Considering moving back to Linux

[rabbit]

What?  I said I'm not sure if he had SP1 or not.  If he did, your point is blown.

[Warrior]

What?  I said I'm not sure if he had SP1 or not.  If he did, your point is blown.

Except, theres a possibility he did. That makes the second part of your sentence incorrect, because the point would of been moot.

[rabbit]

RETORT


That aside, I'll try to remember to ask next Wednesday (the next classtime) if/how long he's had SP1.

[nslay]

Vi?

He means they're using the Microsoft platform for the class, not just the IDE.

What kind of CS class develops on Microsoft platform?  At least, I think thats a bit odd...  Generally when I think CS in academia, I think Unix.

Probably one that wants to use the most advanced language/platform on the planet. :)

Dream on, you can't even get copying files, and print-to-file correct... Windows is advanced at being fragile and time consuming.  Too bad C++ (implied by "most advanced language") wasn't invented by Microsoft...that would require innovation, instead of copying.  I do have to say Microsoft makes some excellent interfaces for their applications...though I wonder whats behind the scenes when 77.1*850 = 100000.

Read the 'copying files' link, its a pcmag column...I love John C Dvorak...he's a hilarious columnist.

If you knew anything about anything, you'd know that rabbit's class is using C++ .NET.
When I said most advanced language and platform I meant the .NET Language family (Mainly C#) and the .NET Platform. Both of which are not specific to Vista.

Yes, university courses are often about Microsoft fads...  I have yet to take one or even see one.  And yes, these things are fads...they come and go.  How do I know?  I grew up with a hardcore Microsoft technologies developer...he's been doing this stuff for IBM for decades and boy I hear, even to this day, what new cool technological fad Microsoft came out with.  What about all those other previous fads?  Microsoft is a business, I expect them to continue to make new fads...they have to make money somehow.  C/C++ on the other hand...these languages are the real thing...they stood the test of time and will continue to persist.  They are among the most powerful languages ever made...and I'm sorry, C#, Java and other newer languages don't compare.  Oh sure, you can make GUIs and other neat toys with them...that suits them quite well.  As for real time and high performance applications...people will laugh at you.

I also love how you seem to imply that alternatives to Windows do not have bugs, or even that any software does not contain bugs. Unless of course you enjoy sounding like a hypocrite?

Unix systems are incredibly simple (as characteristic of early operating systems), their code base are incredibly small compared with Windows' code base.  I'd expect that bugs are far fewer in any Unix system, than the more complicated Windows operating systems.  My experience with it daily, at work, and at home seem to support this point quite well...on even Linux ( and I hate Linux!).  Hey, at least I can print-to-file and copy files pretty easily! :D

Microsoft made 16 billion this quarter, somewhere along the line something is being done correctly. All of this "Vista is a failure", "Vista is not catching on" does not seem to go with what their earnings are reporting :).

Oh, the old market share argument...haven't you learned your lesson from last time?  Market share doesn't mean technological superiority. Windows 9x was a monopoly too...but being a DOS based system, it was technological inferior to everything else of the time.

Microsoft records $16 billion...but nobody is running to CompUSA or Best Buy screaming for Windows Vista copies.  Instead,  Mary and Joe buy a new computer with it pre-installed.  It's not that people want Vista, its that people just end up with it.  Just from general consensus, it seems people and business are staying away from it... Nobody really wants it right now. 

[Joe]

I've been considering Slackware because I've used it before and I like it, but I hate Ubuntu.  I've never tried Gentoo, and I'm not going to go out on a limb with it.  Maybe on a test box a little later?  I don't know.  Thanks.

What's wrong with Ubuntu?  People call it the "noob" distro so it turns you off it? :P

Joe uses it.

Joe uses Vista, Debian, and NetBSD.

@nslay:
If Microsoft teamed with a company (perhaps Motorola?) to build a processor that ran MSIL semi-natively, I think that C# could blow C++ out of the water. Currently, yeah, it's .NET specific, but if .NET becomes more global and can be run on a lower level (IE, outside of a virtual machine), it's got potential -- it's like Visual Basic simplicity on a higher level of power.

The reason I say Motorola is because of their history with PowerPC, a processor much more powerful than Intel Architecture. I think that it wouldn't be too difficult to build a processor that runs MSIL on a level nearly at that of x86's system level assembly. Of course, phasing that in would be hell for everybody, but if it were done properly, it could be really cool.

[Warrior]

Yes, university courses are often about Microsoft fads...  I have yet to take one or even see one.  And yes, these things are fads...they come and go.  How do I know?  I grew up with a hardcore Microsoft technologies developer...he's been doing this stuff for IBM for decades and boy I hear, even to this day, what new cool technological fad Microsoft came out with.  What about all those other previous fads?  Microsoft is a business, I expect them to continue to make new fads...they have to make money somehow.  C/C++ on the other hand...these languages are the real thing...they stood the test of time and will continue to persist.  They are among the most powerful languages ever made...and I'm sorry, C#, Java and other newer languages don't compare.  Oh sure, you can make GUIs and other neat toys with them...that suits them quite well.  As for real time and high performance applications...people will laugh at you.

Why don't you forget for a second that Microsoft birthed .NET and C#, and look at the language/platform for the features it offers as opposed to the logo it's under.

I mean, you're welcome to cite a richer Language/Platform combo in the programming world. It's not a secret that C#/.NET are not high performance, and that is not really their target audience although Java/C# JIT'd code is really competetive in speed when compared to C/C++.

I also think you have a big misunderstanding of the technology world, it's a fast paced ever changing field so "Fads" come and go.

The needs for better languages, with more flexibility and functionality as well as richer APIs is a constant one. There is always better things they can do, especially when it comes to making Managed Programming an easy task.

Unix systems are incredibly simple (as characteristic of early operating systems), their code base are incredibly small compared with Windows' code base.  I'd expect that bugs are far fewer in any Unix system, than the more complicated Windows operating systems.  My experience with it daily, at work, and at home seem to support this point quite well...on even Linux ( and I hate Linux!).  Hey, at least I can print-to-file and copy files pretty easily! :D

I don't care what you "expect", I care about the statistics. Let's look at the security track history for Windows Vista, you'll find that it's pretty good.

This is of course, discounting bugs silently fixed by Microsoft (and obviously, not counted so the results may be a little tainted), and the severity of the exploits and the speed to which they were fixed must be accounted for.

Even so, you'll see the track record is astronomically smaller than Windows XP, and comparable to the likes of OSX and some flavors of Linux.

Oh, the old market share argument...haven't you learned your lesson from last time?  Market share doesn't mean technological superiority. Windows 9x was a monopoly too...but being a DOS based system, it was technological inferior to everything else of the time.

It's you that should learn your lesson, and the lesson is that technological superiority (or at least what you perceive to be technologically superior), really does not matter as much as making an easy to use system and shipping great Applications with it.

You can bad mouth all you want, list whatever alternatives you want, but it's not going to change this fact.

Microsoft records $16 billion...but nobody is running to CompUSA or Best Buy screaming for Windows Vista copies.  Instead,  Mary and Joe buy a new computer with it pre-installed.  It's not that people want Vista, its that people just end up with it.  Just from general consensus, it seems people and business are staying away from it... Nobody really wants it right now. 

Is that a problem? It's the preferred way of buying computers, how many people are really going to sit down and pop in a CD? Honestly.

Microsoft still makes money off of selling OEM licenses (Obvious from their record breaking quarterly earnings), and Vista has still moved 100 million copies in it's first year.

They sold more in a week than there are users of OSX and Linux, and currently hold more marketshare than Linux, OSX and any version of Windows pre-XP combined.

There's absolutely no mind that Vista is moving units, and it's only a matter of time before it's the dominant operating system just as I predicted it would be during early in it's Beta cycle.

[Warrior]

@nslay:
If Microsoft teamed with a company (perhaps Motorola?) to build a processor that ran MSIL semi-natively, I think that C# could blow C++ out of the water. Currently, yeah, it's .NET specific, but if .NET becomes more global and can be run on a lower level (IE, outside of a virtual machine), it's got potential -- it's like Visual Basic simplicity on a higher level of power.

The reason I say Motorola is because of their history with PowerPC, a processor much more powerful than Intel Architecture. I think that it wouldn't be too difficult to build a processor that runs MSIL on a level nearly at that of x86's system level assembly. Of course, phasing that in would be hell for everybody, but if it were done properly, it could be really cool.

This would defeat the purpose of Managed Programming, completely. There is no significant slowdown, the slowdown is purely statistical and for number counters and it's very very competitive in some areas.

With Managed Programming you want runtime checking and adaptation of the code, if you "compile" IL, you can only do so much to assure security and then digitally sign it.

Singularity does some of this, with their MSIL->x86 Instruction converter, and they're able to proof their components which allows them to run everything in kernel mode since the protection of memory is assured by the proof.

It's an interesting field, and I have to go or I'd type more but there's a lot of potential for not just managed code, but for the concepts it brings forth to expand in the world of operating systems.

[nslay]

Yes, university courses are often about Microsoft fads...  I have yet to take one or even see one.  And yes, these things are fads...they come and go.  How do I know?  I grew up with a hardcore Microsoft technologies developer...he's been doing this stuff for IBM for decades and boy I hear, even to this day, what new cool technological fad Microsoft came out with.  What about all those other previous fads?  Microsoft is a business, I expect them to continue to make new fads...they have to make money somehow.  C/C++ on the other hand...these languages are the real thing...they stood the test of time and will continue to persist.  They are among the most powerful languages ever made...and I'm sorry, C#, Java and other newer languages don't compare.  Oh sure, you can make GUIs and other neat toys with them...that suits them quite well.  As for real time and high performance applications...people will laugh at you.

Why don't you forget for a second that Microsoft birthed .NET and C#, and look at the language/platform for the features it offers as opposed to the logo it's under.

I mean, you're welcome to cite a richer Language/Platform combo in the programming world. It's not a secret that C#/.NET are not high performance, and that is not really their target audience although Java/C# JIT'd code is really competetive in speed when compared to C/C++.

Okay, cite: http://basepath.com/aup/
Nah...they're really not competitive with C/C++.  I know you've probably read this in places, but whatever C/C++ code they benchmarked against is probably poorly written (as is characteristic of high level programmers).

I also think you have a big misunderstanding of the technology world, it's a fast paced ever changing field so "Fads" come and go.

But yet so many good historical development tools and languages are still used today.  That Ritchie fellow...what a genius.

The needs for better languages, with more flexibility and functionality as well as richer APIs is a constant one. There is always better things they can do, especially when it comes to making Managed Programming an easy task.

Unix systems are incredibly simple (as characteristic of early operating systems), their code base are incredibly small compared with Windows' code base.  I'd expect that bugs are far fewer in any Unix system, than the more complicated Windows operating systems.  My experience with it daily, at work, and at home seem to support this point quite well...on even Linux ( and I hate Linux!).  Hey, at least I can print-to-file and copy files pretty easily! :D

I don't care what you "expect", I care about the statistics. Let's look at the security track history for Windows Vista, you'll find that it's pretty good.

And how long has Vista been out now?   Besides, the code base for Vista is so much larger than XP...I'd really expect there to be tons of problems with it.

This is of course, discounting bugs silently fixed by Microsoft (and obviously, not counted so the results may be a little tainted), and the severity of the exploits and the speed to which they were fixed must be accounted for.

Even so, you'll see the track record is astronomically smaller than Windows XP, and comparable to the likes of OSX and some flavors of Linux.

XP, the Windows OS of choice for nearly a decade.  I don't know anything about OS X security, but I can tell you Linux security, like anything else, is largely dependent on what its running.  There are a good many 'comes-with-everything' distributions that bundle all kinds of crap that can be exploited...this is really no surprise at all.

Oh, the old market share argument...haven't you learned your lesson from last time?  Market share doesn't mean technological superiority. Windows 9x was a monopoly too...but being a DOS based system, it was technological inferior to everything else of the time.

It's you that should learn your lesson, and the lesson is that technological superiority (or at least what you perceive to be technologically superior), really does not matter as much as making an easy to use system and shipping great Applications with it.

You can bad mouth all you want, list whatever alternatives you want, but it's not going to change this fact.

But it really makes you wonder doesn't it :)

Microsoft records $16 billion...but nobody is running to CompUSA or Best Buy screaming for Windows Vista copies.  Instead,  Mary and Joe buy a new computer with it pre-installed.  It's not that people want Vista, its that people just end up with it.  Just from general consensus, it seems people and business are staying away from it... Nobody really wants it right now. 

Is that a problem? It's the preferred way of buying computers, how many people are really going to sit down and pop in a CD? Honestly.

It's a problem when you want to use this figure to show Vista is somehow popular.  Oh my God!  They'd have to pop a CD in...oh those poor souls.

Microsoft still makes money off of selling OEM licenses (Obvious from their record breaking quarterly earnings), and Vista has still moved 100 million copies in it's first year.

They sold more in a week than there are users of OSX and Linux, and currently hold more marketshare than Linux, OSX and any version of Windows pre-XP combined.

There's absolutely no mind that Vista is moving units, and it's only a matter of time before it's the dominant operating system just as I predicted it would be during early in it's Beta cycle.

It'll be the dominant Windows operating system when XP reaches its EOL in 2010...most likely not before then.

[Warrior]

Okay, cite: http://basepath.com/aup/

I don't know, if you find that easy then I'm afraid to see what that's supposed to simplify.

Nah...they're really not competitive with C/C++.  I know you've probably read this in places, but whatever C/C++ code they benchmarked against is probably poorly written (as is characteristic of high level programmers).

I'm sure I'll take your word for it. Except I won't.

But yet so many good historical development tools and languages are still used today.  That Ritchie fellow...what a genius.

And? My statement is still true.

And how long has Vista been out now?   Besides, the code base for Vista is so much larger than XP...I'd really expect there to be tons of problems with it.

One year in five days, now you can compare it to the first year drops of Windows XP and alternatives. You'll see it's obvious which is more secure.

The fact is, that Vista is rock solid from a security standpoint. The attack surface has been greatly minimized, and viruses have been reduced to user mode threats (Which UAC locks down anyhow).

XP, the Windows OS of choice for nearly a decade.  I don't know anything about OS X security, but I can tell you Linux security, like anything else, is largely dependent on what its running.  There are a good many 'comes-with-everything' distributions that bundle all kinds of crap that can be exploited...this is really no surprise at all.

So do you think Windows is an exception to this rule? It's security is decided by faulty drivers, bad programs, etc.

The improvements in Vista are how the kernel handles these sometimes malicious programs/drivers. With large portions of driver frameworks/stacks being moved to User Mode, the chance they had to crash the Kernel is eliminated.

This goes hand in hand with what I mentioned about minimizing the attack surface.

It's a problem when you want to use this figure to show Vista is somehow popular.  Oh my God!  They'd have to pop a CD in...oh those poor souls.

And wait for the Operating System to install, and configure packages/programs, and locate drivers, and setup firewall rules, and ...

You see where it gets tedious? Not everyone is computer literate, people want to buy a computer, plug it in, and surf the web.

There are OEM Linux distributions as well, so it's fair game.

It'll be the dominant Windows operating system when XP reaches its EOL in 2010...most likely not before then.

I highly doubt that, people said the same thing about Windows 2000 when Windows XP dropped. Same exact complaints.

What happened there? Didn't exactly pan out as you're saying there either.

[nslay]

Okay, cite: http://basepath.com/aup/

I don't know, if you find that easy then I'm afraid to see what that's supposed to simplify.

Nah...they're really not competitive with C/C++.  I know you've probably read this in places, but whatever C/C++ code they benchmarked against is probably poorly written (as is characteristic of high level programmers).

I'm sure I'll take your word for it. Except I won't.

If you really believe Java and C# actually compete with C/C++ you must be brain dead.  I mean take for example a matrix...say you want to do something like the Jacobian-method with it.  In Java, you have to allocate a jagged matrix.  What does this mean?  It means each row points to a different location in memory (and likely not contiguous)...this does NOT preserve locality of reference...this causes cache and possibly TLB misses.   In C/C++ you get 2d<->1d mapping row major array...this preserves locality of reference quite well...and actually, you get a couple orders of magnitude speed up over a Java implementation.  This is just one simple example where languages like C# and Java aren't competitive with C/C++.  Many C/C++ compilers can even take direct advantage of processor features like SSE...some can vectorize and parallelize code too.  And all aside, with tools like OpenMP...Java and C# couldn't hope to get speed comparable to C/C++...it's not even funny how Java and C# aren't competitive with C/C++.

So to summarize:
1) Java is quasi-interpreted
2) Java's memory management does not preserve locality of reference well
3) Java probably can't take advantage of prefetching (because references can change...this problem is in C pointers too, unless you use ISO C99's 'restrict' keyword)

I'm sure similar can be said of C#.

C/C++
1) Can be optimized by the compiler all the way down to the processor level (e.g. vectorization)
2) Use of references or restrict pointers allow the compiler to make more aggressive optimizations (e.g. prefetching)
3) C/C++ support static multi dimensional arrays that are contiguous in memory, thus preserving locality of reference and reducing misses
4) OpenMP makes parallel programming on C/C++ child's play...its totally transparent (i.e. no working with threads/locks).

And more can be said than just those 4 points.

I think this should make my point more obvious.  I think you were originally correct...C# and Java aren't high performance.  I have no idea why you decided to state that they are competitive with C/C++.  They aren't.

But yet so many good historical development tools and languages are still used today.  That Ritchie fellow...what a genius.

And? My statement is still true.

Let's remember, C has been around for 30+ years, C++ has been around for 20+ years...these languages are still in demand.  It means attempts to one-up these languages has failed.  I don't see them ever being replaced.  Oh sure, languages like C# will come and go...but these will persist.

And how long has Vista been out now?   Besides, the code base for Vista is so much larger than XP...I'd really expect there to be tons of problems with it.

One year in five days, now you can compare it to the first year drops of Windows XP and alternatives. You'll see it's obvious which is more secure.

And because Vista has been out "one year and five days", I don't think its fair to compare Windows XP with Windows Vista...XP has been out since 2001.  No, it's not obvious which is more secure...I swear you never learn your lesson.  The same was said about XP when it came out. 

The fact is, that Vista is rock solid from a security standpoint. The attack surface has been greatly minimized, and viruses have been reduced to user mode threats (Which UAC locks down anyhow).

That's great that Windows has finally adopted control over loadable kernel modules (drivers)...operating systems like *BSD have had this capability since their inception in the early 90s.

XP, the Windows OS of choice for nearly a decade.  I don't know anything about OS X security, but I can tell you Linux security, like anything else, is largely dependent on what its running.  There are a good many 'comes-with-everything' distributions that bundle all kinds of crap that can be exploited...this is really no surprise at all.

So do you think Windows is an exception to this rule? It's security is decided by faulty drivers, bad programs, etc.

You totally missed the pont of the very first sentence.  I'm emphasizing that XP has been used for a half decade+...and you automatically compare its track record to a very new Windows operating system.  Well, I don't know if you've noticed...at every windows release, the same crowd (i.e. people like you) all make claims about how secure it is...well it took 2 service packs and 5 years to lock XP down!

The improvements in Vista are how the kernel handles these sometimes malicious programs/drivers. With large portions of driver frameworks/stacks being moved to User Mode, the chance they had to crash the Kernel is eliminated.

And I'd like to this moment to emphasize how other OS have had this capability decades ago.  The original BSD had a lot of its kernel in userspace, Linux has a lot of its kernel services in userspace (e.g. udev).

This goes hand in hand with what I mentioned about minimizing the attack surface.

It's a problem when you want to use this figure to show Vista is somehow popular.  Oh my God!  They'd have to pop a CD in...oh those poor souls.

And wait for the Operating System to install, and configure packages/programs, and locate drivers, and setup firewall rules, and ...

You see where it gets tedious? Not everyone is computer literate, people want to buy a computer, plug it in, and surf the web.

There are OEM Linux distributions as well, so it's fair game.

Not only have I installed every other Windows system all the way back to Windows 3.1...I've even installed Windows Vista.  And I gotta say...its pretty easy.  I think the hardest thing you have to do is specify that your language is English!  Firewall rules?  Computer illiterate people care about firewalls?  Doesn't Vista enable all that junk for you anyhow?  I guess drivers with Vista would be a problem! 

And honestly...admit it, its not tedious.  I'm sorry, I like Unix, but Linux is not as easy as that...even in Ubuntu you'll likely have to open the scary terminal.

It'll be the dominant Windows operating system when XP reaches its EOL in 2010...most likely not before then.

I highly doubt that, people said the same thing about Windows 2000 when Windows XP dropped. Same exact complaints.

What happened there? Didn't exactly pan out as you're saying there either.

It's not that people said this...it's just a fact of life.  Windows licenses cost!  New Windows releases are also notorious for having bugs and security problems that many just want to wait it out.  Companies like IBM...my university even, waited years after XP's release to begin using it.  They probably waited for the EOL on Windows 2000 to justify spending the money for XP.

[Warrior]

I'm sure I'll take your word for it. Except I won't.

^ I think you missed that.

Let's remember, C has been around for 30+ years, C++ has been around for 20+ years...these languages are still in demand.  It means attempts to one-up these languages has failed.  I don't see them ever being replaced.  Oh sure, languages like C# will come and go...but these will persist.

That remains to be seen. Again, I'm not going to take your word for it.

And because Vista has been out "one year and five days", I don't think its fair to compare Windows XP with Windows Vista...XP has been out since 2001.  No, it's not obvious which is more secure...I swear you never learn your lesson.  The same was said about XP when it came out. 

Can you read? I said "now you can compare it to the first year drops of Windows XP and alternatives."
Compare it to the security track records of the first year of other operating systems, it's proven to be extremely resilient.

That's great that Windows has finally adopted control over loadable kernel modules (drivers)...operating systems like *BSD have had this capability since their inception in the early 90s.

What's your point? I never stated the opposite.

You totally missed the pont of the very first sentence.  I'm emphasizing that XP has been used for a half decade+...and you automatically compare its track record to a very new Windows operating system.  Well, I don't know if you've noticed...at every windows release, the same crowd (i.e. people like you) all make claims about how secure it is...well it took 2 service packs and 5 years to lock XP down!

And the Windows Vista codebase is based off of the Windows Server 2003 codebase, which is branched from Windows XP SP2.

You claim yourself that it took that long to "lock XP down", and since Vista stems from this code it makes XP inherently "locked down"

Thank you, you've proven my point. Now run along.

Vista has had half the bugs/exploits that Windows XP has had in it's first year. I'm not sure of the comparisons to other operating systems off the top of my head, but you're welcome to do some research.

And I'd like to this moment to emphasize how other OS have had this capability decades ago.  The original BSD had a lot of its kernel in userspace, Linux has a lot of its kernel services in userspace (e.g. udev).

Again, did I ever say they didn't? I'm citing this as a security enhancement in Vista, if you're trying to downplay this then you automatically downplay their counterparts in other Operating system.

One must ask if you're actually reading before you hit Post.

Not only have I installed every other Windows system all the way back to Windows 3.1...I've even installed Windows Vista.  And I gotta say...its pretty easy.  I think the hardest thing you have to do is specify that your language is English!  Firewall rules?  Computer illiterate people care about firewalls?  Doesn't Vista enable all that junk for you anyhow?  I guess drivers with Vista would be a problem! 

And honestly...admit it, its not tedious.  I'm sorry, I like Unix, but Linux is not as easy as that...even in Ubuntu you'll likely have to open the scary terminal.

It's definitely much harder than it should be on the consumer. It ruins the Out of the Box Experience (The end to end experience Microsoft puts emphasis on, and a lesson that OSS is slowly learning with things like KDE4)

A consumer does not want to be burdened with setting this crap up, and managing drivers for their system, managing updates, setting up User Accounts/Permissions, all the things necessary to getting an Operating System up and running.

An OEM is allowed to customize the Vista SKU they're shipping the product with, and ship their at-times proprietary drivers, customer support software, Firewall/Browser rules, personalized Sidebar Gadgets, etc.

It's more work than the consumer should have to do, and of course there is the option of asking an OEM to include an install disk if you're so inclined.

I don't see what the problem is.

It's not that people said this...it's just a fact of life.  Windows licenses cost!  New Windows releases are also notorious for having bugs and security problems that many just want to wait it out.  Companies like IBM...my university even, waited years after XP's release to begin using it.  They probably waited for the EOL on Windows 2000 to justify spending the money for XP.

Yet, XP caught on extremely fast. Despite what you say, it was quickly the dominant operating system.
The same will happen with Windows Vista, and the same will happen with Windows 7, etc.

For every two Windows releases, OSX will gain about 1% of marketshare, and Linux users will have about 10 more years of false hope.

[Chavo]

If you really believe Java and C# actually compete with C/C++ you must be brain dead.  I mean take for example a matrix...say you want to do something like the Jacobian-method with it.  In Java, you have to allocate a jagged matrix.  What does this mean?  It means each row points to a different location in memory (and likely not contiguous)...this does NOT preserve locality of reference...this causes cache and possibly TLB misses.

I don't like to get involved here, but I just wanted to point out that I doubt most people here are even going to know what a cache miss is.  I'd like to then point out that the performance loss due to the possible miss is going to depend pretty heavily on the processor-memory architecture.

[iago]

Hang on a sec....

...and Linux users will have about 10 more years of false hope.

How do you mean? Are you implying that Linux users want everybody else to be Linux users? And, in fact, are you implying that a higher marketshare is a useful metric of any kind? Because if so, I totally disagree. Just before more people like something doesn't make it good.

[Warrior]

Hang on a sec....

...and Linux users will have about 10 more years of false hope.

How do you mean? Are you implying that Linux users want everybody else to be Linux users? And, in fact, are you implying that a higher marketshare is a useful metric of any kind? Because if so, I totally disagree. Just before more people like something doesn't make it good.

That seems to be the mindset taken by most I've talked to, yes.
They've been claiming that this year is the year of the Desktop Linux, or the year that Linux will from the ashes of irrelevance.

While I think very positive strides are being made (Ubuntu, KDE4, etc) in terms of usability, it still has a long way to go before it can achieve any measurable form of success.

You may not see the world like that, but I think you're someone who sees things from a more passive position which is generally a good thing. I really don't care who uses Linux, use what works. What I do care about is defending a strong language+platform combination, the fact that it degenerated into a blow-for-blow Linux vs Windows vs OSX pissing contest is the direct result of nslay's inflammatory responses.

[iago]

That seems to be the mindset taken by most I've talked to, yes.
They've been claiming that this year is the year of the Desktop Linux, or the year that Linux will from the ashes of irrelevance.

Well, I guess I've been trying to live a non-competitive lifestyle, and my opinions are sort of based on that. I don't see it as a competition, anymore than I see programming languages in competition. People should use whatever's best for the job at hand.

Just because I rarely see Windows or C# as being the best for the current job, that's besides the point... :P

While I think very positive strides are being made (Ubuntu, KDE4, etc) in terms of usability, it still has a long way to go before it can achieve any measurable form of success.

In my opinion, it has already achieved hugely amounts of success. It's widely considered one of the most stable operating systems and is used on servers worldwide.

If you mean in terms of usability, then that's a little different. I don't especially care, though. Linux is for geeks, and nobody else. End of story.

You may not see the world like that, but I think you're someone who sees things from a more passive position which is generally a good thing. I really don't care who uses Linux, use what works. What I do care about is defending a strong language+platform combination, the fact that it degenerated into a blow-for-blow Linux vs Windows vs OSX pissing contest is the direct result of nslay's inflammatory responses.

err, read my first comment. I think I got mixed up. It's all in my newsletter, though.