Everything in RAM does not have to be saved. If you look at a computer with 1GB of RAM you see that much of what is in RAM is just a disk buffer cache and pages of executable code that were paged in. This stuff can simply be trashed as it will be paged back in after the machine wakes up. Saving LESS to RAm can actually make start up FASTER because less data neds to move between flash and RAM.
When I do any kind of analysis and notice that one company does not want benchmarks published I can only assume that is because they will com out on the bottom and don't want that fact to be published. If they came out on top, of course they'd want everyone to know that. OK now prove me wrong but you will have to publish some benchmarks to prove your case.
But in this case, who cares? No one uses VMs because they need to run something faster. VMs have become popular because we have an excess of hardware performance and can therefor consolidate. Or maybe we use VMs as a test tool or whatever but never for performance senitive applications.
But still, it should be clear to anyone why a company would not want bechmarks of it's products published. They can explain their reasons forever but everyone will always suspect the real reaon is that the benchmarks don't look so good.
No, GPS is designed such that no precision timming is needed inside the reciever. The precisionclocks live inside the spacecraft. The spacecraft trans mit digital messages containing only the orbital parameters and the time of day. Recivers only have to look at the low data rate digital messages and compare the time field in each. The reciever then solves a set of equaations and gets location. Solving can take a good deal of time on a low end device so the GPS really tells you where you were a second ago not were you are unless you are not moving.
This device was never inteneded as a comercial product the intended market is radio hobbyests who need a "front end" so they can experimant with the software. There are other front end that cost alot less. The bottle neckis always bandwidth. How much of the spectrum can you stuff into a PC?
If you assume that all drive have equal probibilty of failure then if yo own seven 100Gb drives you have a 7X greater chance of failure then if you owned one drive.
But if you organized those seven 100GB drives into a 500GB RAID5 then you would have to loose two drives before you would loose data.
Typically small file servers do not require a lot of CPU power. The comsumner level NAS boxes you buy for home or small office use use very low end CPUs not much different from what you find on a home router/firewall.
But if youinstall a large scale RAID array and four gigabit eethernet interfaces and hook it up to all 250 PCs in an office where all therusers keep their data on the server then yes you might want a modern server class computer.
It does NOT mater how hard Vista is to install. Hardly anyone will install it. Almost all of the copies of vista will be sold as pre-installed on new hardware. Only the "true geeks" and MS fanboys will buy and install the retail Vista on existing hardware. Vista will be a huge success for the same reason XP was: It will come preinstalled on 90% of all new computers.
I've watch many IT people install OSe on PCs. Every time they simply roll on a preformmated disk image. Most simply keep a stack of pre-image hard drives. If you have 1,000 or so PCs you don'r other with install CDs.
Here is the logic: The goernment pays for the research. If the results are not published it would be a waste of money so of course it gets published. But why let some someone else take your reasearch, turn it into a product and patent it? Rather then that the government can file for the patent and licenese it to any number of companies and use the income from the licenese to do more research. As long as the lincenses are non-exclusive everyone wins. Universities do the same thing
I read the Flash presentation. It's at such a gross high level and lacks any details other then the names of some modules. There is no new information here. OK there are lots of parts and it looks complicated but Linux has all those parts too (except for the bagage that is needed to support old Mac OS9 aps)
My opinion is that mac OSX is very well designed if you beleive that in the future there is be more and more "cores" inside most notebook and desktop computers. OSX is well setup to take advantage of 8-core machines. It should scale well.
Of course your pupil is only 1.5 mm and cameras can already outperform the human eye in many respects.
The above is exactly correct. but those cameras that perform so well have an aperture greater than 1.5mm. Actually the human eye can open up to as wide as 7mm in the dark with 5mm being more typical for many people. But yes 1.5mm is normal in good light
An optical system's resolving power is proportional to the diameter of the optics over the wavelenght of light. Bigger is always better
Yes, if all your data will fit on one disk drive, then you can use one disk drive to back up your data. But let's say you have 5TB of data and the largest disk you can buy is 500GB. Do you plug in 10 USB disk drives every night?
Think about a typical medium sized bussines that has say 100 computers in the office. On any one day there are not likely to be all that many changed files but still you have to do a "level zero" backup periodically. You can't backup 100 computers to one 500GB USB drive. No you are going to need automated tape handling system
The numbers of gigabytes will change over the years but whenever you have more then 5 or so disk drives worth of data to backup you will be wanting a tape system.
So why wouldn't you just use Debian if you want a server linux distro? What will Ubuntu provide over Debian for a server?
Ubuntu adds a layer of quality control and most impotantly a predictable release schedule. With Ubuntu the idea is that you do NOT pull packages out of some Internet site that job is done for you by the people who distribute Ubuntu then they test it for 6 months and then they offer a complete system. It takes a lot of the quality control functions off the end user.
Nothing to stop you from bypassing this layer of QC either. but no Solaris admin would even concider doing an "aptget" then gong live with whatever he got, not they'd run on a parallel test system for months first. Sun's customers are not home/hobby users
A "fork" will happen no matter what. What needs to be prevented in an incompatable fork that is non-standard Java. They need to prevent others from taking over the standard. One way that has worked is the way Ada did it. There are many Ada compilers (including the Open Source gcc Ada front end) but the language spec for Ada is closely followed by everyone. What they should do is hold onto the trademark but offer to let anyone use it whos Java implementation passes a large test suit. By "large" I mean something like a 100,000 lines of Java code. Should there ever be a problem they can extend the test suite.
After all the whole point of OSS it so people can change the software. The first time one guy checks it out of CVS and make an edit you have a "fork" but hopefuly these get checked back into the main brabch after so testing
That's the whole point. If you don't like the UI YOU can fix it on YOUR phone. If you want you can contribute the fix the the world and other users can have it. It Moterola likes your fix they can include it in thier next relese. There is no need to submit your changes and hope they are accepted, yes that's nice as it lets others see your hard work but you can keep them all to yourself too and have a one of a kind phone too.
No more complainning about how the UI could be better. You have the means to fix it. OK but few users have the skill to fix it but itonly takes a few. If you can't write software maybe you can step in and manage some usability studies or write some documentation or something
Building your own is only proffetable if you do it in the "underground economy". If you have to pay rent on the facility, pay payroll and payroll taxes buy insurance, bussines licenese, pay income tax for the bussines and on your salary and so on and so on then you would find it would be cheaper to simply do all this work in China and deliver compleed units to people here.
But yes, building one computer might save you money but building 10,000 willl cost you a lot more money. Bussiness don't buy one computer.
Now consider a microkernel. The filesystem driver is a separate server process. Executing a system call means sending a message to that server and waiting for an answer. Now, what happens if the server is already executing another call ? The calling process blocks, possibly for a long time if there's lots of other requests queued up.
First off I agree with Linus, sort of. I think a hybred would be best. Take a lot of stuff out of the kernel if you can and put it inside it's own addres space. But "balance" and "compromise" is key here the answer is not black or white.
Now to adress the above quote. A DBMS is exactly like a very complex file system. It has many clients reading and writing all at once and performance matters a lot. DBMSes typically run as processes and work well. You can do it by using the process per client method and "pre forking" some server process like Apache does so there is no start up delay. A file system would be a good thing to pull out of the kernel.
I need to look at how Sun implemented ZFS. It is a 128-bit file system that includes the functionality of a volume manager and a RAID system all in one level. Theyu claim it is very fast beating even UFS while adding many, many functions
I would put only hardware device abstraction and process managment inside a kernel but leave where exactly to draw the line up in the air. For example: RAID if you areusing a hardware RAID controller then clearly RAID can't go in userland but what if a user wants to concatinate a SCSI and SATA drive into one logical volume? That can't go into the driver?
OK not black and white.
2500 years ago some greeks invented an attack for which there is no defence. It's called the "Trojan Horse". It goes like this: Make something that looks to be very atractive to your enemy but is actually anything but that. Normal defences and security can't work when the user really _wants_ some program he found to run. The only thing that can work 100% is to take control of the computr away from the user.
Wallmart is not moving the labor back inside it's store. All a customer does when he "builds" a computer is select a few major parts. "Build" is not the right word. "Custom Asembly" might be better. Nobody will be folding sheet metal, spot welding, soldering or stack disk platters in the store. This is smart for Walmart. It lets them offer a wide range of configurations with a smaller inventory.
the interresting part will be to see what software they buy if they have to pay and nothing comes bundled.
I worked with an Army version. They use smoke generators whichare simply a turbine engine with diesel oil spray-injected into the hot exhaust stream. Next they can add very small brass flakes to the oil and make the "smoke" block IR wavelenghts too.
So it works out to $82K per day. But if there are 2000 scientists using the facility that works out to $41 each per day. I suspect that keeping the trash cans empty, roof from leaking and the restrooms clean and stocked costs about the same or maybe more.
A typical overhead rate for the industry is close the 100%. In other words a high tech company typically pays an enginerr his salery and then spend that much again on all the little (and big) things that enable him to do his work, payroll taxes, vacations and so on. $41/day is not much.
Here is how to "put it in hardware". True you really can't but this comes close. You write code to enforce the limits and you burn this into ROM the on-board procesor is designd to execute the ROM code periodically.
This is typical with spacecraft. There is an interval timmer that forces and interrupt that drive the processor to execute coe from ROM periodically. The "safe mode" or "fail safe" or "watchdog" stuf runs there even if a totally bogus set junk gets uploaded in the the RAM. Alows remote recovery from a software error. Same technique could allow for recovery from t driver run amock.
You have a basic problem with transit in Los Angeles. It's a "medium density" city. This means that if you build transit stations such that there are enogh of them so that one is always within reasonable walking distance then each station will searve far to few people and transit user will be anoyed at the slow rate of travel doue to the hundreds of stops at stations where only one or two people get on or off. But if you build fewer stations the travel time is reduced but fewer people can walk to the station.
Ive used mass transit in Tokyo. It works well there because the density is higher
The good thing about a car is that it takes you from door to door. It can follow the absolute best route and does not make any unneeded stops. Mas transit can never have these features because as soon as yu share a ride you have to compromise the route and make more stops. So cars have some good points that people like what are there bad points: (1) they use to much energy. (2) they take up to much road space and cause conjestion. If you could address these two points there would beno need to public transit.
I think #1 and #2 above are both due to the very poor "payload ratio" of the typical car. Basically it is 4,000 pounds of steel used to move 180 pounds of human cargo.
Reduce this by a factor of five or so and you've solved the problem. I can envision little electric powered carbon fiber "cars" that look like bubles around a few seats anf they drive themselves and can even touch bumbers or evan lock "traincar style" on a freeway.
"customers had said that's what they wanted." That's something as a software engineer I learned years ago. Customers don't really know what they want. What you have to do is work with the and get to know them well enough that you get to know what they need.
Aperture was not a total failure. It does most of what isneeded but Version 1.0 was not at all ready to be realeased. Aple should have done what Adobe did with Lightroom. They called the first release "Beta" and made it a free download. Adobe gets comments from real customers and no one is upset with Adobe because they didn't pay anything. But Adobe gets free feedback from real users
The other thing in Apple's favor is that no one knows what one of these kinds of programs should do. Spreadsheets are mature, we know what one should do but these "raw workflow programs"? What are they? Apple was breaking new ground and taking a risk. Get them credit for that.
I'm sure slashdot is not represetative. most people here ould think of a computer as an entertainment device and so $3K would seem expensive. For a normal bussines $3K is not expensive compared to the jack hammer and aircompressor someone I know bought and it's triveal compared to his Catapiler backhoe, trailer and dumptruck. A typical plumber bills at $100/hr to jackhammer a cement floor and replace a drainline. and that's dirt cheap when the restraunt kitchen is shut down with a plugged drain the owner will gladly pay double to have his floor dug up at night. Plumbers are not rich yuppies it's kind of the sterotype blue color job.
Now for me. Yes and no. I had Solaris 10 on the computer but "holy s**t" talk about the time required to "fiddle" with it. Took half a day to figure out how to set up a "zone". I switched to Linux and now things are so much simpler. I used to suggest Mac OX for the desktop and Solaris in the server room. But Solaris 10 is gotten just to hard to admin. To much time. Linux is easy. So yes I just swapped out systems to redce the time spent messing with stuff. But both Solaris and Linux are free Open Sourse OSes So we didn't need to spend money.
About the price: These computers have "Pro" in the name. They are clearly targeted to "Profesionals" that means people that are being paid for their work. If you live in California, have a family, a house and two cars -- prety much the "normal" setup. then SOMEBODY, either your clients or you employer is paysing well over $100 per hour to have you sit at the keybord. So even if that keyboard costs $4,000 and has a three year lifetime the computer adds not even one percent to your cost. It's triveal when compared to the cost of the operator. The dual Xeon system I'm using costs about $10K. Big deal I'll use it to make 25 times that much back. And don't forget the tax deductions for capital expense. Again "Pro" means you can put this on a 3 year schedule so the real out of pocket cost is maybe only $2K.
One the other hand if you buy a $3,000 computer to pay games, surf the web and watch DVDs at home it's a total waste of money. OK maybe you can afford to waste $3k? Lots of people can. Heck I bought a sailboat for 15x that amount. No way on Earth to justify a sailboat other then "I just like them."
Slashdot Mirror
Everything in RAM does not have to be saved. If you look at a computer with 1GB of RAM you see that much of what is in RAM is just a disk buffer cache and pages of executable code that were paged in. This stuff can simply be trashed as it will be paged back in after the machine wakes up. Saving LESS to RAm can actually make start up FASTER because less data neds to move between flash and RAM.
But in this case, who cares? No one uses VMs because they need to run something faster. VMs have become popular because we have an excess of hardware performance and can therefor consolidate. Or maybe we use VMs as a test tool or whatever but never for performance senitive applications.
But still, it should be clear to anyone why a company would not want bechmarks of it's products published. They can explain their reasons forever but everyone will always suspect the real reaon is that the benchmarks don't look so good.
No, GPS is designed such that no precision timming is needed inside the reciever. The precisionclocks live inside the spacecraft. The spacecraft trans mit digital messages containing only the orbital parameters and the time of day. Recivers only have to look at the low data rate digital messages and compare the time field in each. The reciever then solves a set of equaations and gets location. Solving can take a good deal of time on a low end device so the GPS really tells you where you were a second ago not were you are unless you are not moving. This device was never inteneded as a comercial product the intended market is radio hobbyests who need a "front end" so they can experimant with the software. There are other front end that cost alot less. The bottle neckis always bandwidth. How much of the spectrum can you stuff into a PC?
But if you organized those seven 100GB drives into a 500GB RAID5 then you would have to loose two drives before you would loose data.
But if youinstall a large scale RAID array and four gigabit eethernet interfaces and hook it up to all 250 PCs in an office where all therusers keep their data on the server then yes you might want a modern server class computer.
I've watch many IT people install OSe on PCs. Every time they simply roll on a preformmated disk image. Most simply keep a stack of pre-image hard drives. If you have 1,000 or so PCs you don'r other with install CDs.
Here is the logic: The goernment pays for the research. If the results are not published it would be a waste of money so of course it gets published. But why let some someone else take your reasearch, turn it into a product and patent it? Rather then that the government can file for the patent and licenese it to any number of companies and use the income from the licenese to do more research. As long as the lincenses are non-exclusive everyone wins. Universities do the same thing
My opinion is that mac OSX is very well designed if you beleive that in the future there is be more and more "cores" inside most notebook and desktop computers. OSX is well setup to take advantage of 8-core machines. It should scale well.
The above is exactly correct. but those cameras that perform so well have an aperture greater than 1.5mm. Actually the human eye can open up to as wide as 7mm in the dark with 5mm being more typical for many people. But yes 1.5mm is normal in good light
An optical system's resolving power is proportional to the diameter of the optics over the wavelenght of light. Bigger is always better
Yes, if all your data will fit on one disk drive, then you can use one disk drive to back up your data. But let's say you have 5TB of data and the largest disk you can buy is 500GB. Do you plug in 10 USB disk drives every night? Think about a typical medium sized bussines that has say 100 computers in the office. On any one day there are not likely to be all that many changed files but still you have to do a "level zero" backup periodically. You can't backup 100 computers to one 500GB USB drive. No you are going to need automated tape handling system The numbers of gigabytes will change over the years but whenever you have more then 5 or so disk drives worth of data to backup you will be wanting a tape system.
Ubuntu adds a layer of quality control and most impotantly a predictable release schedule. With Ubuntu the idea is that you do NOT pull packages out of some Internet site that job is done for you by the people who distribute Ubuntu then they test it for 6 months and then they offer a complete system. It takes a lot of the quality control functions off the end user.
Nothing to stop you from bypassing this layer of QC either. but no Solaris admin would even concider doing an "aptget" then gong live with whatever he got, not they'd run on a parallel test system for months first. Sun's customers are not home/hobby users
A "fork" will happen no matter what. What needs to be prevented in an incompatable fork that is non-standard Java. They need to prevent others from taking over the standard. One way that has worked is the way Ada did it. There are many Ada compilers (including the Open Source gcc Ada front end) but the language spec for Ada is closely followed by everyone. What they should do is hold onto the trademark but offer to let anyone use it whos Java implementation passes a large test suit. By "large" I mean something like a 100,000 lines of Java code. Should there ever be a problem they can extend the test suite. After all the whole point of OSS it so people can change the software. The first time one guy checks it out of CVS and make an edit you have a "fork" but hopefuly these get checked back into the main brabch after so testing
That's the whole point. If you don't like the UI YOU can fix it on YOUR phone. If you want you can contribute the fix the the world and other users can have it. It Moterola likes your fix they can include it in thier next relese. There is no need to submit your changes and hope they are accepted, yes that's nice as it lets others see your hard work but you can keep them all to yourself too and have a one of a kind phone too.
No more complainning about how the UI could be better. You have the means to fix it. OK but few users have the skill to fix it but itonly takes a few. If you can't write software maybe you can step in and manage some usability studies or write some documentation or something
But yes, building one computer might save you money but building 10,000 willl cost you a lot more money. Bussiness don't buy one computer.
First off I agree with Linus, sort of. I think a hybred would be best. Take a lot of stuff out of the kernel if you can and put it inside it's own addres space. But "balance" and "compromise" is key here the answer is not black or white.
Now to adress the above quote. A DBMS is exactly like a very complex file system. It has many clients reading and writing all at once and performance matters a lot. DBMSes typically run as processes and work well. You can do it by using the process per client method and "pre forking" some server process like Apache does so there is no start up delay. A file system would be a good thing to pull out of the kernel.
I need to look at how Sun implemented ZFS. It is a 128-bit file system that includes the functionality of a volume manager and a RAID system all in one level. Theyu claim it is very fast beating even UFS while adding many, many functions
I would put only hardware device abstraction and process managment inside a kernel but leave where exactly to draw the line up in the air. For example: RAID if you areusing a hardware RAID controller then clearly RAID can't go in userland but what if a user wants to concatinate a SCSI and SATA drive into one logical volume? That can't go into the driver? OK not black and white.
2500 years ago some greeks invented an attack for which there is no defence. It's called the "Trojan Horse". It goes like this: Make something that looks to be very atractive to your enemy but is actually anything but that. Normal defences and security can't work when the user really _wants_ some program he found to run. The only thing that can work 100% is to take control of the computr away from the user.
the interresting part will be to see what software they buy if they have to pay and nothing comes bundled.
I worked with an Army version. They use smoke generators whichare simply a turbine engine with diesel oil spray-injected into the hot exhaust stream. Next they can add very small brass flakes to the oil and make the "smoke" block IR wavelenghts too.
A typical overhead rate for the industry is close the 100%. In other words a high tech company typically pays an enginerr his salery and then spend that much again on all the little (and big) things that enable him to do his work, payroll taxes, vacations and so on. $41/day is not much.
This is typical with spacecraft. There is an interval timmer that forces and interrupt that drive the processor to execute coe from ROM periodically. The "safe mode" or "fail safe" or "watchdog" stuf runs there even if a totally bogus set junk gets uploaded in the the RAM. Alows remote recovery from a software error. Same technique could allow for recovery from t driver run amock.
Ive used mass transit in Tokyo. It works well there because the density is higher
The good thing about a car is that it takes you from door to door. It can follow the absolute best route and does not make any unneeded stops. Mas transit can never have these features because as soon as yu share a ride you have to compromise the route and make more stops. So cars have some good points that people like what are there bad points: (1) they use to much energy. (2) they take up to much road space and cause conjestion. If you could address these two points there would beno need to public transit. I think #1 and #2 above are both due to the very poor "payload ratio" of the typical car. Basically it is 4,000 pounds of steel used to move 180 pounds of human cargo. Reduce this by a factor of five or so and you've solved the problem. I can envision little electric powered carbon fiber "cars" that look like bubles around a few seats anf they drive themselves and can even touch bumbers or evan lock "traincar style" on a freeway.
I assume they have documatation for this hardware, if not how else could their own developers write drivers? Simply post this doc to the web.
They could be different and attract a loyal folloing if they suported Open Source. But it is is I'll ignore them.
"customers had said that's what they wanted." That's something as a software engineer I learned years ago. Customers don't really know what they want. What you have to do is work with the and get to know them well enough that you get to know what they need. Aperture was not a total failure. It does most of what isneeded but Version 1.0 was not at all ready to be realeased. Aple should have done what Adobe did with Lightroom. They called the first release "Beta" and made it a free download. Adobe gets comments from real customers and no one is upset with Adobe because they didn't pay anything. But Adobe gets free feedback from real users The other thing in Apple's favor is that no one knows what one of these kinds of programs should do. Spreadsheets are mature, we know what one should do but these "raw workflow programs"? What are they? Apple was breaking new ground and taking a risk. Get them credit for that.
I'm sure slashdot is not represetative. most people here ould think of a computer as an entertainment device and so $3K would seem expensive. For a normal bussines $3K is not expensive compared to the jack hammer and aircompressor someone I know bought and it's triveal compared to his Catapiler backhoe, trailer and dumptruck. A typical plumber bills at $100/hr to jackhammer a cement floor and replace a drainline. and that's dirt cheap when the restraunt kitchen is shut down with a plugged drain the owner will gladly pay double to have his floor dug up at night. Plumbers are not rich yuppies it's kind of the sterotype blue color job. Now for me. Yes and no. I had Solaris 10 on the computer but "holy s**t" talk about the time required to "fiddle" with it. Took half a day to figure out how to set up a "zone". I switched to Linux and now things are so much simpler. I used to suggest Mac OX for the desktop and Solaris in the server room. But Solaris 10 is gotten just to hard to admin. To much time. Linux is easy. So yes I just swapped out systems to redce the time spent messing with stuff. But both Solaris and Linux are free Open Sourse OSes So we didn't need to spend money.
About the price: These computers have "Pro" in the name. They are clearly targeted to "Profesionals" that means people that are being paid for their work. If you live in California, have a family, a house and two cars -- prety much the "normal" setup. then SOMEBODY, either your clients or you employer is paysing well over $100 per hour to have you sit at the keybord. So even if that keyboard costs $4,000 and has a three year lifetime the computer adds not even one percent to your cost. It's triveal when compared to the cost of the operator. The dual Xeon system I'm using costs about $10K. Big deal I'll use it to make 25 times that much back. And don't forget the tax deductions for capital expense. Again "Pro" means you can put this on a 3 year schedule so the real out of pocket cost is maybe only $2K. One the other hand if you buy a $3,000 computer to pay games, surf the web and watch DVDs at home it's a total waste of money. OK maybe you can afford to waste $3k? Lots of people can. Heck I bought a sailboat for 15x that amount. No way on Earth to justify a sailboat other then "I just like them."