Anyone reading slashdot already has 4GB or more I call bullshit.
desktop versions of 32 bit windows do not support more than 4GB of physical address space meaning ram usable to the OS is limited to some figure below that (exactly how much depends on the exact hardware configuration). Even if you are one of the minority that uses an OS that supports more the chipset may not. For laptops the situation is even worse, even if the chipset and OS support 4GB of ram 2GB sodimms are not cheap and I'm not aware of any laptops that support more than two modules.
On a system with 2GB of ram (which is the most many users can easilly and economically install) 200MB is 10% of your ram, that is not negligable.
It's just like road rage: if vehicles carried some sort of visible, unique identifier, drivers would afford others the same respect and caution that they expect to receive themselves. vehicles do carry a visible unique identifier. However it is a code that most people will find hard to remember and it ties to the vehicle not the driver. It can also be changed (though at a cost).
So there is a certain percieved anonymity on the road, someone paying a lot of attention can take down your number plate and check it against cars they interact with on the road in future but most people won't and most people don't belive others will. The same applies online, someone determined can find your IP address and other info that they can use to recognise you but they won't recognise you by accident.
Humans are set up to recognise faces and most humans know this so behaving badly when your face is clearly visible to people is rationally a bad idea.
are the admins noses really clean though? I can't imagine why someone would set up an exchange site for pirate material if they weren't using it themselves or being funded by people who were.
I was under the impression that they pushed new versions of that tool out through windows update in such a way that they run automatically once installed on a regular basis.
there are fewer product lines with ATI chips than there are any others, since ATI is now only present in the iMac line of products and as a BTO option on the Mac Pro. Nvidia gets the macbook pro and the default choice for the mac pro -- 2 lines Intel gets the mini and the macbook --2 lines ATI gets the imac, the xserve and a choice for the mac pro -- 3 lines
so ATI wins on number of product lines.
but in terms of total units to apple I suspect intel is the winner followed by nvidia.
Distribution has a fairly specific definition, and that ain't it. Indeed.
So what's ethically wrong with that? It is trying to find loopholes to avoid doing what you would otherwise be required to do. I put it in the same categrory as writing a wrapper round a GPL library to let it be used over an interprocess communication mechanism from non GPL apps.
If it's good, everyone should want to use it, right? Every restriction you add to a license has a cost and a benefit. Choosing an appropriate license for a peice of code you release is about balancing costs with benefits. There is no one license that is right for every code release.
So as part of your business, you give shell access to customers. Now they have the right to ask for the source code to Emacs just because they ran it on your server. If it was normal practice for people to offer access to modified versions of console apps on their servers and use them as a selling point while not giving the source to them to the community I would support use of a similar license for console apps. As it stands for most if not all console apps I think such a license would do more harm than good.
I don't consider a software as a service operator to be an "end user", they are a middleman providing users with the ability to use software that runs on thier servers. Software as a service was being used to essentially do an end run arround the GPLs requirement to provide the source code.
The GPL is and has always been passed to DERIVIATIONS of an application The GPL is about enforcing give and take in the "free software" community, you get to use and modify the communities code on condition that when you release an improved version your users get the source to those modifications under the GPL (and hence can feed the code back to the community if they wish which if there is more than a handfull of them one of them probablly will).
The problem has been that companies are making improvements to free software but getting arround the requirement to release the source to those improvements to thier users by operating on a service model and not giving the code to thier users in any form (either source or binary).
Afaict there is no requirement in the license to release your website data or the source to other apps on your site in the license (though things could get tricky for tightly integrated sites, the license doesn't seem to do a good job of handling them).
Traditionally a second was a fraction of a minute which was a fraction of an hour which was a fraction of a day which was a property of our planets orbit that heavilly effected conditions on the surface.
Trouble is that the length of a day is both slightly variable (while the earths orbit arround the sun and it's pre-existing rotation are the dominant factors they aren't the only ones) and tricky to measure so the second was redefined in atomic terms which were consistant and relatively easy to measure accurately. However humans lives are governed by earth time so among those who cared about consistant timekeeping over long periods there was a need for an adjustment system.
However leap seconds are a pain for our IT infrastructure. They are rare enough, unpredictable enough and small enough that few systems handle them properly so most systems end up treating them as time steps. Time steps are bad for anything that relies on the system clock being a linear time counter. So some people are proposing to get rid of them and not worry about the problem of our time getting wildly out of step with earth in centurys to come.
seriously, your brain cramping because the truth is extreme and its easy to just psychologically deny it, but do the math 200W under load / 1.33VDC core voltage = 150 Amps! Your house probably has a 200 Amp panel, the two are comparable current wise; emotively you'd expect the computer to turn into a china-syndrome pool of molten metal. Yeah the current is quite surprising though you are probablly overestimating at least for normal systems (afaict the figures in the toms hardware article linked by the parent are whole system power consumption). Quad core chips probablly involve higher currents though.
Still current is not the only issue, a switched mode converter must rapidly switch the input voltage, easy enough at 5V or 12V and not too bad at 300V (about what you get from the input rectification/rough smoothing stage of a cheap and nasty PC power supply) but as the voltage climbs further it gets harder and harder.
I know controlling 375KVDC at 80 Amps isn't trivial but neither is controlling the same in AC DC is much harder to break than AC because zero crossings tend to kill arcs.
But more important is that with AC you only need to switch it when you are reconfiguring the power routes or dealing with a fault, so switching can be mechanical and fairly slow (e.g. pull an arc and then kill the arc with a burst of gas). With DC you have to switch it many times per second to get a voltage conversion.
There are cases where DC makes sense.
* Backup power systems for elecronics dominated loads where DC lets you eliminate a conversion step and saves you from the issues of inverters driving rectifiers. Telcos have done it this way for years and finally other datacenters are catching up. * Vehicles for similar reasons to above. * Inside equipment where small transformers would require substantial circitry on the output anyway to get a power supply suitable for driving electronics. * Specialist systems at the very high end of power transmission where other advantages of DC (no need to sync grids, no losses caused by capacitance and inductance, better peak-rms ratio) outweigh the cost and inefficiancy of the conversion stations. The main ones are when a very long or undersea run is involved or where two grids are not synced for political (syncing grids requires a LOT more trust and cooperation than merely exchanging power) or historical (different frequency) reasons.
but for general power transmission and distribution systems I still belive that DC would just increase cost and decrease reliability, efficiancy and safety. Utility transformers while initially fairly expensive have a long life even in outdoor conditions, and have very high reliability and efficiancy.
I was more talking about buisness and technical users. Afaict ordinary home users usually use whatever the OEM shoves down thier throats (with varying degrees of grudgingness admittedly) and have done for some time.
companies stop proposing XP as an OS option. I don't know what the status is where you live but here in the UK if you buy from our main PC chain (PC world) or from the consumer branch of dell UK then most if not all of the machines are vista only. It will get even harder to find XP machines once MS stops offering big brand OEM copies and harder still once they stop offering whitebox OEM copies.
if the collective was to be believed, XP was pointless, because win2k was the pinnacle of Windows OS. Yeah
XP had some new features, some nice (boot time was improved a bit) but most half assed (laggy sound support for dos apps as apposed to none at all, fast user switching that was only availible on systems that didn't use any centralised authentication system). It also had some more bloat and some compatibility issues (though I don't think it was as bad as vista has been)
For the most part like vista it was underwhelming, most 2K users saw little reason to upgrade. Despite this a combination of bundling, hardware compatibility issues and MS moving it to extended support (e.g. you still get security updates free but any other updates cost big bucks) meant that most places did eventually end up on XP.
I anticipate a similar slow and grudging migration to vista.
Indeed while afaict MS doesn't sell very many upgrades (in comparison to OEM copies) to end users or small buisnesses they do sell lots of upgrades (or software assurance which basically ammounts to an upgrade though it comes with a few other benifits and entitles you to more upgrades if they come out while it is still current).
I agree with you few buisnesses are likely to skip vista. One thing that could happen though is if a buisness holds off long enough they may be able to get the windows 7 upgrade as part of the same software assurance purchase.
That "spectacular failure" is selling about 300,000 copies per day. Well:
Home users don't tend to get much choice in the matter, they get whatever the salesperson at the local branch of the big computer chain tells them to get. Larger buisness users buying OEM may as well buy vista since it won't cost them any more than XP and they get downgrade rights which they can excercise using thier activation free corp media/key. Afaict volume license customers can't buy older versions they have to buy the latest and excercise thier downgrade rights.
So of course vista will make lots of sales regardless of if it is any good or not.
Despite all this smaller buisnesses have managed to place enough pressure for major vendors to start offering XP as well. That IMO speaks volumes about how badly received vista has been.
Personally I prefered the simplicity of 2000. XP was really a point release of 2K. There were a few minor improvements, a bit more bloat, a bit of new bling and the introduction of product activation which sucked but many people avoided it by using corp (legally or otherwise) or big brand OEM copies. The main change however was how it was marketed. For the first time ever MS targeted an OS from the NT line at home users.
We have to teach every new developer to share and reuse code and generally avoid continually reinventing the wheel On the other hand bringing in code from outside without understanding the license implications can also cause problems.
This may be the least effective method of debugging in existence. I have had to use it once. I was designing a programmer for the pic18f452 (something I don't reccomend btw, using an existing programmer design is much easier). At well over a hundred clocks before anything interesting happened following the programming waveforms on a scope was not feasible so the only option was to repeatedly go through my code comparing it to the programming spec until things work.
If you're failing to dispose of an object, look at the places where it should be freed and make sure that it is. Generally, there aren't a lot of these places. If you have a dangling reference, it will show up in the form of a crash. probablly.
The resulting backtrace will tell you exactly what object contains a dangling reference, at which point you just have to track down why the reference was not cleaned up when the object it referenced was freed. If you are lucky. If you are unlucky you will currupt memory that has been realocated to something else causing a crash in some code path that is totaly independent of the real error.
Slashdot Mirror
Anyone reading slashdot already has 4GB or more
I call bullshit.
desktop versions of 32 bit windows do not support more than 4GB of physical address space meaning ram usable to the OS is limited to some figure below that (exactly how much depends on the exact hardware configuration). Even if you are one of the minority that uses an OS that supports more the chipset may not. For laptops the situation is even worse, even if the chipset and OS support 4GB of ram 2GB sodimms are not cheap and I'm not aware of any laptops that support more than two modules.
On a system with 2GB of ram (which is the most many users can easilly and economically install) 200MB is 10% of your ram, that is not negligable.
looks like that is only for hacked DS's :(
It's just like road rage: if vehicles carried some sort of visible, unique identifier, drivers would afford others the same respect and caution that they expect to receive themselves.
vehicles do carry a visible unique identifier. However it is a code that most people will find hard to remember and it ties to the vehicle not the driver. It can also be changed (though at a cost).
So there is a certain percieved anonymity on the road, someone paying a lot of attention can take down your number plate and check it against cars they interact with on the road in future but most people won't and most people don't belive others will. The same applies online, someone determined can find your IP address and other info that they can use to recognise you but they won't recognise you by accident.
Humans are set up to recognise faces and most humans know this so behaving badly when your face is clearly visible to people is rationally a bad idea.
are the admins noses really clean though? I can't imagine why someone would set up an exchange site for pirate material if they weren't using it themselves or being funded by people who were.
I was under the impression that they pushed new versions of that tool out through windows update in such a way that they run automatically once installed on a regular basis.
there are fewer product lines with ATI chips than there are any others, since ATI is now only present in the iMac line of products and as a BTO option on the Mac Pro.
Nvidia gets the macbook pro and the default choice for the mac pro -- 2 lines
Intel gets the mini and the macbook --2 lines
ATI gets the imac, the xserve and a choice for the mac pro -- 3 lines
so ATI wins on number of product lines.
but in terms of total units to apple I suspect intel is the winner followed by nvidia.
Distribution has a fairly specific definition, and that ain't it.
Indeed.
So what's ethically wrong with that?
It is trying to find loopholes to avoid doing what you would otherwise be required to do. I put it in the same categrory as writing a wrapper round a GPL library to let it be used over an interprocess communication mechanism from non GPL apps.
If it's good, everyone should want to use it, right?
Every restriction you add to a license has a cost and a benefit. Choosing an appropriate license for a peice of code you release is about balancing costs with benefits. There is no one license that is right for every code release.
So as part of your business, you give shell access to customers. Now they have the right to ask for the source code to Emacs just because they ran it on your server.
If it was normal practice for people to offer access to modified versions of console apps on their servers and use them as a selling point while not giving the source to them to the community I would support use of a similar license for console apps. As it stands for most if not all console apps I think such a license would do more harm than good.
Generally whatever the intent of a rule it is pretty natural to take action when you see your competitor violating it.
If rules are not enforced then those who break them will have a competitive advantage over those that follow them.
I don't consider a software as a service operator to be an "end user", they are a middleman providing users with the ability to use software that runs on thier servers. Software as a service was being used to essentially do an end run arround the GPLs requirement to provide the source code.
The ampere is based on the meter, kilogram and second but in a rather weired way.
most of the other units used electrically are a pretty simple derivation of the ampere, the meter the kilogram and the second.
The GPL is and has always been passed to DERIVIATIONS of an application
The GPL is about enforcing give and take in the "free software" community, you get to use and modify the communities code on condition that when you release an improved version your users get the source to those modifications under the GPL (and hence can feed the code back to the community if they wish which if there is more than a handfull of them one of them probablly will).
The problem has been that companies are making improvements to free software but getting arround the requirement to release the source to those improvements to thier users by operating on a service model and not giving the code to thier users in any form (either source or binary).
Afaict there is no requirement in the license to release your website data or the source to other apps on your site in the license (though things could get tricky for tightly integrated sites, the license doesn't seem to do a good job of handling them).
Traditionally a second was a fraction of a minute which was a fraction of an hour which was a fraction of a day which was a property of our planets orbit that heavilly effected conditions on the surface.
Trouble is that the length of a day is both slightly variable (while the earths orbit arround the sun and it's pre-existing rotation are the dominant factors they aren't the only ones) and tricky to measure so the second was redefined in atomic terms which were consistant and relatively easy to measure accurately. However humans lives are governed by earth time so among those who cared about consistant timekeeping over long periods there was a need for an adjustment system.
However leap seconds are a pain for our IT infrastructure. They are rare enough, unpredictable enough and small enough that few systems handle them properly so most systems end up treating them as time steps. Time steps are bad for anything that relies on the system clock being a linear time counter. So some people are proposing to get rid of them and not worry about the problem of our time getting wildly out of step with earth in centurys to come.
have you ever tried just dropping your sim into an unlocked phone?
seriously, your brain cramping because the truth is extreme and its easy to just psychologically deny it, but do the math 200W under load / 1.33VDC core voltage = 150 Amps! Your house probably has a 200 Amp panel, the two are comparable current wise; emotively you'd expect the computer to turn into a china-syndrome pool of molten metal.
Yeah the current is quite surprising though you are probablly overestimating at least for normal systems (afaict the figures in the toms hardware article linked by the parent are whole system power consumption). Quad core chips probablly involve higher currents though.
Still current is not the only issue, a switched mode converter must rapidly switch the input voltage, easy enough at 5V or 12V and not too bad at 300V (about what you get from the input rectification/rough smoothing stage of a cheap and nasty PC power supply) but as the voltage climbs further it gets harder and harder.
I know controlling 375KVDC at 80 Amps isn't trivial but neither is controlling the same in AC
DC is much harder to break than AC because zero crossings tend to kill arcs.
But more important is that with AC you only need to switch it when you are reconfiguring the power routes or dealing with a fault, so switching can be mechanical and fairly slow (e.g. pull an arc and then kill the arc with a burst of gas). With DC you have to switch it many times per second to get a voltage conversion.
There are cases where DC makes sense.
* Backup power systems for elecronics dominated loads where DC lets you eliminate a conversion step and saves you from the issues of inverters driving rectifiers. Telcos have done it this way for years and finally other datacenters are catching up.
* Vehicles for similar reasons to above.
* Inside equipment where small transformers would require substantial circitry on the output anyway to get a power supply suitable for driving electronics.
* Specialist systems at the very high end of power transmission where other advantages of DC (no need to sync grids, no losses caused by capacitance and inductance, better peak-rms ratio) outweigh the cost and inefficiancy of the conversion stations. The main ones are when a very long or undersea run is involved or where two grids are not synced for political (syncing grids requires a LOT more trust and cooperation than merely exchanging power) or historical (different frequency) reasons.
but for general power transmission and distribution systems I still belive that DC would just increase cost and decrease reliability, efficiancy and safety. Utility transformers while initially fairly expensive have a long life even in outdoor conditions, and have very high reliability and efficiancy.
I was more talking about buisness and technical users. Afaict ordinary home users usually use whatever the OEM shoves down thier throats (with varying degrees of grudgingness admittedly) and have done for some time.
companies stop proposing XP as an OS option.
I don't know what the status is where you live but here in the UK if you buy from our main PC chain (PC world) or from the consumer branch of dell UK then most if not all of the machines are vista only. It will get even harder to find XP machines once MS stops offering big brand OEM copies and harder still once they stop offering whitebox OEM copies.
if the collective was to be believed, XP was pointless, because win2k was the pinnacle of Windows OS.
Yeah
XP had some new features, some nice (boot time was improved a bit) but most half assed (laggy sound support for dos apps as apposed to none at all, fast user switching that was only availible on systems that didn't use any centralised authentication system). It also had some more bloat and some compatibility issues (though I don't think it was as bad as vista has been)
For the most part like vista it was underwhelming, most 2K users saw little reason to upgrade. Despite this a combination of bundling, hardware compatibility issues and MS moving it to extended support (e.g. you still get security updates free but any other updates cost big bucks) meant that most places did eventually end up on XP.
I anticipate a similar slow and grudging migration to vista.
Indeed while afaict MS doesn't sell very many upgrades (in comparison to OEM copies) to end users or small buisnesses they do sell lots of upgrades (or software assurance which basically ammounts to an upgrade though it comes with a few other benifits and entitles you to more upgrades if they come out while it is still current).
I agree with you few buisnesses are likely to skip vista. One thing that could happen though is if a buisness holds off long enough they may be able to get the windows 7 upgrade as part of the same software assurance purchase.
That "spectacular failure" is selling about 300,000 copies per day.
Well:
Home users don't tend to get much choice in the matter, they get whatever the salesperson at the local branch of the big computer chain tells them to get.
Larger buisness users buying OEM may as well buy vista since it won't cost them any more than XP and they get downgrade rights which they can excercise using thier activation free corp media/key.
Afaict volume license customers can't buy older versions they have to buy the latest and excercise thier downgrade rights.
So of course vista will make lots of sales regardless of if it is any good or not.
Despite all this smaller buisnesses have managed to place enough pressure for major vendors to start offering XP as well. That IMO speaks volumes about how badly received vista has been.
Personally I prefered the simplicity of 2000.
XP was really a point release of 2K. There were a few minor improvements, a bit more bloat, a bit of new bling and the introduction of product activation which sucked but many people avoided it by using corp (legally or otherwise) or big brand OEM copies. The main change however was how it was marketed. For the first time ever MS targeted an OS from the NT line at home users.
If a delay of a few milliseconds is a serious concern you really need an OS you can tightly control too.
Usually when doing that you would make your smallest block the size of a pointer, otherwise it is rather hard to track free blocks.
now try running on NFS and removing not just the file but it's containing folder as well. Suddenly things stop working.
And that is in the *nix world, in the windows world (which can't be ignored by most of us) things are much worse.
We have to teach every new developer to share and reuse code and generally avoid continually reinventing the wheel
On the other hand bringing in code from outside without understanding the license implications can also cause problems.
This may be the least effective method of debugging in existence.
I have had to use it once. I was designing a programmer for the pic18f452 (something I don't reccomend btw, using an existing programmer design is much easier). At well over a hundred clocks before anything interesting happened following the programming waveforms on a scope was not feasible so the only option was to repeatedly go through my code comparing it to the programming spec until things work.
If you're failing to dispose of an object, look at the places where it should be freed and make sure that it is. Generally, there aren't a lot of these places. If you have a dangling reference, it will show up in the form of a crash.
probablly.
The resulting backtrace will tell you exactly what object contains a dangling reference, at which point you just have to track down why the reference was not cleaned up when the object it referenced was freed.
If you are lucky. If you are unlucky you will currupt memory that has been realocated to something else causing a crash in some code path that is totaly independent of the real error.