2008 - A hard deadline

[cyborg]

Interesting and well researched article I am sure will interest everyone here:

World Domination 201 (http://www.catb.org/~esr/writings/world-domination/world-domination-201.html)

In the 1990s Linus Torvalds used to give a talk called World Domination 101 on the early steps he believed Linux would need to take to achieve "world domination — fast" [1]. We've made a lot of progress since then, but Linux desktop market share remains stuck below 5%, which is too low to garner support from hardware vendors in some critical areas like graphics and wireless hardware, and too small a political base from which to effectively oppose software patents, hardware DRM, and other horrors. For those within the Linux community, World Domination 201 is about how to take advantage of an opportunity to go the rest of the way. For those outside it, this is a warning of what's coming if we fail.

When 8-bit microcomputer hardware stopped selling, it took 8 bit software down with it. The end of 16-bit hardware undid the dominance of DOS, and the end of 32-bit hardware spells the end of the road for Win-32. What will replace 32 bit Windows as the next dominant OS has yet to be decided.

The industry-wide switch to 64-bit hardware is opening a critical transition window during which the new dominant operating system will be determined. This window will close at the end of 2008, a hard deadline. The last such transition completed in 1990, the next one cannot be expected before 2050.

The three contenders for the new 64-bit standard are Windows-64, MacOS X, and Linux. The winner will be determined by desktop market share, the bulk of which consists of non-technical end users.

[Wudang]

the bulk of which consists of non-technical end users.
Who think that 64 bits for a computer sounds pretty cheap.

[QueigBladecaster]

I am an IT systems engineer working in the UK, and I can certainly tell you that many companys are starting to embrace 64-bit hardware and more importantly linux and solaris in preference to windows for server side applications.

But that is a whole other world then getting linux established as a home PC or even a business workstation platform.

Its just too complicated and user unfriendly and not to mention confusing with the huge number of distros available atm for it to be accepted by the home market.

Its going to take a major shift for people to move away from windows. people know it, they are familiar with it and there is tons of software for it. Most casual PC users dont want to look at what is the best OS they want whats easiest for them to use and what they feel comfortable with. Microsoft has the drop on this. there system is attractive, very easy to learn universially supported and most PC users know that is what most of their friends are using.

Most people who use Apple will tell you til they are blue in the face that MacOS is a far superior OS and in many ways they are right. But the pure fact that is less widely used by the average person means that its less likely to be accepted as the defacto standard home OS. And this is a system that is just as easy to use and some say more attractive then the current defacto standard.

Looking at that you can see how huge the mountain is that linux has to climb. The number of people the average person will know who uses Linux is even smaller. The fact that it is so confusing to even choose which distro to go for will put people off. Add to that the less then intuitive way of configuring the systems and adding your own applications, and the fact that you probably will not be able to use that new and cool application that most of your friends are raving about, means that I suspect Linux will stay firmly in its niche market for some time to come.

[cyborg]

Its going to take a major shift for people to move away from windows.

:rolleyes: You need to read the article. Your whole reply reads as if you didn't because it sounds like the standard, and wrong, reasons why Linux/Mac OSX isn't king.

[Jekyll]

What difference does moving to 64-bit code make to the average user?

They're not going to care that the underlying infrastructure has changed, and they'll quite happily keep using widows because they were happy with it before. As long as Microsoft don't screw up windows64 to much no one will notice and there will just be a seamless transition from one windows OS to the next in the market.

Sure, there might be problems running old code, but I don't see that driving Joe Bloggs to linux.

[cyborg]

I see nobody is reading the article.

I highly recommend you do before replying - it does cover these issues in some detail.

For example, "As long as Microsoft don't screw up windows64 to much" is covered. Read it.

[Jekyll]

I see nobody is reading the article.

I highly recommend you do before replying - it does cover these issues in some detail.

For example, "As long as Microsoft don't screw up windows64 to much" is covered. Read it.

I did. His conclusion is Microsoft will screw up because their code base is so big. I am not convinced.

[cyborg]

That's not the conclusion.

The conclusion is that there are several very real organisational and technical problems that are hindering a Windows foray into 64-bit. All the stored money in the world isn't going to make that right by magic. It is simply wrong thinking to assume that their dominance today means a damn thing when the whole point is that history shows otherwise. Microsoft have very little time to turn around their massive corporate inertia towards 64 bit and their current market dominance and code base doesn't help them at all. Dominance of the 32 bit world has almost certainly left them short-sighted to the impending 64 bit change.

[QueigBladecaster]

I have read it and I dont see anything that disagrees with what I have said in fact I quite liked


The general populace thinks Firefox is an alternative operating system, if they understand the concept at all. It's not that they couldn't understand the difference between Windows and Linux if they tried, but that's what they see, and they actively don't want to know anything more. They only notice what's underneath when it fails, or when it needs service, or when it costs them money. Waxing eloquent about the implementation, licensing, or development model is like touting the benefits of dual overhead cams with a manual transmission. If grandmother ever has to pop the hood, she has the wrong car.


All the arguements within the link are perfectly valid and I aggree with the majority of it but it doesnt allow for human nature and natural apathy.

The document is targeted at and written by people who understand Linux. People for whom its a no brainer, of course we should be using it its so much better because...... (insert favorite reason here). But people want to sit at a PC turn it on do what they expect to be able to do quickly and without having to think too much. You just dont get that experience with Linux or MacOS unless you are already used to using Linux and MacOS

There is currently no major driving force for the average user to move away from windows. Even if everything recommended in the paper were done (all of which is perfectly possible) I still say it will take a good 3 years to get Microsofts hooks out of people. There are people still quite happily using windows 98 who will swear by it and are positively dismayed by the thought of giving it up.

In the UK I have only come accross one organisation that has implemented Linux desktops as standard and thats Novell, their offices now pretty much all use there own linux workstation distro. I personally have been involved in evaluation studies for 3 different companys who wanted to use it. Each time it was decided not to adopt purely on the amount of retraining involved, and the difficulty of getting employees to feel comfortable with the system. This is the main reason that business does not take it up. Even when you prove that the TCO including all the retraining involved is a substantial saving, business does not want to loose the comfort zone that windows gives them.

The single biggest thing that will inprove the situation is for systems to be sold with Linux installed, configured and working as soon as the PC is first turned on this turns apathy to its advantage. currently people have to WANT to switch to linux and they have to put the effort in to get there.

For it to succeed you have to create a situation where they have to use it unless they WANT to move to something else, This is what Microsoft has managed to do so successfully.

[cyborg]

All the arguements within the link are perfectly valid and I aggree with the majority of it but it doesnt allow for human nature and natural apathy.

Quite the opposite which I why I think you are missing the point.

The document is targeted at and written by people who understand Linux. People for whom its a no brainer, of course we should be using it its so much better because...... (insert favorite reason here).

I absolutely fail to see how you get that impression.

But people want to sit at a PC turn it on do what they expect to be able to do quickly and without having to think too much.

Again I don't see how you could fail to grasp that the article covers this if you read it.

You just dont get that experience with Linux or MacOS unless you are already used to using Linux and MacOS

Arguments form 'Windows does it X way' doesn't translate into 'Linux and MacOS' are hard to use per se. Besides, you don't get that experience with Windows either. That's a nonsense - there are just as many (far more really but I'm being generous) problems there as anywhere else. They're just problems people are familiar with.

There is currently no major driving force for the average user to move away from windows.

Which is the entire crux of the article - 64 bit IS that force.

Even if everything recommended in the paper were done (all of which is perfectly possible) I still say it will take a good 3 years to get Microsofts hooks out of people. There are people still quite happily using windows 98 who will swear by it and are positively dismayed by the thought of giving it up.

I don't think you are grasping that 2008 is not a swithover deadline where people will magically start using the new king, it is where the network effects will start an inexorable rise to dominance.

business does not want to loose the comfort zone that windows gives them.

That comfort zone is dying. That is the point. There is a new battle on and there is no way in which Windows is a default winner because their current market means squat.

The single biggest thing that will inprove the situation is for systems to be sold with Linux installed, configured and working as soon as the PC is first turned on this turns apathy to its advantage. currently people have to WANT to switch to linux and they have to put the effort in to get there.

That is also covered by the article! Damn! I really don't believe you read it.

For it to succeed you have to create a situation where they have to use it unless they WANT to move to something else, This is what Microsoft has managed to do so successfully.

Sigh. I know. This is covered. This is why 64 bit is important.

[Jekyll]

That's not the conclusion.

The conclusion is that there are several very real organisational and technical problems that are hindering a Windows foray into 64-bit. All the stored money in the world isn't going to make that right by magic. It is simply wrong thinking to assume that their dominance today means a damn thing
I'm sorry but it obviously does. Your man in the street isn't going to find a transition from 32 to 64 bit any more significant than the duo 2 coming out.

Microsoft need to screw up to let other people in through the door, and your article underestimates them, and over estimates the importance of legacy device drivers. Most people don't use old kit with their new desktops, and Vista's printer driver problems don't seem to have driven anyone over to the darkside yet.

[Wudang]

I did read the article. I fail to see any case for 64-bit computing being a great sea change. I thought the section on drivers was naive in the extreme and ignored the historical lessons of the Windows-OS/2 war.

[cyborg]

I'm sorry but it obviously does. Your man in the street isn't going to find a transition from 32 to 64 bit any more significant than the duo 2 coming out.

It DOES NOT assume that they do. Quite the opposite in fact.

Microsoft need to screw up to let other people in through the door, and your article underestimates them,

I do not see that. The estimation appears accurate for a highly entrenched corporation. It's going to require some sort of miracle of vision because currently it doesn't matter how talented the techs are they are hand tied today and there's no reason to see that changing real soon.

and over estimates the importance of legacy device drivers. Most people don't use old kit with their new desktops,

That is a weaker assumption yes.

and Vista's printer driver problems don't seem to have driven anyone over to the darkside yet.

Why would they? Most people aren't using Vista. And if they are it's 32 bit Vista - and as such I feel people are still kinda missing the point...

[cyborg]

I did read the article. I fail to see any case for 64-bit computing being a great sea change.

Why? It clearly is.

No, the end user doesn't care about 64 bit vs 32 bit.

They do care about WHAT 64 bit can do that 32 bit CAN'T even if they don't know why there is a difference. They cared when it was 16 vs 32, and 8 vs 16. Knowing what the numbers mean is missing the point.

I thought the section on drivers was naive in the extreme and ignored the historical lessons of the Windows-OS/2 war.

Explain.

[Jekyll]

Why would they? Most people aren't using Vista. And if they are it's 32 bit Vista - and as such I feel people are still kinda missing the point...

The main complaint with people running vista seems to be a lack of support of printer drivers.

The main response from these people is to wait for it to sort itself out, not to switch to linux.

[QueigBladecaster]

The conclusion is that there are several very real organisational and technical problems that are hindering a Windows foray into 64-bit. All the stored money in the world isn't going to make that right by magic. It is simply wrong thinking to assume that their dominance today means a damn thing when the whole point is that history shows otherwise. Microsoft have very little time to turn around their massive corporate inertia towards 64 bit and their current market dominance and code base doesn't help them at all. Dominance of the 32 bit world has almost certainly left them short-sighted to the impending 64 bit change.

Microsoft could (and have) completely screw up implemetation of new technology and people would still stick with it.

I dont see how historical changes can be used as a comparison. We have never had a time when computer use has been so integrated into peoples lives. The shift to 32 bit technology is not even the same playing field as the shift from 32 to 64 will be.

Arguments form 'Windows does it X way' doesn't translate into 'Linux and MacOS' are hard to use per se. Besides, you don't get that experience with Windows either. That's a nonsense - there are just as many (far more really but I'm being generous) problems there as anywhere else. They're just problems people are familiar with.


Its the fact that they are problems that people are familiar with which is what keeps them tied to windows. If you had a total novice learning to use a PC for the first time I personally feel that that could learn more and faster using a Mac.



Quote:
But people want to sit at a PC turn it on do what they expect to be able to do quickly and without having to think too much.

Again I don't see how you could fail to grasp that the article covers this if you read it.
Quote:
The single biggest thing that will inprove the situation is for systems to be sold with Linux installed, configured and working as soon as the PC is first turned on this turns apathy to its advantage. currently people have to WANT to switch to linux and they have to put the effort in to get there.

That is also covered by the article! Damn! I really don't believe you read it.


Quote:
For it to succeed you have to create a situation where they have to use it unless they WANT to move to something else, This is what Microsoft has managed to do so successfully.

Sigh. I know. This is covered. This is why 64 bit is important.



I never said any of this wasn't covered in the article.

I am not arguing against the article, I just think the power of human apathy has been underestimated.

The is still a long way to go before Joe Public even realises that there is a benefit to 64 bit technology. I take you point about this not being a switchover and this being the start of the process, I aggree. I just dont think it will be an easy or a fast process

[cyborg]

The main response from these people is to wait for it to sort itself out, not to switch to linux.

*Sigh* I really, really don't think you get the point here. You are not arguing against the case made in the article. People DO NOT switch OSes on a whim. That is why 2008 is important because it represents one of the turning points where it is actually possible for a new OS to displace Windows-32.

[Wudang]

Why? It clearly is.


I've worked on changes from 24 to 31 to 64 bit addressing on MVS/os390/zos so I'm quite familiar with the technology, thanks. The article and you take it as written that users will care. You do not provide any argument to support this.

Manufacturers focussed on the market that was there. MS sold them on what windows 95 would do, it was where the market was so they produced the drivers. IBM paid many companies to produce drivers for OS/2 and many of them were crap.
And what will the TV ads be telling people? You need a 64-bit chip? Or will they talk about the "experience"? Were you around when MS launched win95? All that fuss about not very much. A triumph of style over substance.

[Jekyll]

*Sigh* I really, really don't think you get the point here. You are not arguing against the case made in the article. People DO NOT switch OSes on a whim.

I agree with this, but it just seems kind of whimsical to switch desktop environments because a new chip has come out.

[cyborg]

I've worked on changes from 24 to 31 to 64 bit addressing on MVS/os390/zos so I'm quite familiar with the technology, thanks. The article and you take it as written that users will care. You do not provide any argument to support this.

The article is not written as if users care about the technology. I do not understand how you can get that - it is pretty implicit about users not caring, since both MacOS X and Linux are objectively better technologies than Windows. Having a superior product is irrelevant. The article is written in this context - from the 101 document.

Manufacturers focussed on the market that was there. MS sold them on what windows 95 would do, it was where the market was so they produced the drivers. IBM paid many companies to produce drivers for OS/2 and many of them were crap.

I fail to see the analogy. Linux doesn't really have technical problems - as the article says the major problems are legal and logistical.

And what will the TV ads be telling people? You need a 64-bit chip? Or will they talk about the "experience"?

The later. But this doesn't change the fact they're going to be getting 64-bit does it? The point is that despite the perceived wisdom otherwise it is simply not clear cut that MS is going to win by default.

Were you around when MS launched win95? All that fuss about not very much. A triumph of style over substance.

Indeed. But I think it is wrong to conclude that ultimately that is what made W95 successful. There is a strong case here that the network effects had already taken care of that through Win3.1.

[cyborg]

I agree with this, but it just seems kind of whimsical to switch desktop environments because a new chip has come out.

But that is NOT what the article is suggesting. The thrust is network effects. As such it is NOT implying that in 2008 everyone will switch to a new OS. It is saying that in 2008 the winner will already have been decided.

[MortFurd]

One big point that the article was trying to make (and which obivously flew over most people's heads and sunk in the Pacific) was that a 32 bit OS is not capable of using the resources that the 64bit architecture of the new machines can provide - most specifically, 32 Vista will not be able to make good use of RAM above 4GB.
PCs are currently being delivered with 1GB installed, some people need more. 2GB is good, but at 4GB on 32 bit hardware you start having problems such that large portionsof the 4th GB aren't usuable with 32 bit hardware.

The new 64Bit hardware lets you make effective use of more than 4GB RAM, BUT a 32 bit OS on 64 bit hardware can't deliver that additional RAM to user programs.

The 2008 deadline is the point at which a 32 bit OS (Windows Vista 32) can't effectively make use of commonly purchased hardware (64bit architecture with more than 4GB RAM.) This is the point at which joe average will start noticing the problem - and the high end users (those who are buying 4GB systems today) will have been dealing with the problem for years.

Those users who need to use more than 4GB RAM today are having to face the choice of using Windows 32 (which doesn't let them make full use of their RAM,) using Windows 64 (which has problems,) or using Linux 64 (which handle 64 bit hardware well, but lacks some of the spiffy programs like iTunes.)

That's what it was all about. The article is saying that the hardware has outstripped the OS, and the Microsoft doesn't have an effective replacement OS in hand.

That situation opens the possibility that the market leadership will change. The article then lays out the requirements for Microsoft, Linux, or Mac OSX taking the lead in the 64 bit era. Those requirements then include many of the arguments people are trotting out against Linux and OSX.


We could now go on to argue about whether or not PC manufacturers will actually deliver hardware (64bit systems with more than 4GB) that outstrips the capabilities of the OS. I think they will, as to do otherwise would break their beloved upgrade cycles and cut their sales. They've preached "this year's model is bigger and better than last year's" for so long that customers may just fail to upgrade if this year's model really isn't any bigger or better.

[Jekyll]

But that is NOT what the article is suggesting. The thrust is network effects. As such it is NOT implying that in 2008 everyone will switch to a new OS. It is saying that in 2008 the winner will already have been decided.

It relies on the idea that switching to a new chip will open up the desktop market. This will only happen if people start looking for alternatives to windows as they switch to 64 bit.

[cyborg]

Jekyll - you miss the point.

YOU MUST LOOK FOR AN ALTERNATIVE TO WIN32.

Windows 32 is useless for 64 bit.

People need to stop thinking that Windows 32 == Windows 64.

I'm glad to see that someone gets the point MortFurd.

[Almo]

Linux is still too hard to use for the average user.

[cyborg]

Linux is still too hard to use for the average user.

-1 Redundant.

:rolleyes:

[QueigBladecaster]

Those users who need to use more than 4GB RAM today are having to face the choice of using Windows 32 (which doesn't let them make full use of their RAM,) using Windows 64 (which has problems,) or using Linux 64 (which handle 64 bit hardware well, but lacks some of the spiffy programs like iTunes.)


When I was still a saleman rather than an engineer I would always be amused at the windows 95 users paying out for extra memory that their OS just couldnt use.

For the home market most of the bigger, better, faster number chasing is driven by games software which is still very windows orientated (another big hurdle for linux) And even in that arena the trend is away from PC's to consoles.

Im actually having a hard time thinking of a killer app that would make the home user want to shell out for a specific 64 bit platform that isnt likely to be taken over by the console market.

[Jekyll]

Jekyll - you miss the point.

YOU MUST LOOK FOR AN ALTERNATIVE TO WIN32.

Windows 32 is useless for 64 bit.

People need to stop thinking that Windows 32 == Windows 64.

I'm glad to see that someone gets the point MortFurd.

No, you still don't get it. No one on this tread thinks that win32 = win64. What I have been saying is that consumers won't care about the differences. They'll just keep using windows (of what ever flavour) unless Microsoft cock up win64 so much that it entirely unusable.

[QueigBladecaster]

Almo's comment is not redundant.

If Linux is to make any headway at all in this battle for the 2008 deadline then this is exactly the kind of perception that needs to be fought.

[voidx]

I do not understand how you can get that - it is pretty implicit about users not caring, since both MacOS X and Linux are objectively better technologies than Windows.
Define better? I find many people are a touch ignorant on the real breadth of the "Windows" platform. Are you talking desktops or servers? And if it comes to Servers, lets look at business level productivity suites. The list is huge, but I think you'll find that Microsoft dominates a lot of it. Many people mistakenly think that the OS / Desktop alone is what drives business and ultimately Microsofts market share. Business use a wide range of products, that ultimately drive the decision of what platform, or mix of platforms they use. On top of this desktop / Server OS and the run of the mill Office suites there exists database backends, enterprise desktop management, group policies, CRM, virtualization, etc.

In this sense, I'm sorry, Microsofts market share very much DOES matter. I'm not saying that the impact of 64bit hardware outstripping the current OS's won't open up opportunities. But this impression that Microsoft is lumbering and in grave danger seriously under-estimates a company that has been smart enough to dip its fingers into a vast range of business, home use, and entertainment applications.

What you'll also find is that hardware issues tend to become lessened on corporate desktops simply because the vast majority of users do not require the latest and greatest to send email, use Word and surf the net. And as a result, most companies don't go out of their way to setup the standard user with a whiz bang ultra-leading edge PC, because they don't feel like paying for it.

[cyborg]

Almo's comment is not redundant.

It is entirely redundant:

1) It is covered by the article.
2) It is covered in this thread.

If Linux is to make any headway at all in this battle for the 2008 deadline then this is exactly the kind of perception that needs to be fought.

No, it's not. Most people have no perception of Linux. Most people don't even really know what it is - even those who think they do.

[cyborg]

Are you talking desktops or servers?

The underlying technology is better either way. Linux is simply a better operating system technology - *nix is a better technology: period.

Many people mistakenly think that the OS / Desktop alone is what drives business and ultimately Microsofts market share.

Well that is certainly not what I nor the article is implying.

In this sense, I'm sorry, Microsofts market share very much DOES matter.

No it doesn't. The upheaval in question negates the market share - if you are forced to change and you cannot simply upgrade then you are free to evaluate all possibilities.

This is the entire thrust of the article.

But this impression that Microsoft is lumbering and in grave danger seriously under-estimates a company that has been smart enough to dip its fingers into a vast range of business, home use, and entertainment applications.

Have you seen how poorly they are doing in this vast range of businesses? I think people are seriously over-estimating how important being a big lumbering giant is. One only needs look to IBM to see how quickly a technology giant can go from ubiquitous to irrelevant. They had to completely reinvent themselves to get back on track.

MS is currently plodding on regardless. I'd be worried if I were a shareholder.

[QueigBladecaster]

No, it's not. Most people have no perception of Linux. Most people don't even really know what it is - even those who think they do.Today 06:06 PM

But surely the only way for Linux to become more widely used and supported is for more people to know of it and understand why its better for them and their bright shiney new 64 bit PC platform.

If Linux is to "win" its not only a major problem for it that few people know what it is, its also a fairly big issue that there are perceptions like almo's.

You can give it all the hardware drivers in the world if you like if the prevailing perception of the product is "its more difficult to use then what I currently use" people will just not use it.

Linux not only has to be better or at least as good, it has to deal with the Microsoft spin machine.

[cyborg]

But surely the only way for Linux to become more widely used and supported is for more people to know of it and understand why its better for them and their bright shiney new 64 bit PC platform.

No. The way for Linux to be more widely used is for it to become the de facto preinstalled OS that can actually make use of your shine new 64 bit PC.

As has already been pointed out people just don't change their OSes on a whim. People do not choose Windows any more than they choose not to use Linux.

If Linux is to "win" its not only a major problem for it that few people know what it is, its also a fairly big issue that there are perceptions like almo's.

almo's perception is based on ignorance - he conflates OS with application. If one is going to consider Browser + Word Processor = Computer, because frankly that encompasses a great majority of users, then I fail to see how one can make a cogent argument that Linux is fundamentally harder to use compared to Windows when the applications run on both in identical ways.

There are some bad examples of UI in Linux land. There are also plenty in Windows land. The main UI issues as far as Linux goes are to do with administrative issues, because traditionally there has been no need for non-technical people to have anything to do with this in user land.

Of course he didn't bother to read the article which goes into this but instead just did a knee jerk, "Linux is hard."

Linux not only has to be better or at least as good, it has to deal with the Microsoft spin machine.

And it is. By producing better software.

[QueigBladecaster]

A man goes into PC world looks at 2 dell computers, both have the same CPU, memory, etc, etc. They are the same price, in fact identicle in every way except for the software.

One is installed with Windows64 one has Linux 64.

The man doesnt consider himself computer literate but sometimes uses it to work from home, looks at the internet, writes the odd document and keeps records of his coin collection on it.

He has been using windows XP for the last 5 years and is in the main happy that he can do the work on it that he wants to do. He is only changeing his computer in the first place because his last one is broken and the parts for it are no longer in production.

He has never used Linux but has heard rumours that it is difficult to use and just for techies.

Which one do you think he buys?



Incorrect assumptions, plain ignorance and knee jerk reactions are the very things Linux has to counter if its to convince hardware manufacturers to install it as default on their machines.

If customer surveys show that people are less likely to buy a system preinstalled with linux rather then windows then the manufacturers will choose to bundle windows. It makes no difference how far from the truth the public opinion is that is what they will sell, and therefore what the customers get cornered into buying.

[cyborg]

A man goes into PC world looks at 2 dell computers,

Dell computers in a computer shop?

Umm... okay.

One is installed with Windows64 one has Linux 64.

For the same price? No.

He has never used Linux but has heard rumours that it is difficult to use and just for techies.

Which one do you think he buys?

If the man is in the damn shop he can:

1) ASK
2) TRY

:rolleyes:

Besides, one could easily adjust your rumour to:

"I've heard there's this software that runs a PC better, doesn't crash as much, and is free."

This whole, "Windows is familiar" argument breaks down completely with Vista. The consumer isn't getting any benefit from familiarity and Windows has a poor history when it comes to how it makes people perceive computers - as flaky things.

Incorrect assumptions, plain ignorance and knee jerk reactions are the very things Linux has to counter if its to convince hardware manufacturers to install it as default on their machines.

No, it is not. You severely overestimate the visibility of Linux to the general user-base. This is really not the problem you think it is. Nor is the FUD necessarily as effective as you think.

If customer surveys show that people are less likely to buy a system preinstalled with linux rather then windows then the manufacturers will choose to bundle windows.

The public don't want Windows or Linux or Mac. They want to DO things. They will choose whatever it is that lets them do this.

Besides MS has generated enough bad will with their Window's tax that manufacturers are more than willing to look at alternatives.

[Soapy Sam]

If Win 63.9 is crap
and
If Linux 64 is not,
Then
Microsoft will buy a Linux 64 distro, relabel it Windows 64.1GTi
And the market will plug on as usual.

In short, history will repeat itself.

[drkitten]

Jekyll - you miss the point.

YOU MUST LOOK FOR AN ALTERNATIVE TO WIN32.

Windows 32 is useless for 64 bit.

The problem is -- there is one. It's called Windows 64.

ESR believes that Windows 64 will not be commercially successful.

If he's right, then something other than Windows 64 will be, almost by definition. But none of the reasons that he's cited are, in my view, credible. He suggests that the existing Windows codebase is so large as to be unmaintainable, which is directly contradicted by the fact that Windows Vista does exist, bad as it is. The idea that Microsoft will not be able to delliver something functional as a 64-bit operating system is ludicrous. It may not be best of breed, but it will work.

So the question becomes -- will Windows 64 suck badly enough to drive the customers away? And given that most customers neither need nor notice the OS-level features of their systems, the answer is probably, "no, it won't suck that badly."

[cyborg]

If Win 63.9 is crap
and
If Linux 64 is not,
Then
Microsoft will buy a Linux 64 distro, relabel it Windows 64.1GTi
And the market will plug on as usual.

In short, history will repeat itself.

Um, in a Linux world MS cannot maintain dominance. This is why OSS scares the bejesus out of them.

You are talking about a situation in which MS goes from giant to irrelevance. You do realise how devastating that would be to their market position right?

[cyborg]

which is directly contradicted by the fact that Windows Vista does exist, bad as it is.

I fail to see how all the many problems with getting Vista out contradicts that. If anything the numerous dropped promises, slippages AND shoddiness of the final product suggest otherwise.

The idea that Microsoft will not be able to delliver something functional as a 64-bit operating system is ludicrous. It may not be best of breed, but it will work.

Win98 anyone? That was broken from day one. They may be able to get something out of the door and claim 64 bit but actually working is another thing.

[aerosolben]

Have you seen how poorly they are doing in this vast range of businesses?
Funny, IDC is reporting that Windows Server market share continues to grow faster than the server market (http://home.businesswire.com/portal/site/google/index.jsp?ndmViewId=news_view&newsId=20070226006224&newsLang=en),and has been doing so consistently for at least 4 years now. In fact, in 2005, it took the top spot from Unix.

SQL Server continues to gain market share (http://arstechnica.com/journals/microsoft.ars/2006/5/31/4163) as well (although Linux and open source solutions are also a growth market in this space). I could probably go on.

You may have your opinions about the quality of the product, but your misconceptions about the success of Microsoft's server business are naive.

[geni]

Um, in a Linux world MS cannot maintain dominance. This is why OSS scares the bejesus out of them.

Not really software patents would allow them control prety much indefenitely.

Ever tried converting a .wmv file to a .ogg with Theora codec?

How may people here can even play this file:

http://commons.wikimedia.org/wiki/Image:Blender3D_BouncingSoftbodyBall.ogg

[aerosolben]

Win98 anyone? That was broken from day one. They may be able to get something out of the door and claim 64 bit but actually working is another thing.
Even assuming you're correct, this doesn't even remotely point to the failure of Windows as a platform.

[Soapy Sam]

Um, in a Linux world MS cannot maintain dominance. This is why OSS scares the bejesus out of them.

You are talking about a situation in which MS goes from giant to irrelevance. You do realise how devastating that would be to their market position right?

No, I'm not.
Windows , to the average user , is Windows, because it says "Windows" on the box. The average user does not know or care if it is 16 bit, 32 bit, 64 bit or 126.4 bit. Nor does he care if it uses FAT or NTFS or some other file system.
He cares only that it is similar enough at the GUI level to the version he is already familiar with, that he can do his job/ save his files/rip his CDs.

Does the average user know the difference between Win9X and Win NT?
Like hell.

What I'm saying is that if M$ needs to buy and rebadge Linux 64 as Windows 64, they will do so and business will carry on as usual.

For the technerds, that will be a day to rejoice that M$ has embraced the wonder of Linux, or to grieve that Linux has gone over to the dark side.

To the 99% of people who don't give a damn either way, it will make no difference at all.

[drkitten]

I fail to see how all the many problems with getting Vista out contradicts that. If anything the numerous dropped promises, slippages AND shoddiness of the final product suggest otherwise.

Let's see. You suggest that because Microsoft can't produce a good quality product, they won't be able to ship something that sucks and have people buy it anyway.

Microsoft Vista sucks -- and people buy it anyway.

That's a pretty firm contradiction.





Win98 anyone? That was broken from day one.

Yeah, so what else is new. And despite being broken, people bought it anyway.


They may be able to get something out of the door and claim 64 bit but actually working is another thing.

On the contrary. Windows 98 works. It doesn't work well, but it works. It works well enough to have held a 37% market share on the end-user desktop in 2002.

What's going to stop the same 37% of people who bought a lousy, broken, operating system in 2002 to buy a lousy, broken, operating system in 2008? They'll buy the same lousy, broken, operating system for the same reason -- because it says "Windows" on the box, and because there's really fundamentally no reason at the user level for them to change.

Windows 64 will work. You won't like how it works. But it will work.

[Stainless_Steel_Rat]

Linux certainly COULD take control of the market. That's a long long ways from WILL though. Certainly the article makes good points.

There are so many hurdles it must go through though. It has to do everything Windows can do. :) I'm a pretty high end computer user, and the first thing I'm going to ask is, "Will it play [insert name of shiney new game]?" "Will I be able to do everything I could do before?"

Now it's certainly possible that we'll be able to say yes to questions like these, but what's the possibility that this will happen before next year? Linux's biggest strength is also its weakness. There's no "leader" telling people what they have to do. It's people just deciding what they want to do. This causes redundencies and gaps in what gets developed. *shrug*

Interesting read, but I doubt that Linux developers can organise well enough to take advantage of the envisioned opportunity.

SSR
*Edit* forgot to add that I'm a XP64 user currently, and have seen no compelling reason to change as of yet* I also run an Ubuntu (speel?!?) box occassionaly, but don't have the time/inclination to figure out much on it.*

[cyborg]

He cares only that it is similar enough at the GUI level to the version he is already familiar with, that he can do his job/ save his files/rip his CDs.

Are people even reading my replies because it seems like they are not.

What I'm saying is that if M$ needs to buy and rebadge Linux 64 as Windows 64, they will do so and business will carry on as usual.

Please explain how MS is supposed to buy Linux.

(Hint: they cannot).

If MS releases an OS based on Linux their dominance is over. It cannot be business as usual in this situation. They cannot possibly maintain their current revenue streams this way. It is game over.

Not really software patents would allow them control prety much indefenitely.

PEOPLE ARE NOT ***** READING THE ARTICLE ARE THEY?

I suggest you look towards the bottom where the codec/patent issue is covered.

Microsoft Vista sucks -- and people buy it anyway.

You are aware that sales are not what MS would hope for right?

They'll buy the same lousy, broken, operating system for the same reason -- because it says "Windows" on the box, and because there's really fundamentally no reason at the user level for them to change.

It is a simple fact that most people simply have the OS that their computer came with.

[Soapy Sam]

Microsoft Vista sucks -- and people buy it anyway.

Out of interest Dr- what makes you say that Vista sucks?

[geni]

Please explain how MS is supposed to buy Linux.

(Hint: they cannot).

That isn't quite true and in any case they don't have to. Given the amount of cash MS have on hand building a linux style OS from scratch would not be imposible.



I suggest you look towards the bottom where the codec/patent issue is covered.


I did. Handwaveing and the hope the BBC will rescuce them. The abandoment of Dirac (pity it looked like quite a nice system and for once patents didn't appear to be an issue) and the more recent deal with youtube should kill that one off quite nicely.

[cyborg]

That isn't quite true and in any case they don't have to. Given the amount of cash MS have on hand building a linux style OS from scratch would not be imposible.

Again, if MS makes a Linux OS it is game over for them. Their business model CANNOT support this situation.

This is not a solution to the problem MS is facing - it is admitting defeat. MS producing a Linux OS is the end for their dominance - it's as simple as that.

I don't know what magical mechanism you people think MS branding a Linux distro would allow the status quo.

ETA: A *nix style OS is supposedly what NT is supposed to be - or at least something that would kill it.

I did. Handwaveing and the hope the BBC will rescuce them.

That was not even slightly the conclusion.

[Wudang]

Hells teeth, I feel like I'm back in the 90's on comp.os advocacy group. Look back at the history of NT and the pile of festering dingos kidneys it was before they hired one of the architects of VMS to re-architect the kernel. NT was a POS until , what, 3.5? Look how long it lasted.

[Dave1001]

Interesting and well researched article I am sure will interest everyone here:

World Domination 201 (http://www.catb.org/~esr/writings/world-domination/world-domination-201.html)

The next hardware transition can't be expected before 2050? How did they come to that number? It sounds counter-intuitive to me.

[jeremyp]

The article loses credibility for me right here:

The three contenders for the new 64-bit standard are Windows-64, MacOS X, and Linux. The winner will be determined by desktop market share, the bulk of which consists of non-technical end users. (my italics).

Mac OS X will not be the dominant 64 bit operating system because, in order to run it you need to buy a Macintosh. This is not going to change because many of the advantages of Mac OS X depend on Apple's ability to control the hardware, and, of course, Apple sees itself as a hardware company and OS X as being part of what is needed to make the hardware run.

Also Mac OS X is currently a 32 bit operating system with support for 64 bit user processes. As of now, the kernel and drivers are 32 bit. I understand 10.5 will be addressing this.

I find it quite amusing that the article has a section called "Surviving the Killer App", because ultimately, the killer app is what it's all about. Too many people forget that the operating system is only a means to an end, it's a necessary evil. If you could run Microsoft Office or World of Warcraft on the bare hardware, there would be no operating system. Currently for PCs, killer apps fall into two categories: Microsoft Office and PC games. Linux is not going to dominate the desktop until a) Microsoft port Office to it and b) WoW (or whatever other popular game) becomes available for it.

[geni]

Again, if MS makes a Linux OS it is game over for them. Their business model CANNOT support this situation.

Linux like. Not linux itself I think it is fairly safe to say that microsoft view the GPL as something that happens to other people. Given the ammount of effort that has been put into porting MS stuff to linux it would appear that microsoft could survive under those conditions.

Certianly macs appear to not have been killed off by Free BSD.


That was not even slightly the conclusion.

There was a conclusion? Describe it in your own words.

[cyborg]

Linux like. Not linux itself I think it is fairly safe to say that microsoft view the GPL as something that happens to other people.

I'm not quite sure how you think they're going to get something 'Linux like' in such short order and how that's going to overcome the fact that everyone else has a massive head start.
And it's not merely a case of technical difficulties - any such a move is basically admitting that all the FUD and such they've been spewing is nonsense. It is a severe loss of face. Any top man in MS proposing such an approach will have to have huge balls. MS hasn't demonstrated that it has that sort of vision.
That scenario is a marketing manager's worst nightmare to deal with.

Given the ammount of effort that has been put into porting MS stuff to linux it would appear that microsoft could survive under those conditions.

You don't seem to understand.

It's game over under these conditions. They lose dominance. They cannot possibly keep their revenue streams as high as they need. The house of cards will tumble down under this situation.

Survive? Sure, IBM survived their big mistake. But it will knock the crap out of MS if that happens.

There was a conclusion? Describe it in your own words.

Namely that Linux MUST HAVE THE CODECS BEING USED, MP3, DVD etc... Quite the opposite of, "only open source codecs please". How you could miss that part unless you are selectively reading is beyond me.

[Wudang]

NT
VMS like. Not VMS itself

[cyborg]

Mac OS X will not be the dominant 64 bit operating system because, in order to run it you need to buy a Macintosh. This is not going to change because many of the advantages of Mac OS X depend on Apple's ability to control the hardware, and, of course, Apple sees itself as a hardware company and OS X as being part of what is needed to make the hardware run.

*Sigh*

If you had bothered to read the Mac OS X analysis you would have seen that this is covered.

Currently for PCs, killer apps fall into two categories: Microsoft Office and PC games. Linux is not going to dominate the desktop until a) Microsoft port Office to it and b) WoW (or whatever other popular game) becomes available for it.

Games are not as important as you think - consoles are far and away more popular.
Nor is Office a problem either. There's few desktop users in business or at home who need more than 10% of the features of MS Office. These are all in the alternatives.
The Office suite doesn't have to be MS Office.

Either way you don't seem to have gotten the fact that this chicken and egg scenario is covered. I don't see why you think it's funny that the article has the title.

[Wudang]

Nor is Office a problem either. There's few desktop users in business or at home who need more than 10% of the features of MS Office. These are all in the alternatives.

Sorry but it is a problem and I say this as someone who's used a lot of different office/DTP/etc software. As a real example, I was able to produce my CV in Open Office, looked great. Viewed it with the free MS word viewer tool, looked a bit sloppy. Loaded it in word and it looked crap, margins just that little bit misaligned etc. Businesses care about those little details.

[cyborg]

Sorry but it is a problem and I say this as someone who's used a lot of different office/DTP/etc software. As a real example, I was able to produce my CV in Open Office, looked great. Viewed it with the free MS word viewer tool, looked a bit sloppy. Loaded it in word and it looked crap, margins just that little bit misaligned etc. Businesses care about those little details.

Hell, I've had such problems between versions of Word. Still, conversion IS possible even if adjustments are required and of course it matters not a jot here if you take out the culprit causing the trouble in this chain (i.e. MS Office).

[Wudang]

Yes, but businesses don't care that conversions are possible, they want the most cost effective solution that meets their requirements. And how many of their staff are going to be people who freak out at those adjustments?
So you take out MS office? So you factor in the cost of auditing all the files that may need to be converted, possible business impact, legacy files due to SOX regulations, consistency with business partners etc.
Possible isn't the issue, transparent is.

[JamesM]

The next hardware transition can't be expected before 2050? How did they come to that number? It sounds counter-intuitive to me.
It's the estimated amount of time before hardware will have sufficient RAM beyond that addressable with a 64-bit architecture.

[cyborg]

Possible isn't the issue, transparent is.

Yes, but you are forgetting: this is a technical problem, as such it is in the easiest of categories to fix and also in the category of problems OSS people are best at fixing.

Technical problems are not problems, so to speak. Once identified they'll be fixed in short order.

[geni]

I'm not quite sure how you think they're going to get something 'Linux like' in such short order and how that's going to overcome the fact that everyone else has a massive head start.


By throwing money at it.


And it's not merely a case of technical difficulties - any such a move is basically admitting that all the FUD and such they've been spewing is nonsense. It is a severe loss of face. Any top man in MS proposing such an approach will have to have huge balls. MS hasn't demonstrated that it has that sort of vision.
That scenario is a marketing manager's worst nightmare to deal with.


Nah pretty trivial really. You market it as a complete redesign of the code base in order to give more choice to the consumer.


You don't seem to understand.

It's game over under these conditions. They lose dominance. They cannot possibly keep their revenue streams as high as they need. The house of cards will tumble down under this situation.

Survive? Sure, IBM survived their big mistake. But it will knock the crap out of MS if that happens.


MS are not IBM. The 64 bit version of vista certianly appears to be solid enough. People are used to printers not working to well. The switch over to PCI Express should also reduce the amount of legacy hardware. For the time being the problem appears to be that XP is good enough.

I also doubt you will be able to run alan wake on a linux system (although there is the issue that running it on any currently existing system would be a bit of a problem).


Namely that Linux MUST HAVE THE CODECS BEING USED, MP3, DVD etc...


It also admits that in some cases at least that is imposible (there is a reason why flash video is the problem. H.263 is protected by quite a few patents).

[aerosolben]

Yes, but you are forgetting: this is a technical problem, as such it is in the easiest of categories to fix and also in the category of problems OSS people are best at fixing.

Technical problems are not problems, so to speak. Once identified they'll be fixed in short order.
It's been 7 years. How's that coming along?

That's less a criticism of Open Office, and more a criticism of writing off technical problems.

[geni]

Games are not as important as you think - consoles are far and away more popular.

One of which MS owns.

Other than games what needs more RAM than a 32 bit system can support? While new PCs will go to 64 bit the second hard market will keep 32 bit systems going for a long time because for most aplications they are good enough.

[cyborg]

By throwing money at it.

I don't think you get it.

Throwing money at it won't solve the fundamental problem. MS is out numbered by an order of magnitude of developers. They have internal organisational issues. There is a serious problem of internal inertia to overcome.

Money doesn't solve a goddamn thing automatically and it is entirely naive to think it can. MS's money is not useful for development - you cannot purchase development - what it is useful for is playing the business game where money does help. If they cannot play that game they are screwed.

What you propose is not a salvation for MS, it would be them admitting defeat and stepping down.

Nah pretty trivial really. You market it as a complete redesign of the code base in order to give more choice to the consumer.

Choice?! Man, choice is the one thing MS cannot afford to make anyone think they have. This is the whole point. They might just choose in a way MS wouldn't like!

Admitting that MS has to do things the *nix way is simply defeat - they've spent a lot of market effort trying to pull down the *nix arena after all.

MS are not IBM.

Au contraire - they are currently sitting in the same position IBM did in the late 70's before the microcomputer knocked them for six.

For the time being the problem appears to be that XP is good enough.

*Sigh* I refer you to one of the previous comments about why 64 bit is an important threshold.

I also doubt you will be able to run alan wake on a linux system (although there is the issue that running it on any currently existing system would be a bit of a problem).

Alan wake?

It also admits that in some cases at least that is imposible (there is a reason why flash video is the problem. H.263 is protected by quite a few patents).

*Sigh* So are the other techs. The problem is getting the licensing - the object of the patent here after all is not to prevent people seeing flash video. The only possible solution, as proposed, is to have some sort of per-seat license just the same as what happens in Window's land (without most people realising that they have in fact paid a small bit for the privilege of MPEG2).

[cyborg]

One of which MS owns.

So? It's a loss leader BTW.

Other than games what needs more RAM than a 32 bit system can support?

:rolleyes:

Applications will expand to fill capacity. Always have. There is great irony here in you not seeing this lesson of history when that is what the article is all about.

While new PCs will go to 64 bit the second hard market will keep 32 bit systems going for a long time because for most aplications they are good enough.

*Sigh* Are you ever going to get the point that there is no assumption otherwise?

[geni]

I don't think you get it.

Throwing money at it won't solve the fundamental problem. MS is out numbered by an order of magnitude of developers.


Meaningless. The correct way of counting is devloper hours.


They have internal organisational issues. There is a serious problem of internal inertia to overcome.

So do most linux projects (although I think Iceweasel is kinda cute)


Money doesn't solve a goddamn thing automatically and it is entirely naive to think it can. MS's money is not useful for development - you cannot purchase development


You can. MS have. We live in a capitalist system. Pretty much anything can be purchased.


- what it is useful for is playing the business game where money does help. If they cannot play that game they are screwed.


I think MS have shown they know how to play the business game.


What you propose is not a salvation for MS, it would be them admitting defeat and stepping down.

Not at all. Claiming to have totaly reworked the code is is good marketing.


Choice?! Man, choice is the one thing MS cannot afford to make anyone think they have. This is the whole point. They might just choose in a way MS wouldn't like!

And yet strangely you can run almost anything on windows. MS are quite happy to let you have choice as long as that choice is with an MS enviroment.


Admitting that MS has to do things the *nix way is simply defeat - they've spent a lot of market effort trying to pull down the *nix arena after all.


No the effort has gone into trying to remove GPL systems and other competitors. That they tend to be unix based isn't significant.


Au contraire - they are currently sitting in the same position IBM did in the late 70's before the microcomputer knocked them for six.


IBM had no Vienna equiv.


*Sigh* I refer you to one of the previous comments about why 64 bit is an important threshold.



Seen them. I think Windows Mobile shows that MS can deal with them.

However for your normal user office works fine under 32 bit systems as does pretty much everything else. There is no technical need to switch untill 2038.


Alan wake?


Game by the people behind Max Payne. I suspect it is meant to get gamers to switich to DirectX 9 based systems in the way halo did with the 360.


*Sigh* So are the other techs. The problem is getting the licensing - the object of the patent here after all is not to prevent people seeing flash video.

The only possible solution, as proposed, is to have some sort of per-seat license just the same as what happens in Window's land (without most people realising that they have in fact paid a small bit for the privilege of MPEG2).

Pay for linux?

In any case microsoft could just buy the patents.

[geni]

So? It's a loss leader BTW.

Gives them annother area to work from.



:rolleyes:

Applications will expand to fill capacity. Always have. There is great irony here in you not seeing this lesson of history when that is what the article is all about.


Extrapolateing when conditions don't stay constant is not legit. Other than games and some graphics stuff nothing comes close to pushing my current system to it's limits and existing aplications for the most part do a good enough job.



*Sigh* Are you ever going to get the point that there is no assumption otherwise?

since they assume the switch over time remains constant I would argue that there is.

[strathmeyer]

Is Vista 'desktop ready', yet?

[Dave1001]

It's the estimated amount of time before hardware will have sufficient RAM beyond that addressable with a 64-bit architecture.

A link to how the estimate is calculated and criticisms of it?

[JamesM]

A link to how the estimate is calculated and criticisms of it?
It's from the OP.


How long will the new 64-bit platform last?

It took 50 years to exhaust the first 32 bits, from the introduction of the Univac through 2005, which roughly matches Moore's Law's estimate of 48 years. It took 18 years (1987 to 2005) to go from 16 bits to 32 bits. Using the next 32 bits (to exhaust 64 bits) can thus be expected to take anywhere from 36 to 50 years.

[cyborg]

Meaningless. The correct way of counting is devloper hours.

Absolutely and totally WRONG.

The man month is mythical. This has been shown time and again.

You cannot say project X will take Y hours and then assign P people working T hours to complete it. It doesn't work that way. Development is not like construction work.

So do most linux projects (although I think Iceweasel is kinda cute)

I really don't think you understand the basics of organisational theory.

You can. MS have. We live in a capitalist system. Pretty much anything can be purchased.

*Sigh* This is plain wrong. I don't care what you think is possible - I am telling you IT CANNOT BE DONE. You do not understand the nature of the development process - this is why you continue to make basic assumptions that are wrong.

I think MS have shown they know how to play the business game.

Did I imply otherwise? They know every dirty trick in the book. They have made a lot of enemies by doing so as well.

Not at all. Claiming to have totaly reworked the code is is good marketing.

To a *nix based architecture? No. That is pure fantasy. You don't seem to know the history at work here.

And yet strangely you can run almost anything on windows. MS are quite happy to let you have choice as long as that choice is with an MS enviroment.

!!! Are you really that dense? You can run open standards on windows BECAUSE THEY ARE OPEN STANDARDS!!!. MS certainly does not care if you do so if you are running in an MS world. (And have you heard of embrace, extend and extinguish? It's the reason why the web is in the mire of **** standards compliance it is now). (Oh and not forgetting you can run most anything on Linux, even if it is illegal to do so).

The choice they DO NOT want you to make is to choose not to be in an MS world. They therefore use everything at their disposal to ensure that choice is difficult.

No the effort has gone into trying to remove GPL systems and other competitors. That they tend to be unix based isn't significant.

YES IT IS - that's the competition! They've spent a lot of effort explaining WHY they are better than *nix.

Producing a *nix compliant system would effectively lower the barrier to move to other *nix systems. I really don't think you get why MS wouldn't do this.

Seen them. I think Windows Mobile shows that MS can deal with them.

I fail to see how that works.

However for your normal user office works fine under 32 bit systems as does pretty much everything else. There is no technical need to switch untill 2038.

*Sigh* That does NOT help MS. THEY NEED the revenue streams. Everyone using the same systems they are today would still leave MS in the crapper.

Game by the people behind Max Payne. I suspect it is meant to get gamers to switich to DirectX 9 based systems in the way halo did with the 360.

Again, PC gaming is just not as important as you think. It's simply not as big as on other platforms.

Pay for linux?

Yes. If you think that's a contradiction then you do not understand what you are arguing against.

In any case microsoft could just buy the patents.

:rolleyes: I really think you should take a look at how the real world works rather than how you think it should work.

[Wudang]

Absolutely and totally WRONG.

The man month is mythical. This has been shown time and again.

You cannot say project X will take Y hours and then assign P people working T hours to complete it. It doesn't work that way. Development is not like construction work.

Indeed. And that's why linux having 10 times the developers is irrelevant. One of the finest products I ever worked on was by the smallest very very talented and experienced team.

*Sigh* This is plain wrong. I don't care what you think is possible - I am telling you IT CANNOT BE DONE. You do not understand the nature of the development process - this is why you continue to make basic assumptions that are wrong.

Argument by assertion. When NT was in trouble getting going they hired away the chief VMS architect to sort the kernel.

To a *nix based architecture? No. That is pure fantasy. You don't seem to know the history at work here.

How did they sell NT?

!!! Are you really that dense? You can run open standards on windows BECAUSE THEY ARE OPEN STANDARDS!!!.

Which have to be implemented on each platform. I was very impressed with the team that got posix on os/390 OE.

The choice they DO NOT want you to make is to choose not to be in an MS world. They therefore use everything at their disposal to ensure that choice is difficult.

Yes. Though look at the recent MS/Novell deal. That surprised me.

YES IT IS - that's the competition! They've spent a lot of effort explaining WHY they are better than *nix.

Producing a *nix compliant system would effectively lower the barrier to move to other *nix systems. I really don't think you get why MS wouldn't do this.

And I think you vastly underestimate the power of marketing and following the market. Look at all the excitement that was generated by a few visual frills you can get today with stuff like windowblinds.

*Sigh* That does NOT help MS. THEY NEED the revenue streams. Everyone using the same systems they are today would still leave MS in the crapper.

Yes. MS shareholders and market anaylsts have come to expect certain growth figures and could cause a massive (if possibly ill-advised) drop in MS stock.

Again, PC gaming is just not as important as you think. It's simply not as big as on other platforms.

Do you have figures? I'm curious as I thought it was quite big but I could be wrong

[richardm]

The Office suite doesn't have to be MS Office.


It doesn't - but people want it to be at the moment. This is what was argued earlier. People are familiar and comfortable with MS products, and unless they produce a series of total turkeys it's going to be a job to persuade people to move away. They don't care what's under the skin so long as it has the reassuring Microsoft Tattoo on it. Corporations especially have a lot of inertia, and I think that while the article makes some fair points it is simply overlooking the marketing power that Microsoft can bring to bear. Lack of technical superiority is simply not an impediment.

Do you have figures? I'm curious as I thought it was quite big but I could be wrong

Me too - however this data (http://www.theesa.com/facts/sales_genre_data.php) indicates that in recent years video games shifted 5 times as many units as PC games.

[geni]

Absolutely and totally WRONG.

The man month is mythical. This has been shown time and again.


So why have debian taken to employing people?


You cannot say project X will take Y hours and then assign P people working T hours to complete it. It doesn't work that way. Development is not like construction work.

However a lot of the problems (such as updateing drivers and the like).


*Sigh* This is plain wrong. I don't care what you think is possible - I am telling you IT CANNOT BE DONE.

Which laws of physics would it violate.


You do not understand the nature of the development process - this is why you continue to make basic assumptions that are wrong.


That would be an appeal to authority logical fallacy.


To a *nix based architecture? No. That is pure fantasy. You don't seem to know the history at work here.

What I know doesn't matter. It is what their customers know. Their customers don't really know what unix is.


!!! Are you really that dense? You can run open standards on windows BECAUSE THEY ARE OPEN STANDARDS!!!.


No you can run them because MS lets you. It would be fairly trivial for MS to prevent them from running.


MS certainly does not care if you do so if you are running in an MS world. (And have you heard of embrace, extend and extinguish? It's the reason why the web is in the mire of **** standards compliance it is now).

Nah that is everyone's fault (well with the posible exception of opera).


(Oh and not forgetting you can run most anything on Linux, even if it is illegal to do so).


Rome: Total War.


The choice they DO NOT want you to make is to choose not to be in an MS world. They therefore use everything at their disposal to ensure that choice is difficult.

Of course but at the same time once you are in that MS world they have no problem with choice.


YES IT IS - that's the competition! They've spent a lot of effort explaining WHY they are better than *nix.

Producing a *nix compliant system would effectively lower the barrier to move to other *nix systems. I really don't think you get why MS wouldn't do this.


Again MS can fall back on patents.


I fail to see how that works.


It was a new platform that required a new OS. MS did rather better than linux.


*Sigh* That does NOT help MS. THEY NEED the revenue streams. Everyone using the same systems they are today would still leave MS in the crapper.


Lot of systems out there don't use XP yet. No the effect is that there is much greater momentum to use something with a windows sticker on it.


Again, PC gaming is just not as important as you think. It's simply not as big as on other platforms.

It is pretty much unique to microsoft. Macs linux and BSD are not gameing platforms in the way windows is. Secondly gamers are the group that will upgrade early so to an extent will lead the way. Thirdly they are a set the hardware manufactors would really rather not lose.


Yes. If you think that's a contradiction then you do not understand what you are arguing against.


Oh I understand it happens.


:rolleyes: I really think you should take a look at how the real world works rather than how you think it should work.

People with patents tend to want money. Microsoft has money. In the real world the result is fairly straight forward. Would hardly be the first time microsoft has purchased outside stuff that they wanted.

[Jekyll]

Me too - however this data (http://www.theesa.com/facts/sales_genre_data.php) indicates that in recent years video games shifted 5 times as many units as PC games.

Unfortunately ,that doesn't tell us anything about how many PC users expect to be able to play games on their computer.

[Almo]

-1 Redundant.

:rolleyes:

But true.

:rolleyes:

[Almo]

Is Vista 'desktop ready', yet?

Not really. MS OS releases are usually pretty unsuable. Wait till Service Pack 1 or 2. It will be more stable by then.

[QueigBladecaster]

Again, PC gaming is just not as important as you think. It's simply not as big as on other platforms.


While I agree that PC gaming is being taken over by consoles Its one of the big reasons 64 bit technology has found its way into the home market so fast.

If MS wishes to throw its hat into the Linux ring it could do it very easily. It would be fairly simple matter to buy a popular linux distro, throw some money and marketing at it and get it to a point where you can get business to pay for support licences. Novell are doing quite nicely at this with SUSE.

The fact is they dont need to, windows works (im not saying it works well but it does work). Unless windows stops working the situation is unlikely to change very much. Windows XP64 is already compatible with active directory, exchange, MS office, Citrix, etc, etc, etc. There is no reason to suspect any future version of Windows will not be. This means that business has a clear and easy upgrade path to it. For the move to Linux desktops there is a huge amount application testing and research that a business needs to do before it even considers the high retraining and support costs involved. Linux is not 'free' by any stretch of the imagination in the business environment.

What linux currently lacks is a clear direction or marketing plan, and its something it will never have because of its nature. The open source community is not going to pay to advertise, it has to rely on word of mouth. In the same way the open nature of linux means that things are developed for a percieved need in the community not for a specific business case. The only distro that comes close to having that is Red Hat and their current marketing plan seems to be to push further into the server and virtualisation market.

[jimlintott]

MS wouldn't have to buy a Linux distro, they could just make their own. So can any of us. It would be viewed by many as a defeat. I would think of it that way.

I don't think it would suit their current business model. MS doesn't really support very well. They prefer to just sell software to the Dells, Gateways, etc. They let those companies provide support. I know MS provides support but that isn't really their game.

While MS will cope with 64bit just fine the article does point out some of the problems they face. Legacy hardware drivers that they don't have source code for and no way to get it. They have to do what the OSS guys did and reverse engineer (if the old hardware companies are gone) to make proper 64 but drivers. An emulation layer is possible but clumsy. My guess is that if you call them to say that XYZ hardware doesn't work they would tell you to buy new hardware.

Another problem I see cropping up for MS is new platforms. If someone revolutionises the CPU with a new platform how quickly could MS support it. Their single platform support has probably actually retarded growth in CPU performance. Look how easy it was for Apple to switch to Intel. The BSD kernel made that easy. Linux is also cross platform and can easily shift to another CPU. Linux can boot a Playstation 3 which might make it the first desktop suited OS to run on a cell processor.

I think in the future MS will find it necessary to shift platforms and be able to hit a moving target. They've kept the target frozen for too long.

My own opinion is that there will be some shaky times ahead for MS. I doubt they will die, they are too big and strong but they had better diversify for when the desktop marketshare becomes more fragmented. Up until recently they had virtually zero competition.

That's my two cents, anyway.

[QueigBladecaster]

MS doesn't really support very well

This is very true, but there is a huge volume of cheap support professionals and semi professionals which means they dont have to.

Its currently not the same situation for unix and linux. There is a premium for engineers/consultants with sun solaris/unix/linux skills purely because there are less people with those skills and experience.

The situation is definately improving, but far too slowly.

[bruto]

Absolutely and totally WRONG.

The man month is mythical. This has been shown time and again.

You cannot say project X will take Y hours and then assign P people working T hours to complete it. It doesn't work that way. Development is not like construction work.


A little sideways drift here, but I'm always reminded of Ambrose Bierce's example (illustrating logic): a man can dig a post hole in sixty seconds, therefore, sixty men can dig a post hole in one second.

[aerosolben]

Its currently not the same situation for unix and linux. There is a premium for engineers/consultants with sun solaris/unix/linux skills purely because there are less people with those skills and experience.

The situation is definately improving, but far too slowly.
Evidence?

As I posted earlier, Windows is increasing their market share in servers, and essentially holding pat in workstations. It seems counter-intuitive to me that the *nix skills base is growing (relative to the Windows skill base).

[Stainless_Steel_Rat]

Games are not as important as you think - consoles are far and away more popular.

Do you have evidence to back this up? Sure, consoles are taking over B&M locations, becuase it's nearly impossible to resell PC games. And consoles draw in a lot of midrange gamers. But PC's have both the casual and hardcore markets well in hand. Many people believe the Myth that gaming on the PC is dying, but it's far from it, just changing. Casual games like all the Popcap series and with new distro models like Steam and EA link, PC games are still raking in the dough.

Without PC games to push the equipment to it's limits, what's the use of a more powerful computer? The only people who will care are animators and CAD users.
Sure, the industry will start selling 64bit systems with 4GB of ram, but unless there an application that people WANT/NEED for it, why upgrade?

SSR

[Soapy Sam]

Woiuldn't it be nice if a CPU came with the OS already integrated into the architecture?
So when you switched it on, it was already booted and could run any damn software you happened to have?

[QueigBladecaster]

Evidence?

Personal experience in the industry. More and more of our clients are implementing red hat and solaris for their server platforms. Our sun consultants have never been busier, and were seeing far more project work involving linux solutions then we have in the last couple of years.

I dont doubt for a second that the windows server market share is increasing, and I dont expect Linux to take over anytime soon but the increased interest is encouraging.

The other area that is growing quickly is virtualisation and we are having a big push atm to train up engineers and consultants on VMware and XEN.

[aerosolben]

Personal experience in the industry. More and more of our clients are implementing red hat and solaris for their server platforms. Our sun consultants have never been busier, and were seeing far more project work involving linux solutions then we have in the last couple of years.

I dont doubt for a second that the windows server market share is increasing, and I dont expect Linux to take over anytime soon but the increased interest is encouraging.
Are you seeing them leave Unix, or is it just new systems? I believe the Linux server market had a big growth spurt a couple years ago; It's cooled a little it since then, but I think it's still growing.

The other area that is growing quickly is virtualisation and we are having a big push atm to train up engineers and consultants on VMware and XEN.
Now THAT I definitely believe. :)

[ob986s]

Personal experience in the industry. More and more of our clients are implementing red hat and solaris for their server platforms. Our sun consultants have never been busier, and were seeing far more project work involving linux solutions then we have in the last couple of years.

I dont doubt for a second that the windows server market share is increasing, and I dont expect Linux to take over anytime soon but the increased interest is encouraging.

The other area that is growing quickly is virtualisation and we are having a big push atm to train up engineers and consultants on VMware and XEN.

I am seeing the same thing, of course as full disclosure I work for Vmware

I find this entire thread funny because I believe at VMware we are making the OS obsolete anyway. If you look at what MS is doing as a business right now they seem to be more worried about what we are doing then what Linux may do on the desktop. I would argue that in the next few years many will be running VMware with multiple operating systems running in Virtual Machines on the same box.

IMO MS has more to worry about with us separating the Hardware from the OS than they do of losing the desktop


Jon

[jimlintott]

I am seeing the same thing, of course as full disclosure I work for Vmware

I find this entire thread funny because I believe at VMware we are making the OS obsolete anyway. If you look at what MS is doing as a business right now they seem to be more worried about what we are doing then what Linux may do on the desktop. I would argue that in the next few years many will be running VMware with multiple operating systems running in Virtual Machines on the same box.

IMO MS has more to worry about with us separating the Hardware from the OS than they do of losing the desktop


Jon

Very interesting position. I have a few questions though (I'm not being argumentative. I want to learn more.)

How is virtualization making the OS obselete? I thought it was so you could run several different OS on the same machine? I don't see how this eliminates the OS unless Vmware becomes an OS in itself.

The article linked to in the OP is mostly about the desktop and home use. Isn't virtualization really a server issue. Some users may need it (software and web developers come to mind) but most desktop users will always be single OS. I suppose in the future that OS may actually be virtualized on a bigger server but how long have we been waiting for the dumb terminal now? (I do think dumb terminals are the future for businesses.)

Also much of the article talks about how they expect MS to have trouble switching to 64bit properly. I think MS would have a hard time switching to a new platform. But if this is true doesn't virtualization actually solve that problem for them?

If Vmware was the actual solution to Microsoft's problem then it won't be long until it is MS Vmware.

If you have some shares you might be able to retire that day. ;)

[aerosolben]

If Vmware was the actual solution to Microsoft's problem then it won't be long until it is MS Vmware.
Microsoft already has in-house virtualization: Virtual Server and Virtual PC.

I don't think MS has a pure hypervisor available. I would not be surprised to find out that one was being developed, though.

[geni]

Microsoft already has in-house virtualization: Virtual Server and Virtual PC.

I don't think MS has a pure hypervisor available. I would not be surprised to find out that one was being developed, though.

Microsoft are not talking about psot vista much Vienna may be working towards hypervisors.

http://www.infoworld.com/article/07/02/09/HNvistafollowup_1.html

[ob986s]

So to answer the last three responses

Microsoft is working on a hypervisor, code named Veridian which is expected to ship 6 months after "longhorn" which is the next Server OS. Longhorn is targeted for late 2007 so I expect it will be 2008 when we will see the MS Hypervisor. though I do not expect it to be completely bare metal as our ESX Server platform is as MS's needs the OS to remain important.

On the Dumb terminal side at least 20% of my current opportunities are on the desktop (I am east coast US Sales) which means hosting XP (or SLED, RH) Desktops on servers, fed out to thin or fat clients. And unlike the thin client initiatives of the late 90's this actually works

On making the OS obsolete that was a very poor choice of words on my part while on a conference call :) I would like to retract that and say we are breaking the bond of the OS to the Hardware (and scaring the pants of MS in the process as so much of their $ is tied to hardware) and decreasing the importance of the OS. What does it matter what OS you need when you can run a dozen OS's on the same box?

And as far as Becoming MS Vmware, we are already 100% owned by EMC, who has announce that they will be IPOing us in the summer so i find it unlikely that even MS could buy us now.


Jon

[rockoon]

64-bit doesnt mean anything.

Very few people are even considering throwing around more than 2 gigs of data in realtime, let alone actualy doing it.

More data implies more content. More content implies bigger development costs. Games already have massive development costs due to content and developers are looking to alternatives such as procedural runtime generated content.

The big wakeup call is in how well an OS allows the software which runs on it to scale up with more processors/cores. Linux doesnt seem to offer any real advantages especialy since its written in mostly C++ (just like windows) which still has absolutely no threading standard.

[MortFurd]

64-bit doesnt mean anything.

Very few people are even considering throwing around more than 2 gigs of data in realtime, let alone actualy doing it.

More data implies more content. More content implies bigger development costs. Games already have massive development costs due to content and developers are looking to alternatives such as procedural runtime generated content.

The big wakeup call is in how well an OS allows the software which runs on it to scale up with more processors/cores. Linux doesnt seem to offer any real advantages especialy since its written in mostly C++ (just like windows) which still has absolutely no threading standard.
So when was the last time you did any video editing? That's a (realtively) common task that uses up piles of RAM and disk space.

The kernel of Linux is written in C. Applications may be written in C, C++, Fortran, Assembler, Perl, Python, or any one a of a bazzilion other languages.

[rockoon]

So when was the last time you did any video editing? That's a (realtively) common task that uses up piles of RAM and disk space.


At 248 megabytes per second for the comming HDTV resolutions... and 1 gig of memory costing ~$100 .. how is 2008 even a reasonable year for end user video editing considerations?

Pulling a number out my ass, 9999 out of 10000 computer users dont edit video.

[geni]

So when was the last time you did any video editing? That's a (realtively) common task that uses up piles of RAM and disk space.


Disk space hasn't been an isue yet and 1.5 gig of ram appears to be enough for the moment.

[RAMS]

I am an IT systems engineer working in the UK, and I can certainly tell you that many companys are starting to embrace 64-bit hardware and more importantly linux and solaris in preference to windows for server side applications.

But that is a whole other world then getting linux established as a home PC or even a business workstation platform.

Its just too complicated and user unfriendly and not to mention confusing with the huge number of distros available atm for it to be accepted by the home market.

Its going to take a major shift for people to move away from windows. people know it, they are familiar with it and there is tons of software for it. Most casual PC users dont want to look at what is the best OS they want whats easiest for them to use and what they feel comfortable with. Microsoft has the drop on this. there system is attractive, very easy to learn universially supported and most PC users know that is what most of their friends are using.

Most people who use Apple will tell you til they are blue in the face that MacOS is a far superior OS and in many ways they are right. But the pure fact that is less widely used by the average person means that its less likely to be accepted as the defacto standard home OS. And this is a system that is just as easy to use and some say more attractive then the current defacto standard.

Looking at that you can see how huge the mountain is that linux has to climb. The number of people the average person will know who uses Linux is even smaller. The fact that it is so confusing to even choose which distro to go for will put people off. Add to that the less then intuitive way of configuring the systems and adding your own applications, and the fact that you probably will not be able to use that new and cool application that most of your friends are raving about, means that I suspect Linux will stay firmly in its niche market for some time to come.



Good report. Fair too. For fun I did linux on my twin screw mac and it was fun but was exactly as you described. I also played with a few others and under emulation linux was better.

But, I do FX and visuals (samples here: http://www.behold-the-rage.com/the_images.html) so stick with OSX. Have been Mac for 27 years (then it was a Apple II GS, switching from a Wang---jeezzz, that is ancient...), and happy with it for its speed and stability.

Again, good post.

RAMS

[PixyMisa]

At 248 megabytes per second for the comming HDTV resolutions... and 1 gig of memory costing ~$100 .. how is 2008 even a reasonable year for end user video editing considerations?

End users don't edit uncompressed HD video. End users generally don't even have access to uncompressed HD video.

Pulling a number out my ass, 9999 out of 10000 computer users dont edit video.

You need a better ass.

[rockoon]

End users don't edit uncompressed HD video. End users generally don't even have access to uncompressed HD video.


If its compressed non-HD video that we are talking about, 1 gigabyte will hold an entire feature length movie. 64-bit is moot in that case too.

This was so obvious that it didnt need mention because I assumed the person I was communicating with wasnt an idiot. Too bad you didnt make similar assumptions. Thanks pal.

[PixyMisa]

This was so obvious that it didnt need mention because I assumed the person I was communicating with wasnt an idiot. Too bad you didnt make similar assumptions. Thanks pal.

Ah, so you were making a point that was entirely irrelevant to the topic? My apologies then.

[PixyMisa]

Anyway, to directly address the subject:

No, Linux isn't going to make any particular inroads on the desktop because of the transition to 64 bits.

64-bit Windows is available now. So some old hardware won't work because of driver issues. That just means that some people will stick with 32-bit Windows until their old hardware dies.

As for being "well-researched", well. 32-bit desktop systems arrived in 1983, or 1984 if you insist on mass-market penetration. Windows didn't go fully 32-bit until 1995. The hardware/software overlap window is a lot longer and a lot more blurred than the article tries to make out. There's no hard deadline, not when viewed across the broader market.

The people who need more than 4GB now, who aren't already on Linux, will go with Vista 64 (or Windows 2003) and just live with the problems - at least 98% of the time.

Another key point is that 32-bit operating systems running on 32-bit hardware can access more than 4GB of memory. They just can't assign more than 4GB to a single user-space application at once. This is already built into various versions of Windows server OS's; I see no reason it couldn't be rolled into an upgrade of Vista if that became necessary.

[rockoon]

Ah, so you were making a point that was entirely irrelevant to the topic? My apologies then.

The person I was communicating with brought up video editing as it applied to the need for 64-bit addressing (which is the thread topic.)

I addressed that very issue.

Memory is currently too expensive for the common enduser to have more than 4 gigabytes. The topic of the entire thread of course is that there will be more than a superficial demand for more than 4 gigabytes by the end of 2008. This may be correct but its not correct in regards to video editing. Far less than 1 in 10,000 users engage is the practice of editing video data that might actualy require 4 or more gigabytes of data (such as feature length HD video.)

..as I pointed out, which you apparently didnt know, a 2 hour DVD quality movie under the latest video compression schemes (divx, xvid, mp4) fits in under 1 gigabyte. The entire movie, sound and all, fits in memory on even old machines like a premium "internet ready" dell computer from 5 years ago.

YOU brought up the irrelevant crap, pal.

Just so that you are aware, 64-bit addressing is exploitable by programs with datasets that are more than 3 gigabytes (under windows) or 4 gigabytes (under linux.) It is not expliotable by programs that use smaller datasets than that. Its is completely irrelevant to those programs. Meaningless sophistry that you tried to engage in.

Additionaly, just so that you are aware, video editors are crippled by current processing power. Memory doesnt even factor into it and wont until the processing power deficiency is solved.

[scribble]

I'd just like to point out that while it's a rare program that can currently benefit from addressing > 4gb RAM, that's not the only benefit of today's 64-bit processors.

There are several, but the what comes to mind as being directly a result of the 64-bit architecture is the additional registers and their larger size. That can and does benefit many, many applications.

[rockoon]

I'd just like to point out that while it's a rare program that can currently benefit from addressing > 4gb RAM, that's not the only benefit of today's 64-bit processors.

There are several, but the what comes to mind as being directly a result of the 64-bit architecture is the additional registers and their larger size. That can and does benefit many, many applications.

This is indeed a benefit of x86-64 and they (the extended registers) are available even while in 32-bit mode just as 32-bit registers were available in 16-bit mode and 16-bit registers were available even when the x86 series was still 8-bit.

Many people dont even know that they have a 64-bit CPU simply because they arent marketed as such. Thats how little 64-bit means to the consumer.

[PixyMisa]

This is indeed a benefit of x86-64 and they (the extended registers) are available even while in 32-bit mode just as 32-bit registers were available in 16-bit mode and 16-bit registers were available even when the x86 series was still 8-bit.

No. They're not, and they weren't. It's a minor point, but your statement is untrue in every particular.

See this diagram (http://en.wikipedia.org/wiki/AMD64#Operating_modes).

[PixyMisa]

The person I was communicating with brought up video editing as it applied to the need for 64-bit addressing (which is the thread topic.)

I addressed that very issue.

Perhaps you did, but not very clearly.

Far less than 1 in 10,000 users engage is the practice of editing video data that might actualy require 4 or more gigabytes of data (such as feature length HD video.)

Yes, agreed.

..as I pointed out, which you apparently didnt know, a 2 hour DVD quality movie under the latest video compression schemes (divx, xvid, mp4) fits in under 1 gigabyte.

Yes.

Just so that you are aware, 64-bit addressing is exploitable by programs with datasets that are more than 3 gigabytes (under windows) or 4 gigabytes (under linux.) It is not expliotable by programs that use smaller datasets than that. Its is completely irrelevant to those programs. Meaningless sophistry that you tried to engage in.

Chill, dude or dudette.

You mentioned that uncompressed video streams are far too large to fit in memory for any reasonable consumer-level system for the next few years; perfectly true. I noted - which you didn't make clear in your original post - that consumers don't edit uncompressed video streams. Your statement that
9999 out of 10000 computer users dont edit video.
lacking as it does the word "uncompressed", is not exactly a paragon of precision.

Additionaly, just so that you are aware, video editors are crippled by current processing power. Memory doesnt even factor into it and wont until the processing power deficiency is solved.

Depends on the editing taks, of course.

Where video editing does need lots of memory is disk buffering. If you want to jump around in the video making cuts and splices and so on - which aren't CPU intensive the way effects and compression are - you want as much of the video in memory as possible. And you don't want the video compressed down to sub-1GB per hour when you're editing, because repeated recompression at that level can cause some serious artifacts.

Of course, 32-bit processors can address more than 4GB of memory; they just (as you note) can't assign more than 4GB to a single process. So a 32-bit processor could easily have 8GB of memory attached, and could greatly speed up video editing responsiveness.

Look, obviously we agree on this. So can we leave the aggression out?

[scribble]

This is indeed a benefit of x86-64 and they (the extended registers) are available even while in 32-bit mode just as 32-bit registers were available in 16-bit mode and 16-bit registers were available even when the x86 series was still 8-bit.

I'm trying to be generous here, so rather than saying soemthing that may be taken as ad hominem, let me try to explain where you've become mistaken.

Our current 64-bit processors can run 32-bit code; that is, code that was written and compiled for a 32-bit machine. If this is what you are referring to, then you are mistaken. The actual machine code that makes up the program only knows about the 32-bt registers that would have existed on the 32-bit processor it was compiled for. (More often than not, this is the 386 -- the literal 80386 spec, I'm not making that up, yes, most programs don't take advantage of anything newer than that, really.) If you take this program and run it on any other machine, it will continue to use only those registers that would have existed on the processor it was coded for.

You might be able to take a program written in a high-level language like C or even higher up that has previously been compiled for 32-bits, and simply compile it for 64-bits, and expect the newly compiled code to use the 64-bit registers. But at that point, you have a 64-bit program, and bringing 32-bits into the conversation just confuses the matter. It's not a 32-bit program in any sense, at all.

Are the registers accessable when you have the processor working in 32-bit mode? I don't know. It doesn't matter because the code that is executing will not be able to use them, even if they're technically available.

[scribble]

Your statement that

lacking as it does the word "uncompressed", is not exactly a paragon of precision.

I'm not a video editor, so explain to me if i'm mistaken here. I must imagine that every video that is edited is uncompressed at the time of edit, even if that fact is abstracted away from the user. The mathematics of literally editing data that is compressed, without decoding, editing, then re-encoding would boggle the mind. I would love to see evidence if there is, in fact, any program or even unimplemented algorithm for literally editing compressed video streams (i.e. while compressed).

My belief -- having not studied it in any depth -- is that all video editors must work on uncompressed data, and therefore the distinction between editing compressed videos and uncompressed videos is moot. They're all uncompressed at the point of edit.

Am I wrong?

If I'm not, I woudl think the distinction would properly be made between an editor that requires the entire stream to be decompressed up-front and an editor which is happy to take a compressed stream and re-encode it on the fly in order to get the job done without all the disk space needs. This would, of course, also explain disk thrashing when such an editor is used; it would be a foregone conculsion by the programmer that the entire uncompressed video would never be in memory at any given time.

Let me use an analogy. When I edit a photo with GIMP, let's say, it requires X*Y*bits-per-pixel bits in memory -- let's say it's 50MB. If I save the photo as a JPEG, and then re-open it to edit again it will still be 50MB in memory because in order for GIMP to edit it, it must be the uncompressed pixel data in memory. Now, the quality of that image may have degraded in the meantime, but that doesn't make any difference, since it's still X pixels by Y pixels, just like before, it still require X*Y*bits-per-pixel bits in memory to edit.