FreeBSD maintains the same kind of stability WITH a more current release schedule.
FreeBSD doesn't have packages for most of things and for a few platforms. Compare that with releasing 12000 packages (14 CDs, IIRC?) for 10-12 architectures. Is not that FreeBSD sucks, they work great, but is not fair to compare two things that are not really the same. And BTW, the 4.X -> 5.3 step has not been exactly "fun".
(and don't come saying "this is the proof that ports > packages. Time has showed everybody that packages are valuable, I don't want to start recompiling libc or X.org because of a critical security bug when I have a spike load, ok?)
Even more, releasing their code not only won't hurt them, it will actually help them. Google's revenue comes from ads not from selling software, sharing the code means other people can improve and help to maintain it.
Actually, linux scales, and not just on clusters. Some people, like say, the NASA guys, run it on 512 simultaneous processors, people also runs it four, eight, 32 or whatever, and all with the same kernel, not a cluster but a big computer. And of course it musr run estable.
Actually, some of those Enterprises (like cisco or dell) don't have a server OS to say "our option is better, that's why you shouldn't use linux". Why should Dell tell you what you need to run? Linux is the fatest growing server platform, so they should shut up their mouth and working on better linux support if they want to remain profitable.
The one contenders there are Microsoft and Sun. We know that windows don't runs more than 64 cpus so they also should shut their mouth up and work on that. Then there's Sun, who have a real OS, but thinks that Linux is "no loss" for them, despite of having lost lots of customers to Red Hat's hands.
When I enter gmail.com i see "gmail BETA by google". It's still beta, the fact that they're giving away lots of invitations does NOT mean they're not beta.
Some of us want an OS that can run on 128 simultaneous processors as well as one or four or twelve all with the same kernel. Not a cluster. One big computer.
That's fine. Some people, like say, the NASA guys, want an OS that can run on 512 simultaneous processors as well as one or four, or eigth, or 32, or whatever, they want it stable and all with the same kernel. Not a cluster. One big computer. A really big one.
If it doesn't have enough permissions, it doesn't even know the registry key is there. Which is why it says "not found".
Well.. and how could then try to open something that it can't see? Obviously, there's something
Besides, the error is When you use an Internet Information Server (IIS) ASP page to access a database, the connection may fail with the following error message:
Microsoft OLE DB Provider for ODBC Drivers error '80004005'
[Microsoft][ODBC Driver Manager] Data source name not found and no default driver specified
ie: it doesn't says anything about a registry key.
Apparently you haven't used windows a lot. Use XP, go to the even viewer, look at any error. It'll say you "even number foo, check http://www.microsoft.com/foobar for more details". Great help, what if the event happens to be a network card error and I can't visit the site??
Call me when Microsoft starts including the documentation in the OS instead of giving me meaningless numbers.
Oh, and I don't think that a support site for servers that says you "click in start -> run and type regedt32.exe" instead of "modifiy registry key HKLM/blah" is a good support site. In that microsoft support article you'll see lot of text, but having a lot of text doesn't means the article is good.
The one thing I've clear is that Microsoft OLE DB provider gave me a error, and instead of saying me "I couldn't access $THIS registry key, not enought permissions", which would have gave me a clue and I could have figured out the fix myself, it gave me a meaningless error number and I had to go trought the web to see what was happening (of course OSS documentation is usually inexistent, so even that is good i suposse...)
In my country (spain) you are allowed to download music for "private use"
You're not allowed to download illegal software even for "private use", you're forced to do whatever the license says (which is what music should allow to do, put it a license)
I bet normal users would benefit from this too. I'm not asking them to distribute "beta patches", just make them available
Right now, bugs are found and fixed between SP releaes, but you don't see them available either. They put a KB entry in their site, and you've to use your phone to ask them for the fix - it' doesn't really mind that they already have the fix, they won't make it publicy available until the SP release, leaving people without the fix for more than a year. Now they're doing the same with security fixes. Join that to the blatant stupidity that is their "patch release policy": "We only release security fixes on Tuesday". WTF? I'm surprised people is still using Microsoft products, I pay a windows XP license and I get this kind of support? Fuck them.
The current version is 1.0-Beta, and it's much better than any alternative I've seen in the OSS world, much better than mozilla's equivalent. Take a look or download it.
With kernels getting huge, microkernels could be easier to write and maintain since they have to be bugfree and stable.
Yes, kernels have been getting more complex since a couple of decades ago. Still, the most complex kernels are those from unix systems (solaris, linux, etc) and their uptime is measured in YEARS. I don't see how "macrokernels" are more unreliable
Microkernel supporters have always been biased toward monolithic kernels.
(Mac OS X is NOT a microkernel, BTW, it's just derived from one and hast lost the most important thing from a microkernel - running everything in userspace, mac os x runs many things that a microkernel runs in userspace in kernelspace, just like NT, both are Mach-derived, with Mac os x having influences from bsd and NT from VMS. They're far from being microkernels)
Microkernel supporters have always been biased towards monolithic kernels about "manteinability", "modularity", etc. Modularity does no come from passing messages between processes. It encourages it, but there's no reason which stops you from writing a really bad interface in a microkernel and do it less modular than a monolithich kernel. Modularity comes from the internal structure of the code, and it's perfectly possible to write a modular monolithich kernel. How do you think commercial unixes, BSDs, Linux, have been able to cope with things like SMP, etc, in the past decades. It's exactly what Linus have been saying for years: Modularity is not a feature of a microkernel, is something outside of the scope of microkernels or even kernels, still in most of microkernel sites you'll find comments about "modularity" and the apparent "inhability" of monolithich kernels to deal with complexity
Another rant I have with microkernels is the fact that they can hang the system. Any driver can do it: touching registers of a graphics card, IDE controller etc can hang your machine. It's in fact what happens with X - it implements drivers (the 2D ones) in userspace (like a microkernel) and can hang your machine by touching the wrong register. No software can save you from hanging your machine - as long as a driver has a bug, it can hang your machine, be it in userspace, kernelspace, or not. The amount of people who really thinks that microkernels can avoid that is amazing, despite of being clearly wrong.
Goneme was a project started in 2004 by someone who didn't like the placement of "accept" and "cancel" buttons and who spent countless hour trolling in osnews/slashdot. The only patch released is from July 2004, and it weights 24 KB. As it can be seen, the mailing list is full of everything except patches.
I only can define it as "dead project" - you really have to have something more than "button order preferences is wrong", "I hate windows registry" and "spatial nautilus is broken" to fork a project. Wow, "Mac OS X is better" - what a surprise. Tell me something I don't know. Not using gecko, use KHTML? Well...wow.
I'm not against forking projects, but this fork is ridiculous. No real reasons, real gnome problems are not mentioned, half of it can be solved by changing the default preferences and no code, etc etc
How ironic would it be if Microsoft eventually abandoned.Net and Mono was the only remaining development environment that supported C#?
It still would be great. It seems like the criticism is made to the platform built on top of C#, ie:.NET. C# continues being a good language. We still can build our own FOSS platform with NO implementation of.NET.
AMD pioneered the addition of 64-bit extensions to x86 in 2003 with its Opteron. Intel followed suit halfway through 2004. Despite AMD's earlier arrival, more revenue came from servers using Intel's 64-bit Xeon chips, McLaughlin said: $1.3 billion for Xeon servers, compared with $838 million for Opteron servers.
And the situation is not better in the desktop world. Intel has only lost a 2% with the "Opteron effect" - from 82% to 80% of market share. As soon as intel starts selling 64-bit enabled CPUs (ie: now), most of the x86 desktop boxes with 64 bit extensions will be the ones from intel, not AMD.
Note that in the x86 server world quoted above, intel's market share is even higher, more than 90% I think. Before opteron, the one serious x86 option was Intel. After Opteron, many people has switched to AMD (no suprise, opteron it's just faster and the individual memory bus for each CPU is great for SMP machines) but many people don't care for speed, they like being able to buy mainboards with a intel or serverworks chipset instead of a nvidia nforce crap. Many people cares about the "platform", they don't care if the CPU runs a ridiculous 10-20-percent slower or even single-number percentages.
I've a screenshot with the windows update downloader making progress - the integration with the IE engine works great, I updated my voodoo3 drivers without even clicking in the IE icon.
It doesn't happen with opensource drivers. Linus (and most of linux users) don't care about closed-source drivers. They don't give anything back, we don't give them anything.
Slashdot Mirror
FreeBSD maintains the same kind of stability WITH a more current release schedule.
FreeBSD doesn't have packages for most of things and for a few platforms. Compare that with releasing 12000 packages (14 CDs, IIRC?) for 10-12 architectures. Is not that FreeBSD sucks, they work great, but is not fair to compare two things that are not really the same. And BTW, the 4.X -> 5.3 step has not been exactly "fun".
(and don't come saying "this is the proof that ports > packages. Time has showed everybody that packages are valuable, I don't want to start recompiling libc or X.org because of a critical security bug when I have a spike load, ok?)
Some sites (Won't mention what ones ;-) upload worms to my box through java, with nothing but Norton warning my that it was even being loaded.
Well, Big fucking deal there. Nobody asked you to install java, There's nothing Firefox can (and will) do with 3rd party plugins.
Even more, releasing their code not only won't hurt them, it will actually help them. Google's revenue comes from ads not from selling software, sharing the code means other people can improve and help to maintain it.
Actually, linux scales, and not just on clusters. Some people, like say, the NASA guys, run it on 512 simultaneous processors, people also runs it four, eight, 32 or whatever, and all with the same kernel, not a cluster but a big computer. And of course it musr run estable.
Actually, some of those Enterprises (like cisco or dell) don't have a server OS to say "our option is better, that's why you shouldn't use linux". Why should Dell tell you what you need to run? Linux is the fatest growing server platform, so they should shut up their mouth and working on better linux support if they want to remain profitable.
The one contenders there are Microsoft and Sun. We know that windows don't runs more than 64 cpus so they also should shut their mouth up and work on that. Then there's Sun, who have a real OS, but thinks that Linux is "no loss" for them, despite of having lost lots of customers to Red Hat's hands.
When I enter gmail.com i see "gmail BETA by google". It's still beta, the fact that they're giving away lots of invitations does NOT mean they're not beta.
Some of us want an OS that can run on 128 simultaneous processors as well as one or four or twelve all with the same kernel. Not a cluster. One big computer.
That's fine. Some people, like say, the NASA guys, want an OS that can run on 512 simultaneous processors as well as one or four, or eigth, or 32, or whatever, they want it stable and all with the same kernel. Not a cluster. One big computer. A really big one.
If it doesn't have enough permissions, it doesn't even know the registry key is there. Which is why it says "not found".
Well.. and how could then try to open something that it can't see? Obviously, there's something
Besides, the error is When you use an Internet Information Server (IIS) ASP page to access a database, the connection may fail with the following error message: Microsoft OLE DB Provider for ODBC Drivers error '80004005' [Microsoft][ODBC Driver Manager] Data source name not found and no default driver specified
ie: it doesn't says anything about a registry key.
Apparently you haven't used windows a lot. Use XP, go to the even viewer, look at any error. It'll say you "even number foo, check http://www.microsoft.com/foobar for more details". Great help, what if the event happens to be a network card error and I can't visit the site??
Call me when Microsoft starts including the documentation in the OS instead of giving me meaningless numbers.
Oh, and I don't think that a support site for servers that says you "click in start -> run and type regedt32.exe" instead of "modifiy registry key HKLM/blah" is a good support site. In that microsoft support article you'll see lot of text, but having a lot of text doesn't means the article is good.
The one thing I've clear is that Microsoft OLE DB provider gave me a error, and instead of saying me "I couldn't access $THIS registry key, not enought permissions", which would have gave me a clue and I could have figured out the fix myself, it gave me a meaningless error number and I had to go trought the web to see what was happening (of course OSS documentation is usually inexistent, so even that is good i suposse...)
In my country (spain) you are allowed to download music for "private use"
You're not allowed to download illegal software even for "private use", you're forced to do whatever the license says (which is what music should allow to do, put it a license)
I bet normal users would benefit from this too. I'm not asking them to distribute "beta patches", just make them available
Right now, bugs are found and fixed between SP releaes, but you don't see them available either. They put a KB entry in their site, and you've to use your phone to ask them for the fix - it' doesn't really mind that they already have the fix, they won't make it publicy available until the SP release, leaving people without the fix for more than a year. Now they're doing the same with security fixes. Join that to the blatant stupidity that is their "patch release policy": "We only release security fixes on Tuesday". WTF? I'm surprised people is still using Microsoft products, I pay a windows XP license and I get this kind of support? Fuck them.
The composer is alive and it's being maintained actively.
The current version is 1.0-Beta, and it's much better than any alternative I've seen in the OSS world, much better than mozilla's equivalent. Take a look or download it.
With kernels getting huge, microkernels could be easier to write and maintain since they have to be bugfree and stable.
Yes, kernels have been getting more complex since a couple of decades ago. Still, the most complex kernels are those from unix systems (solaris, linux, etc) and their uptime is measured in YEARS. I don't see how "macrokernels" are more unreliable
Microkernel supporters have always been biased toward monolithic kernels.
(Mac OS X is NOT a microkernel, BTW, it's just derived from one and hast lost the most important thing from a microkernel - running everything in userspace, mac os x runs many things that a microkernel runs in userspace in kernelspace, just like NT, both are Mach-derived, with Mac os x having influences from bsd and NT from VMS. They're far from being microkernels)
Microkernel supporters have always been biased towards monolithic kernels about "manteinability", "modularity", etc. Modularity does no come from passing messages between processes. It encourages it, but there's no reason which stops you from writing a really bad interface in a microkernel and do it less modular than a monolithich kernel. Modularity comes from the internal structure of the code, and it's perfectly possible to write a modular monolithich kernel. How do you think commercial unixes, BSDs, Linux, have been able to cope with things like SMP, etc, in the past decades. It's exactly what Linus have been saying for years: Modularity is not a feature of a microkernel, is something outside of the scope of microkernels or even kernels, still in most of microkernel sites you'll find comments about "modularity" and the apparent "inhability" of monolithich kernels to deal with complexity
Another rant I have with microkernels is the fact that they can hang the system. Any driver can do it: touching registers of a graphics card, IDE controller etc can hang your machine. It's in fact what happens with X - it implements drivers (the 2D ones) in userspace (like a microkernel) and can hang your machine by touching the wrong register. No software can save you from hanging your machine - as long as a driver has a bug, it can hang your machine, be it in userspace, kernelspace, or not. The amount of people who really thinks that microkernels can avoid that is amazing, despite of being clearly wrong.
looks like spain is full of shit.
"and we're proud of it!"
Goneme was a project started in 2004 by someone who didn't like the placement of "accept" and "cancel" buttons and who spent countless hour trolling in osnews/slashdot. The only patch released is from July 2004, and it weights 24 KB. As it can be seen, the mailing list is full of everything except patches.
I only can define it as "dead project" - you really have to have something more than "button order preferences is wrong", "I hate windows registry" and "spatial nautilus is broken" to fork a project. Wow, "Mac OS X is better" - what a surprise. Tell me something I don't know. Not using gecko, use KHTML? Well...wow.
I'm not against forking projects, but this fork is ridiculous. No real reasons, real gnome problems are not mentioned, half of it can be solved by changing the default preferences and no code, etc etc
GTK#, QT#...yet another programming language
A very small portion of .NET is "standardized"
.NET. I was talking about C#
I don't care about
C# looks much nicer, and unlike Java it's a ECMA standard. Why would I want to use Java?
How ironic would it be if Microsoft eventually abandoned .Net and Mono was the only remaining development environment that supported C#?
.NET. C# continues being a good language. We still can build our own FOSS platform with NO implementation of .NET.
It still would be great. It seems like the criticism is made to the platform built on top of C#, ie:
Uh? Did I read well?
Quoting from http://news.com.com/IBM+extends+lead+in+server+ma
AMD pioneered the addition of 64-bit extensions to x86 in 2003 with its Opteron. Intel followed suit halfway through 2004. Despite AMD's earlier arrival, more revenue came from servers using Intel's 64-bit Xeon chips, McLaughlin said: $1.3 billion for Xeon servers, compared with $838 million for Opteron servers.
And the situation is not better in the desktop world. Intel has only lost a 2% with the "Opteron effect" - from 82% to 80% of market share. As soon as intel starts selling 64-bit enabled CPUs (ie: now), most of the x86 desktop boxes with 64 bit extensions will be the ones from intel, not AMD.
Note that in the x86 server world quoted above, intel's market share is even higher, more than 90% I think. Before opteron, the one serious x86 option was Intel. After Opteron, many people has switched to AMD (no suprise, opteron it's just faster and the individual memory bus for each CPU is great for SMP machines) but many people don't care for speed, they like being able to buy mainboards with a intel or serverworks chipset instead of a nvidia nforce crap. Many people cares about the "platform", they don't care if the CPU runs a ridiculous 10-20-percent slower or even single-number percentages.
I've a screenshot with the windows update downloader making progress - the integration with the IE engine works great, I updated my voodoo3 drivers without even clicking in the IE icon.
# Google Chat # Google IM
.Net Passport
Same thing
# Google portal
Google news?
# Google hosting
They've recently adquired the capacity of sellling domains, it has sense
# Google Forum's
Don't reinvent the wheel. Google news fits perfectly - it works somewhat like gmail, it's great, everyone could use it.
# A Google version of
Already there, called "Google account"
Who cares? It will attract users, that's good.
Ugly.
Exactly. That's what Windows-brained people will like.
At least a couple of Microsoft guys has go to google, one of the XAML guys was mentioned in some blogs.
It doesn't happen with opensource drivers. Linus (and most of linux users) don't care about closed-source drivers. They don't give anything back, we don't give them anything.