memory, but makes a rather poor destkop. With all the improvements that have happened over the past year, kernel 2.5 makes for a damn good desktop or media workstation machine, in addition to handling its previous tasks without too much increased overhead. In the 2.7 kernel series, we're going to see major improvements to the NUMA architecture, so kernel 2.7 is going to be great for large-scale NUMA machines. Windows, on the other hand, is rather limited in what it can do. It doesn't handle real small platforms very well, or big NUMA setups. Even on the desktop, its VM (which desperately needs a rewrite) can cause problems.
The performance hit really matters on what you're doing. First, having attributes adds no performance hit --- XFS and ext3 already do it. Second, the indexing overhead really only comes into play for certain cases. Mainly, there is a performance impact on benchmarks that modify file metadata. Operations like file reads and writes shouldn't be affected. For the types of files that users really want to maintain in the index (documents, media files, etc) metadata updates don't happen often enough for the indexing mechanism to become a bottleneck. On the other hand, for something like a Squid cache, where file create/delete performance is critical, you don't really care if those files are in the index, because the user won't be searching for them. If indexing is applied on a per-attribute basis (and the user takes care not to tag stuff like cache files with attributes that will be indexed) the performance hit of indexing should be minimal.
Actually, in this case, Linux might very well do it first. Reiser4 has all the bits and pieces in there to be a killer DBFS back-end. They're trying to get good file performance out of objects as small as individual XML elements. Throw on a database layer on top (doesn't have to be that featureful for something like this) and you've got yourself a WinFS competitor. I think MS might have done themselves a disservice by throwing these ideas out there this early. Longhorn isn't due out for almost two years. In that time, it may very well be that many of the features touted for Longhorn have been in Linux for months. Remember, this time frame we're talking about is about how long it took the KDE guys to totally rewrite KDE in the 1.0 to 2.0 transition. In the last year, Linux has had several major subsystems totally overhauled. In a year and a half, Linux changes as much as Windows does in a full release cycle (2-3 years). I would not be surprised if Longhorn has some serious competition when it debuts.
Not really. A database FS is one of the few new advances that add real power to systems. If you think about the human thought process, they don't tend to think of things in terms of strict hierarchies. They tend to look at a set of items from multiple viewpoints, possibly creating a hierarchy within a given subset, but rarely subject all items to the same hierarchy. There are lots of places where a hierarchical filesystem doesn't really make sense. Is there any sane method of organizing a documents directory? Whenever I try to do it, I generate huge, complex hierarchies with only a few files in each directory. Organizationally, its elegant, but its a pain to traverse. Instead, if I simply shove all documents into a single store, I can use the database mechanism to view the file set in a way that is most useful for the task at hand. A media file directory is another good example. Its tedious to organize media files into a strict hierarchy, and very limiting to have to deal with that hierarchy every time, especially if you want a different view of the data. For example, usually, I organize music files by artist, then album. But what if I want to quickly access files based on catagory (blues) or period (80's)? Since the artist trait is usually orthogonal to the catagory and period traits, a simple hierarchy fails to adequately describe the situation. This is why jukeboxes like iTunes or MusicMatch are getting so popular: they are very flexible with how songs are presented to the user. The whole idea here is to take the power and efficiency of something like the iTunes interface, and apply it to file management in general. If a simple query language is supported, such a scheme could potentially be very usable via a command line as well.
First, XP, without the SQL or CLR is still not a microkernel. It's less of a microkernel than Linux (which isn't a microkernel) because it runs more stuff in kernel space. The "rumor" is just a bit of additional information that shows that MS is adding even more cruft to the kernel.
OS X uses something that used to be a microkernel, but its been totally changed. The BSD server runs entirely in kernel mode, and message passing between Mach and the BSD server has been replaced by function calls.
NT is hardly a microkernel. A microkernel, according to strict definitions, doesn't include anything like drivers, paging or the filesystem. QNX fits this definition --- the filesystem runs in userspace, and even drivers run as seperate processes that communicate via message passing. In Win2k and WinXP, almost everything runs in kernel space. Heck, in the next version, rumor has it that large parts of SQL and the .NET runtime are going in kernel space! And OS X isn't a microkernel either. It uses Mach, but the BSD server runs in kernel space, and message passing between the two has been replaced by procedure calls.
Two points:
1) As others have pointed out software rarely has 100% of the features you need. Overall, though, let's assume that OSS software only has 90% of the features of the commercial version. You're generally going to have to work around the software anyway, so the main point with OSS software is that you'll have to do some more working around.
2) It's not just about the ability to edit code. It's the overall package of flexibility, low cost, and, most importantly, freedom from vendor schedules that OSS software provides.
It all depends on the criteria you use to define "better." If the openness of the license is a high priority, then an OSS piece of software is "better" by your criteria.
The classic flaw of pragmatism. Concerned only with near-term convenience and practicality, while ignoring the big picture. The idea is that if something meets 90% of your needs, you can usually work around the 10% that is lacking. If that initial investment in work arounds means that, down the road, you gain a much larger pay-off, then its a worthwhile tradeoff.