Now that Linux superclusters have almost swallowed the high-end scientific computing market...
While some simulations parallelize very well to cluster environments, there are still plenty tasks that don't split up like that.
The reason clusters make up a lot of the Top 500 list is that they are relatively cheap and you can make them faster by adding more nodes - whereas traditional supercomputers need to be deisgned from the ground up.
I just feel that i haven't touched my film slrs since i got my digital rebel. Although i see your point that they are a cheap investment, particularly if you've already got the lenses.
Learn about fstops and exposure with your digital, it's liberating to be able to see instantaneous results. Has made me a lot better about how and when to use exposure compensation.
I bought a Shen Hao 4x5 from Badger Graphic a couple of years ago and i'm very pleased with it - although i still don't use it as much as i'd like.
I'd recommend looking at that camera since it seems very well made and relatively inexpensive.
1) Log into staging environment with one username/password 2) Ssh to final environment using my own username password 3) Receive challenge code at login prompt 4) Enter my pin + challenge code into secureid 5) Give one time password to server.
It's a royal pain in the ass, but you'd need to intercept:
1) My account details/password 2) Our team account details/password 3) My secureid card 4) My secureid pin 5) My secureid login - not the same as my *nix one.
I've gotten fairly good results from an Epson 2450.
It sucks for 35mm negatives, but if you are working with a decently large film size then the shortcomings in scanner resolution are much less of an issue.
It's not as sharp as any real film scanner, but unless you work soley in 35mm or APS then film scanners are disgustingly expensive.
I'm not sure what the US standards are for population coverage. I live in a highly populated area and lose t-mobile reception just by walking to the windows on the other side of my office.
Likewise i need to keep my head at a particular angle to keep reception in my appartment.
Perhaps if you aggregate all networks then you get 80% coverage - but i doubt any one does.
The black cart has gone up to 2.19 but the colored ones are only 1.69 a piece. I like the convienece of being able to proof images in seconds and it's certainly not vastly more than costco charge.
I wanted reasonable photo output at a good price and bought an i560.
You can get non-canon cartridges for it for under $2 that produce pretty reasonable results - pushes my cost per page down so low that i can print away to my hearts content:)
You can control all the compiler flags, so you can build binaries that are optimized for size and memory footprint - rather then the desktop world where things are usually speed optimized (funroll-loops please:)
You can also just built the parts of the application that are relevant to your product.
Using any source that's not your own (whether you compile it or not) is a liability - fortunately law makes sure that corporations dont really have to worry about that.
Unless you live in a major metro area, the chances of you having wired phone access would be even lower than your chances of cell access if it hadn't been for the government putting down the cash to install a phone network.
I don't mind the private sector but i do think that broadband providers should have to do an all or nothing approach. Making sure that all their customers have DSL availability.
More often than not, pruported journalists aren't providing journalism anyway.
Many 'news' sources are so biased that they broadcast/publish little but opinion anyway.
Blogs are a step up from the sorts of news sources that blindly reproduce AP articles, but cant ever catch up with those who actually keep correspondants all over the world and try to present what's actually going on.
I do the same thing with my credit cards, but do you really think your average mugger will accept that you shredded the pin because you wanted to avoid the cash-advance interest rates?
Most people carry ATM cards therefore it negates a lot of the possible security arguments about Chip and Pin.
I've just moved to the USA and things are so much more chaotic here. People virtually never check your signature on your credit card, let alone checking that the digits on the card match the stripe. Smart chips are virtually unheard of, but since quite a few stores have smart chip readers but haven't trained their staff on them, i can't use my british cards unless i can explain to them how to use their own equipment....
It would be nice however if there was an independant body who could certify application versions and kernels. They could assess reliability, interoperability and general code standards.
I know that's essentially what redhat do, but as more closed source linux apps and linux hardware come out, it'd be a bonus to be able to have an "Designed for Linux 2.6" badge and have an assurance that it'd all work.
I know fine well that I can inspect the source code of every app running on my laptop, but the despite being a capable coder... i dont. Partly I don't have time and partly I dont care enough.
I'm happy with the mindset of "other people use it so it must be pretty solid".
However, businesses don't really think that way. It makes sense to have a badge that individual distributions use that assures managers that it's reached a particular standard of quality.
Getting hold of someone's pin may not be that difficult, but the pin is no use without the card.
Cloning a smartcard is orders of magnitude harder than cloning a mag stripe. That's not to say that it cant be done, but it presently would require hundreds of thousands of dollars of equipment... unless of course there is some stupid vulnerability in this particular chip design.
From the sketchy detials i've seen, it seems like your PIN gets fed into your CHIP and then your chip releases it's account information along with some sort of code to verify the transaction.
Your chip will never release the private keys that it uses to create said verification code, therefore even if you could created a new card, you'd need to manually extract the keys using something like a tunnelling electron microscope.
For petty thieves the only solution is to steal the actual card, but that raises the stakes and i'm not sure that many credit card thieves would go there.
I can see the more general problem, but it's still a lot better to have one line of SPARC chips instead of two.
Re:It's all about the hardware
on
Sun-isms Debunked
·
· Score: 4, Interesting
Sun are still pushing the high end. They are working on massively multicore (maybe 16 or 32 core) sparc chips, which will then be put into massively multicpu machines. x86 is probably a ways off being able to provide thousands of cores in a single box with a single memory architecture.
However, having a unified OS and desktop across their entire product line will be a bonus.
If you are a small software shop with an employee or two then you are an unlikely litigation target for any broad reaching patent, since the holder is far more likely to go after someone with real money.
The flipside is that if you do become sufficiently successful then you may have to spend some of that to fight them in court.
If you establish a corporation of some kind to do your work as, then you should be able to protect your personal assets from any litigation.
Slashdot Mirror
Now that Linux superclusters have almost swallowed the high-end scientific computing market...
While some simulations parallelize very well to cluster environments, there are still plenty tasks that don't split up like that.
The reason clusters make up a lot of the Top 500 list is that they are relatively cheap and you can make them faster by adding more nodes - whereas traditional supercomputers need to be deisgned from the ground up.
I just feel that i haven't touched my film slrs since i got my digital rebel. Although i see your point that they are a cheap investment, particularly if you've already got the lenses.
Learn about fstops and exposure with your digital, it's liberating to be able to see instantaneous results. Has made me a lot better about how and when to use exposure compensation.
I bought a Shen Hao 4x5 from Badger Graphic a couple of years ago and i'm very pleased with it - although i still don't use it as much as i'd like.
I'd recommend looking at that camera since it seems very well made and relatively inexpensive.
When oil costs spike 10% it has a knock on effect throughout the entire economy.
Being self-sustaining would surely have positive economic effects.
For secure environments i've got to deal with.
1) Log into staging environment with one username/password
2) Ssh to final environment using my own username password
3) Receive challenge code at login prompt
4) Enter my pin + challenge code into secureid
5) Give one time password to server.
It's a royal pain in the ass, but you'd need to intercept:
1) My account details/password
2) Our team account details/password
3) My secureid card
4) My secureid pin
5) My secureid login - not the same as my *nix one.
I'd skip the 35mm - i haven't run a single roll of film since i bought my dSLR a couple of months ago.
:)
Large format is a far better place to invest your money
People do seem to loose sight of the "best technology for the job".
We just dont have anything that can capture 3 gigabytes of data in hundredth of a second... doing the equivilent with film is so easy.
However working with 100 mpixel scans in photoshop is way too painful for me... i think i need to get a 4x5 enlarger.
I've gotten fairly good results from an Epson 2450.
It sucks for 35mm negatives, but if you are working with a decently large film size then the shortcomings in scanner resolution are much less of an issue.
It's not as sharp as any real film scanner, but unless you work soley in 35mm or APS then film scanners are disgustingly expensive.
I'm not sure what the US standards are for population coverage. I live in a highly populated area and lose t-mobile reception just by walking to the windows on the other side of my office.
Likewise i need to keep my head at a particular angle to keep reception in my appartment.
Perhaps if you aggregate all networks then you get 80% coverage - but i doubt any one does.
Lets hope they do something like the UK where 3G bidders are obliged to cover 80% of the population by 2007.
That way citizens of the US get more out of the deal.. as a european living in the us i'm appalled at how frequently i loose reception.
http://www.cellular-news.com/3G/uk.shtml
I bought mine from these guys...
http://www.bulkinkjetcarts.com/cai5inkta.html
The black cart has gone up to 2.19 but the colored ones are only 1.69 a piece. I like the convienece of being able to proof images in seconds and it's certainly not vastly more than costco charge.
I wanted reasonable photo output at a good price and bought an i560.
:)
You can get non-canon cartridges for it for under $2 that produce pretty reasonable results - pushes my cost per page down so low that i can print away to my hearts content
You can control all the compiler flags, so you can build binaries that are optimized for size and memory footprint - rather then the desktop world where things are usually speed optimized (funroll-loops please :)
You can also just built the parts of the application that are relevant to your product.
Using any source that's not your own (whether you compile it or not) is a liability - fortunately law makes sure that corporations dont really have to worry about that.
Well my laptop emerging kde is a far more effective toaster than the funky chrome one in my kitchen.
Unless you live in a major metro area, the chances of you having wired phone access would be even lower than your chances of cell access if it hadn't been for the government putting down the cash to install a phone network.
I don't mind the private sector but i do think that broadband providers should have to do an all or nothing approach. Making sure that all their customers have DSL availability.
More often than not, pruported journalists aren't providing journalism anyway.
Many 'news' sources are so biased that they broadcast/publish little but opinion anyway.
Blogs are a step up from the sorts of news sources that blindly reproduce AP articles, but cant ever catch up with those who actually keep correspondants all over the world and try to present what's actually going on.
In theory doesn't that put you at more risk?
I do the same thing with my credit cards, but do you really think your average mugger will accept that you shredded the pin because you wanted to avoid the cash-advance interest rates?
Most people carry ATM cards therefore it negates a lot of the possible security arguments about Chip and Pin.
I've just moved to the USA and things are so much more chaotic here. People virtually never check your signature on your credit card, let alone checking that the digits on the card match the stripe. Smart chips are virtually unheard of, but since quite a few stores have smart chip readers but haven't trained their staff on them, i can't use my british cards unless i can explain to them how to use their own equipment....
I agree with you completely.
It would be nice however if there was an independant body who could certify application versions and kernels. They could assess reliability, interoperability and general code standards.
I know that's essentially what redhat do, but as more closed source linux apps and linux hardware come out, it'd be a bonus to be able to have an "Designed for Linux 2.6" badge and have an assurance that it'd all work.
But don't you carry an atm card at the moment?
Violent muggers have walked people to cash machines for years and forced them to withdraw cash - really no difference here.
I know fine well that I can inspect the source code of every app running on my laptop, but the despite being a capable coder... i dont. Partly I don't have time and partly I dont care enough.
I'm happy with the mindset of "other people use it so it must be pretty solid".
However, businesses don't really think that way. It makes sense to have a badge that individual distributions use that assures managers that it's reached a particular standard of quality.
Of course in practice that's highly non-trivial.
Getting hold of someone's pin may not be that difficult, but the pin is no use without the card.
Cloning a smartcard is orders of magnitude harder than cloning a mag stripe. That's not to say that it cant be done, but it presently would require hundreds of thousands of dollars of equipment... unless of course there is some stupid vulnerability in this particular chip design.
From the sketchy detials i've seen, it seems like your PIN gets fed into your CHIP and then your chip releases it's account information along with some sort of code to verify the transaction.
Your chip will never release the private keys that it uses to create said verification code, therefore even if you could created a new card, you'd need to manually extract the keys using something like a tunnelling electron microscope.
For petty thieves the only solution is to steal the actual card, but that raises the stakes and i'm not sure that many credit card thieves would go there.
Well the thing is that the single monster architecture does work for Oracle and SAP. And that's enough to justify the huge price size.
The upside to that is that Sun and Fujitsu are partnering on the SPARC front so they can combine their chip and development resources.
_ fujitsu/
http://www.theregister.co.uk/2004/06/01/sun_sparc
I can see the more general problem, but it's still a lot better to have one line of SPARC chips instead of two.
Sun are still pushing the high end. They are working on massively multicore (maybe 16 or 32 core) sparc chips, which will then be put into massively multicpu machines. x86 is probably a ways off being able to provide thousands of cores in a single box with a single memory architecture.
However, having a unified OS and desktop across their entire product line will be a bonus.
Hell, I'd pay extra if they could guarantee a minimum number of minivans swallowed...
I doubt you'll get a Service Level Agreement with a consumer broadband connection...
If you are a small software shop with an employee or two then you are an unlikely litigation target for any broad reaching patent, since the holder is far more likely to go after someone with real money.
The flipside is that if you do become sufficiently successful then you may have to spend some of that to fight them in court.
If you establish a corporation of some kind to do your work as, then you should be able to protect your personal assets from any litigation.