Or on the other hand, if you had no knowledge of computers and walked into a store only to see a machine for $200 and right next to it the same machine for $500-$700 which would you buy? Your figures are greatly exaggerated. Even in rip off britan XP home/vista home basic (which are the editions are a cheap shit box would come with) whitebox OEM are arround £50 ($100) including VAT (our equivilent of sales tax), it is widely believed that the big brand OEMs pay even less.
The problem is that the same time Joe Average is picking out that computer they are also looking at the software shelf loaded with Microsoft centric crap. The moment they pick up that shareware disk for $5.00 and ask, "will this work on that box I'm buying?" will be the kiss of death on that sale. Add in the fact that sales people at WalMart aren't the pick of the crop and mess up even Windows technical issues and it is a recipe for a PR disaster. Agreed, selling linux succesfully requires educating the buyer that it is not windows and what it's advantages (zero cost, less vulnerability to shitware, availibility of a lot of very good free software from the distros repositries) and disadvantages (inability to run the software they are used too and they see on the shelves in every computer related shop, lack of availibility of support from your more geeky but still MS using friends) are and letting them make an informed choice. Sadly this is hard when they don't even know what an OS is.
Tricking people into buying linux when it is not right for them will only breed resentment, especilly when they have to pay three times as much to buy windows after the fact as they would have to buy it with the PC.
I would only reccomend linux on the desktop to anybody if I knew appropriate software for the tasks in hand was availible and either: * I was going to be supporting it * I knew someone with appropriate linux knowlage was arround to support it. * The box was being used for a very limited set of tasks with little prospect that it was going to be used for more
Also even if I reccomended linux on the desktop if I thought there was a reasonable (more than 1 in 3) chance the box will ever be used for something requiring windows then I would have to reccomend getting the windows license anyway due to the aforementioned huge price differential between OEM and retail (afaict most if not all windows volume licences are upgrade/downgrade only).
The trouble is afaict only failures in places like nuke plants tend to get analised rigorously enough to diagnose this sort of thing.
There are so many possible causes of failure in electronics and there is so little reporting of how long equipment lasts and how it fails that drawing meaningfull conclusions on whether this is having a significant effect on the lifetime of consumer electronics is difficult to measure.
Most non-x86 architectures are moving back to in-order execution. Compilers are good enough that they put instructions far enough away to avoid dependencies (something much easier to do when you have lots of registers) and the die space savings from using an in-order core allows them to put more cores on each chip. OTOH most non x86 architectures are used in environments where it is feasible to compile for the specific chip.
to win in the PC market chips must be able to perform reasonablly well on code compiled by compilers targetting older chips as that represents the code that most people will be running.
returning pages of memory to the OS is possible at least on windows on linux.
The problem with making a memory manager (defined here as a peice of code that allocates pages from the OS and suballocates them to meet the demands of other parts of the application) return memory to the OS is that memory has to be allocated and returned to the OS in complete pages. In other words the already complex buisness of managing a heap that stores random small locations gets even more complex as the manager must try to empty pages so they can be returned to the OS.
Yeah, if he actually filed usable bugs, they'd get fixed, and then he'd have nothing to whine about any more. The problem is it is very hard to file usable bug reports for "works fine at first but tends to suck up ever more memory and grind to a halt over long sessions in my environment". You really don't know if it is firefox core or some extension or plugin that is to blame nor if it is firefox core what part of it.
If a developer can't reproduce a bug quickly and reliablly it is very hard for them to find where the bug is and to confirm they really have it fixed.
Also mozilla is known to ignore reproducable crash bugs because they don't affect enough people. The fact they are having to do that speaks volumes about thier code quality IMO.
CIH was destructive to motherboards and certain disk structures (which were generally repairable provided the hard drive was fat32 and the first partition was over a certain size) but that was about all the damage it did.
Far more insiduous would be malware built into the bios by the manufacturer which allowed secret commands to be used to read out data through the systems onboard network adaptors.
Why oh why does Microsoft still automatically run software off any disk that's inserted into your PC? Surely decades of floppy-carried virii should have convinced them of what a frigging stupid idea that is? When CD autorun was first introduced CDs were read only media produced only by big and presumablly trusted companies. So the risk was pretty low. Hard drives equally were rarely shared with anyone but very trusted people so they were pretty low risk too. I don't think floppy and similar drives ever had autorun.
Of course things have changed now putting MS between a rock (vulnerability to malware) and a hard place (pissing off users who have become used to autorun).
If you think this is bad you might like to know that on risc OS just viewing the contents a directory was enough to have code automatically run on your system.
And then under EU law they can't fill that post for a certain length of time (1-2 years?) But is there anything stopping them just reorganising a bit so that no one person is doing the job of the person let go or setting up a shell company to employ the person and have the work done on a contract basis by the shell company?
I've added lots of drives to Windows machines and it never occured to me to try to access them without formatting them. Do these come preformatted? In my experiance bare drives don't but drives ready mounted up in USB caddies do.
Sure you could reformat it to remove stuff but by the time you get to the format screen you are probablly already infected.
agreed. I know a woman in her early 20s who does not work because she has calculated she would have to make at least $12/hr full time to make up for food stamps and welfare and pay for child care while at work. This seems to be a symptom of the benefit system not taking account of expenses related to holding down a job (childcare, fuel etc) and/or not having a sane taper off system.
To them, it's either a full-blown welfare state or Dickensian workhouses. The idea that life isn't binary, and that there might be alternatives between extremes seems unfathomable. A lot of issues seem to gather extremists on both sides.
A good policymaker needs to identify problems in the current system (and they undoubtablly are there) and try and sort them out but equally they need to avoid following extremists who demand rash changes without really thinking them through.
What do you think would happen to theese slums if the governement suddenly stopped the welfare and the subsidised housing as some people seem to be suggesting? Do you think the people would mostly find gainfull employment or do you think there would be a huge rise in low level crime, begging and people living on the streets? I would bet on the latter.
by encouraging work and discouraging poor decisions. On the other hand once people end up living on the street begging to survive afaict it is very very difficult for them to get back into work because they won't have an address for mail, they wont be able to keep properly clean and so on.
I guess thats one way to try and persude people to fork out for the upgrade. Unfortunately since many of them probablly won't anyone who develops in java and has OS-X as a target is probablly going to be stuck with 1.5.x for some time.
Also it looks like that timeline is just speculation based on past behaviour not actual data from apple despite the title.
Letting ordinary users set the system time isn't particularlly smart imo. If time configuration is done by someone who dosn't understand timezones you will end up with fucked up timestamps.
As ATI has started doing with the R500 and R600 series. So far the technical docs they've released basically only cover the frame buffer, but they've stated that the 3D docs will be forthcoming. (2D acceleration is done using the 3D engine.) Lets hope they really follow through. If they do and a workable opensource 3D graphics driver comes out of it and gets included in the major linux distros I will be specifying ATI rather than Nvidia when I buy machines.
For now, it looks like the 3D graphics hardware with the best open source support is the Intel GMA-X3000 integrated graphics in the G965 and GM965 chipsets. The performance is lower than the bleeding-edge ATI and Nvidia parts, but it's adequate for most purposes. I know whichever intel chipset apple put in the macbook isn't even capable of running planetpenguin-racer acceptablly:(. Looking at the wikipedia article it seems to be one of the chipsets you mentioned. Even a £20 geforce FX5200 runs it fine so it isn't just bleeding edge cards the intel integrated is losing to. They are probablly good enough for desktop eye candy though.
For now i'm sticking with nvidia for any machines where I care about graphics performance because the impression I get from linux forums is that they are less hassle to set up on linux than ATI ones.
Using apt-get dist-upgrade blindly is indeed nuts but if you read the release notes and pay attention to what apt plans to do you are pretty safe and unlike with fedora upgrading with the package manager is the supported and reccomended way to upgrade.
That is macrovisions most famous defective restricted media system (mainly because it was one of the first defective restricted media systems created) but it is far from thier only one.
IIRC DSL is woody based. If that is still the case then unless they have been doing a lot of work themselves the software is probablly littered with security holes. Maybe acceptable for a livecd but not for a main desktop OS.
Most video display technologies need an amount of stabilization to display images accurately. That requires a constant current load, unless you want to go back to the 50's, 60's, and 70's where the TVs have to warm up before you use them. I call BS, I don't think I have ever seen a TV or monitor new enough to have standby that couldn't come up to an acceptable picture in a couple of seconds from a cold (no power) start.
Slashdot Mirror
Or on the other hand, if you had no knowledge of computers and walked into a store only to see a machine for $200 and right next to it the same machine for $500-$700 which would you buy?
Your figures are greatly exaggerated. Even in rip off britan XP home/vista home basic (which are the editions are a cheap shit box would come with) whitebox OEM are arround £50 ($100) including VAT (our equivilent of sales tax), it is widely believed that the big brand OEMs pay even less.
The problem is that the same time Joe Average is picking out that computer they are also looking at the software shelf loaded with Microsoft centric crap. The moment they pick up that shareware disk for $5.00 and ask, "will this work on that box I'm buying?" will be the kiss of death on that sale. Add in the fact that sales people at WalMart aren't the pick of the crop and mess up even Windows technical issues and it is a recipe for a PR disaster.
Agreed, selling linux succesfully requires educating the buyer that it is not windows and what it's advantages (zero cost, less vulnerability to shitware, availibility of a lot of very good free software from the distros repositries) and disadvantages (inability to run the software they are used too and they see on the shelves in every computer related shop, lack of availibility of support from your more geeky but still MS using friends) are and letting them make an informed choice. Sadly this is hard when they don't even know what an OS is.
Tricking people into buying linux when it is not right for them will only breed resentment, especilly when they have to pay three times as much to buy windows after the fact as they would have to buy it with the PC.
I would only reccomend linux on the desktop to anybody if I knew appropriate software for the tasks in hand was availible and either:
* I was going to be supporting it
* I knew someone with appropriate linux knowlage was arround to support it.
* The box was being used for a very limited set of tasks with little prospect that it was going to be used for more
Also even if I reccomended linux on the desktop if I thought there was a reasonable (more than 1 in 3) chance the box will ever be used for something requiring windows then I would have to reccomend getting the windows license anyway due to the aforementioned huge price differential between OEM and retail (afaict most if not all windows volume licences are upgrade/downgrade only).
A key for XP corp?
The trouble is afaict only failures in places like nuke plants tend to get analised rigorously enough to diagnose this sort of thing.
There are so many possible causes of failure in electronics and there is so little reporting of how long equipment lasts and how it fails that drawing meaningfull conclusions on whether this is having a significant effect on the lifetime of consumer electronics is difficult to measure.
Most non-x86 architectures are moving back to in-order execution. Compilers are good enough that they put instructions far enough away to avoid dependencies (something much easier to do when you have lots of registers) and the die space savings from using an in-order core allows them to put more cores on each chip.
OTOH most non x86 architectures are used in environments where it is feasible to compile for the specific chip.
to win in the PC market chips must be able to perform reasonablly well on code compiled by compilers targetting older chips as that represents the code that most people will be running.
I always thought that the reason MS bought hotmail was to jumpstart the passport network by making sure a lot of people already had accounts.
returning pages of memory to the OS is possible at least on windows on linux.
The problem with making a memory manager (defined here as a peice of code that allocates pages from the OS and suballocates them to meet the demands of other parts of the application) return memory to the OS is that memory has to be allocated and returned to the OS in complete pages. In other words the already complex buisness of managing a heap that stores random small locations gets even more complex as the manager must try to empty pages so they can be returned to the OS.
Yeah, if he actually filed usable bugs, they'd get fixed, and then he'd have nothing to whine about any more.
The problem is it is very hard to file usable bug reports for "works fine at first but tends to suck up ever more memory and grind to a halt over long sessions in my environment". You really don't know if it is firefox core or some extension or plugin that is to blame nor if it is firefox core what part of it.
If a developer can't reproduce a bug quickly and reliablly it is very hard for them to find where the bug is and to confirm they really have it fixed.
Also mozilla is known to ignore reproducable crash bugs because they don't affect enough people. The fact they are having to do that speaks volumes about thier code quality IMO.
CIH was destructive to motherboards and certain disk structures (which were generally repairable provided the hard drive was fat32 and the first partition was over a certain size) but that was about all the damage it did.
Far more insiduous would be malware built into the bios by the manufacturer which allowed secret commands to be used to read out data through the systems onboard network adaptors.
Why oh why does Microsoft still automatically run software off any disk that's inserted into your PC? Surely decades of floppy-carried virii should have convinced them of what a frigging stupid idea that is?
When CD autorun was first introduced CDs were read only media produced only by big and presumablly trusted companies. So the risk was pretty low. Hard drives equally were rarely shared with anyone but very trusted people so they were pretty low risk too. I don't think floppy and similar drives ever had autorun.
Of course things have changed now putting MS between a rock (vulnerability to malware) and a hard place (pissing off users who have become used to autorun).
If you think this is bad you might like to know that on risc OS just viewing the contents a directory was enough to have code automatically run on your system.
And then under EU law they can't fill that post for a certain length of time (1-2 years?)
But is there anything stopping them just reorganising a bit so that no one person is doing the job of the person let go or setting up a shell company to employ the person and have the work done on a contract basis by the shell company?
I've added lots of drives to Windows machines and it never occured to me to try to access them without formatting them. Do these come preformatted?
In my experiance bare drives don't but drives ready mounted up in USB caddies do.
Sure you could reformat it to remove stuff but by the time you get to the format screen you are probablly already infected.
agreed. I know a woman in her early 20s who does not work because she has calculated she would have to make at least $12/hr full time to make up for food stamps and welfare and pay for child care while at work.
This seems to be a symptom of the benefit system not taking account of expenses related to holding down a job (childcare, fuel etc) and/or not having a sane taper off system.
To them, it's either a full-blown welfare state or Dickensian workhouses. The idea that life isn't binary, and that there might be alternatives between extremes seems unfathomable.
A lot of issues seem to gather extremists on both sides.
A good policymaker needs to identify problems in the current system (and they undoubtablly are there) and try and sort them out but equally they need to avoid following extremists who demand rash changes without really thinking them through.
What do you think would happen to theese slums if the governement suddenly stopped the welfare and the subsidised housing as some people seem to be suggesting? Do you think the people would mostly find gainfull employment or do you think there would be a huge rise in low level crime, begging and people living on the streets? I would bet on the latter.
by encouraging work and discouraging poor decisions.
On the other hand once people end up living on the street begging to survive afaict it is very very difficult for them to get back into work because they won't have an address for mail, they wont be able to keep properly clean and so on.
BTW vista buisness OEM comes with downgrade rights to XP pro and doesn't seem to be any more expensive than XP pro.
IIRC it was primerally there to prevent ripping of the CD.
That post looks like it is complete speculation. Do you have any evidence to the contary?
I guess thats one way to try and persude people to fork out for the upgrade. Unfortunately since many of them probablly won't anyone who develops in java and has OS-X as a target is probablly going to be stuck with 1.5.x for some time.
Also it looks like that timeline is just speculation based on past behaviour not actual data from apple despite the title.
No but you can download up to date JVMs for windows from sun, you can't do that for OS-X.
How much of this is suns fault and how much of it is apples fault I have no idea.
Letting ordinary users set the system time isn't particularlly smart imo. If time configuration is done by someone who dosn't understand timezones you will end up with fucked up timestamps.
As ATI has started doing with the R500 and R600 series. So far the technical docs they've released basically only cover the frame buffer, but they've stated that the 3D docs will be forthcoming. (2D acceleration is done using the 3D engine.)
:(. Looking at the wikipedia article it seems to be one of the chipsets you mentioned. Even a £20 geforce FX5200 runs it fine so it isn't just bleeding edge cards the intel integrated is losing to. They are probablly good enough for desktop eye candy though.
Lets hope they really follow through. If they do and a workable opensource 3D graphics driver comes out of it and gets included in the major linux distros I will be specifying ATI rather than Nvidia when I buy machines.
For now, it looks like the 3D graphics hardware with the best open source support is the Intel GMA-X3000 integrated graphics in the G965 and GM965 chipsets. The performance is lower than the bleeding-edge ATI and Nvidia parts, but it's adequate for most purposes.
I know whichever intel chipset apple put in the macbook isn't even capable of running planetpenguin-racer acceptablly
For now i'm sticking with nvidia for any machines where I care about graphics performance because the impression I get from linux forums is that they are less hassle to set up on linux than ATI ones.
Using apt-get dist-upgrade blindly is indeed nuts but if you read the release notes and pay attention to what apt plans to do you are pretty safe and unlike with fedora upgrading with the package manager is the supported and reccomended way to upgrade.
That is macrovisions most famous defective restricted media system (mainly because it was one of the first defective restricted media systems created) but it is far from thier only one.
IIRC DSL is woody based. If that is still the case then unless they have been doing a lot of work themselves the software is probablly littered with security holes. Maybe acceptable for a livecd but not for a main desktop OS.
Most video display technologies need an amount of stabilization to display images accurately. That requires a constant current load, unless you want to go back to the 50's, 60's, and 70's where the TVs have to warm up before you use them.
I call BS, I don't think I have ever seen a TV or monitor new enough to have standby that couldn't come up to an acceptable picture in a couple of seconds from a cold (no power) start.