Do your homework and read which devices need to support the flag. The devices are TVs, VCRs and other recording devices, etc. How is making my TV able to understand the broadcast flag going to change anything? Well, it only makes sense if the point is to prevent the viewing or recording of flagged content.
Do you honestly believe that anyone can stop me from sending an encrypted copy of a TV show with the broadcast flag in it over the Internet to you? Do you believe anyone can stop you from viewing it?
Their "corporate culture" did create it -- they gave you the tools, the time and the resources to come up with the idea. You don't think they hired you to work on an assembly line did you? They're hiring you for your brain of course.
The vast majority of innovative companies wouldn't exist if it wasn't implicit that employee innovations become corporate inventions. If this doesn't seem fair to you, don't work for a big company and innovate on your own, then tell me the corporate culture wouldn't have helped you do it better.
There's nothing wrong with how Google does things in this sense, to my understanding.
I personally, I might add, didn't take a job recently that required signing over all my ideas and inventions to my employer, including those I worked on in my spare time or that I'd already come up with, but the idea you're presenting here is different altogether.
I remember when Intel first tried to get soft modems introduced to sell faster CPUs. I kid you not -- the internal memos were that if you could sell soft modems which used CPU power to run the modem instead of hardware DSPs and whatnot, you'd save the consumer money on parts you don't produce (modems) and convince them to buy faster CPUs in the long run.
USB has similar links to CPU horsepower, as does IDE, and AGP. Intel has always promoted such technologies to promote their CPU business, nothing wrong with that, but take their ideas with a grain of salt.
That said, I prefered when my sound card had a DSP on board that could produce *quality* Yamaha music or do its own mixing well, without CPU time. How about modems that don't slow down your online gaming?
You do realize that TCP checksumming on a 1Gbit ethernet link takes about 1GHz worth of CPU power?
I have better things for my CPUs to do. CPUs *do not scale efficiently*. That's why we're all moving to multi-core, remember? Dedicated DSPs are much more efficient, even if redundant in cases where they go unused.
In other words, sell soft modems for $10 and real modems for $80 and let the consumer buy the one that works better for them. Ditto for network cards. My home PC is only a Barton 2500+ but with a quality video card and $100 server network card, it performs better in online gaming that most of my friends' 3.2GHz machines.
Many of us played MOOs and MUDs and games like TradeWars even for years. There was a great planet-conquering game too where you would have a province for your local BBS with groups of players forming small governments and you'd launch attacks against other provinces (BBSs) in town and stuff.
You don't see the point of someone who's done the research and leg work and written actual, working code for large scale file distribution critiquing someone else's proposal for doing something similar?
Cohen knows what he's talking about (that is to say, his software works, so he's either very lucky, or understands the system in question), and as such his critiques are probably valid, if not biased.
Images of the returning coffins are not hurtful to the families as they are unnamed in photos; nobody knows who's kids they are, but they're *somebody's*. Seeing that many dead bodies coming back destroys the home front's resolve.
To keep the people resolved, disinformation is produced by the GOP. Its the way things are done... it may be wrong, but it works.
This is just silly -- so since north american whites enslaved blacks for years, there should be no sympathy for the white mother who is slaughtered by an attacker in her home (of any colour)?
These links are fallacious as another poster points out -- past behaviour to another group (or even to yourself) does not justify cruel and unusual torture or destruction of personal property.
It is possible that some Japanese at the time *deserved* to be attacked by bombs, but that the vast majority of those who suffered in the ways described deserved it is unsupportable.
The math shows that A^B is irrelevant was my point.
It may be equally valuable, but not *more* valuable, and requires (in the end) either more network resources or more CPU resources, or at the very least (in a perfectly efficient system), equal resources on both counts to doing things the current way.
What we *need* is good algorithms for determining which blocks are most likely to go out of visibility soon.
I'd love to see a "I'm shutting this off in a few minutes" button on BitTorrent clients so that peers know to prioritize blocks that client has that others don't.
Redundant, by definition, means unnecessary -- its *additional* data.
It is *possible* that you could download the *exact* blocks needed with the *exact* PAR data needed to reconstruct the missing original data and have downloaded the *exact* same number of bits as the original file would have been, yes (if that's what you're getting at). It is, however, highly unlikely to happen.
Not only is it unlikely, but the original file publisher (and all seeds thereafter) have to share *more* data to the average person than they would if they just shared the original file, see?
If I make 2.5GB of FC4 available by BitTorrent, and 1GB of PAR2 data as well, I'm sharing 3.5GB of data needlessly. Why not just share the 2.5GB of actual data?
Well, because you're afraid I'll log out before you're done. That's great, except that as you stated yourself (I think), the best possible outcome for you is to download 2.5GB of data, whether of original data or some mix of data and PAR redundancy. So why not just target the original data?
See my other reply to someone else in this thread for the rest of this rant.
Very nice comment, great idea, and completely false.
If I send you A, and I send someone else B, then you can get B from someone else while they get A from you and I can go on to sending C to someone else again.
This is how BitTorrent works in super-seeder mode.
All those bits are available on the network already, with the same number of bits being transmitted. Your comment only looks good because you haven't counted the bits yet -- how many total bits do I need to send to get the result?
If I send A and A^B to you, I could've just sent A and B instead.
If I send A^B to you instead of A or B, I could've sent B since you already have A, or vice versa.
Remember that the BitTorrent protocol tends to request the least available chunks first, so that on average, all the chunks are equally available to the overall group.
The A^B issue doesn't help. Do the math and prove it, or leave it.
I agree, and I'm a die-hard Linux fan. I'm also a die-hard OpenBSD fan, but for different things.
I love Linux, and as such, I feel pain every time someone says "just fix it yourself". I'm tired of people who know the code inside and out and can't be bothered to document it or help people do something with it or just do it themselves instead of playing with new features nobody cares much about.
Its just selfish, is what it is.
Did you write your software for you and then give it away? Great. Now its in use by a million people, do you care about their opinions? If not, you're a jerk. That's all there is to it. If you don't want the opinions, don't publish it. Period.
We offer colocated corporate E-mail accounts to clients. We routinely reject blacklisted SMTP servers from a couple sources as well as our own internal spam source collection system.
All of them thank us for how little spam they get.
Its mathematically impossible to do this with less data than an original already-compressed stream.
PAR data is additional redundant data to allow reconstruction of files for which not all the original blocks are any longer available.
This is a *real* problem in some cases, mind you, but it requires sending *more* data, not less.
The additional data is either padded onto each block (as they describe it) or as additional blocks (the way RAID5 or PAR works). Either way, you're talking about having *more* data on average.
If no seeds become available *and* all the available peers do not combined have all of the blocks you each need *and* the blocks that are present are sufficient to reconstruct (from their redundant bits) the missing blocks, this becomes useful.
When an ISP is receiving and storing terabytes of E-mail a day of SPAM and wasting their resources doing so when almost none of their users want it, automatically blocking those messages and adding a statement to that effect to their SLA is perfectly legit.
What is actually entertaining is to see software authors claiming its the fault of the person who used -O2 instead of -O on the compiler, and not their own fault for not testing their software well.
I'll admit that there are complexities to certain pieces of code (notably the Linux kernel) that require certain compiler versions and/or bugs to be assumed.
That said, any "ordinary" piece of software should be able to compile with most reasonable optimizations turned on and still function properly. If they don't, either the compiler's broken, or the software is. Its not the Gentoo user's fault for turning on the flag.
Yes, in some cases the flag is silly, in some cases it wouldn't make any speed difference and the user doesn't understand that, but in those cases, the author should bother to document "tested with -O, -O2 and -funit-at-a-time, -O3 made no difference, use other flags at your own discretion" or something.
I'm not a Gentoo fan, but you're obviously not capable of having an actual discussion with someone either -- he did make real arguments for Gentoo and they're valid from his perspective. You'll be right about your views when you actually *understand* his, and still disagree.
This is a common mistake -- although you *could* outsource to some other country for your cotton, the result would be to make that country spontaneously richer than it had been, resulting in massive inflation and then wage hikes and then a loss of savings on the purchase of said goods from said country.
Countries the size of the US *can* have this much of an effect on a foreign country that is significantly smaller. China is notable here -- although *huge* even compared to the USA, China's economy has grown considerably with textile outsourcing, etc. even with state limitations imposed on personal revenues.
In a completely chaotic free market 3rd world country, try to picture the effect of suddenly putting *that* much demand out on the market for product.
Lobbyists are also a problem however, since they prevent actual thinking and debate from happening -- think about the sugar industry. The USA is losing many major candy companies to Canada where sugar is about 1/4 the price it is in the protectionistic USA.
(And my history tells me Canada was always blamed for being protectionists)
Slashdot Mirror
Do your homework and read which devices need to support the flag. The devices are TVs, VCRs and other recording devices, etc. How is making my TV able to understand the broadcast flag going to change anything? Well, it only makes sense if the point is to prevent the viewing or recording of flagged content.
Do you honestly believe that anyone can stop me from sending an encrypted copy of a TV show with the broadcast flag in it over the Internet to you? Do you believe anyone can stop you from viewing it?
I don't.
Their "corporate culture" did create it -- they gave you the tools, the time and the resources to come up with the idea. You don't think they hired you to work on an assembly line did you? They're hiring you for your brain of course.
The vast majority of innovative companies wouldn't exist if it wasn't implicit that employee innovations become corporate inventions. If this doesn't seem fair to you, don't work for a big company and innovate on your own, then tell me the corporate culture wouldn't have helped you do it better.
There's nothing wrong with how Google does things in this sense, to my understanding.
I personally, I might add, didn't take a job recently that required signing over all my ideas and inventions to my employer, including those I worked on in my spare time or that I'd already come up with, but the idea you're presenting here is different altogether.
I call FUD.
I remember when Intel first tried to get soft modems introduced to sell faster CPUs. I kid you not -- the internal memos were that if you could sell soft modems which used CPU power to run the modem instead of hardware DSPs and whatnot, you'd save the consumer money on parts you don't produce (modems) and convince them to buy faster CPUs in the long run.
USB has similar links to CPU horsepower, as does IDE, and AGP. Intel has always promoted such technologies to promote their CPU business, nothing wrong with that, but take their ideas with a grain of salt.
That said, I prefered when my sound card had a DSP on board that could produce *quality* Yamaha music or do its own mixing well, without CPU time. How about modems that don't slow down your online gaming?
You do realize that TCP checksumming on a 1Gbit ethernet link takes about 1GHz worth of CPU power?
I have better things for my CPUs to do. CPUs *do not scale efficiently*. That's why we're all moving to multi-core, remember? Dedicated DSPs are much more efficient, even if redundant in cases where they go unused.
In other words, sell soft modems for $10 and real modems for $80 and let the consumer buy the one that works better for them. Ditto for network cards. My home PC is only a Barton 2500+ but with a quality video card and $100 server network card, it performs better in online gaming that most of my friends' 3.2GHz machines.
Many of us played MOOs and MUDs and games like TradeWars even for years. There was a great planet-conquering game too where you would have a province for your local BBS with groups of players forming small governments and you'd launch attacks against other provinces (BBSs) in town and stuff.
Great fun.
You don't see the point of someone who's done the research and leg work and written actual, working code for large scale file distribution critiquing someone else's proposal for doing something similar?
Cohen knows what he's talking about (that is to say, his software works, so he's either very lucky, or understands the system in question), and as such his critiques are probably valid, if not biased.
The license keys are on that FTP site in two places.
I found the instructions lame myself too, till I searched around a little.
Images of the returning coffins are not hurtful to the families as they are unnamed in photos; nobody knows who's kids they are, but they're *somebody's*. Seeing that many dead bodies coming back destroys the home front's resolve.
... it may be wrong, but it works.
To keep the people resolved, disinformation is produced by the GOP. Its the way things are done
This is just silly -- so since north american whites enslaved blacks for years, there should be no sympathy for the white mother who is slaughtered by an attacker in her home (of any colour)?
These links are fallacious as another poster points out -- past behaviour to another group (or even to yourself) does not justify cruel and unusual torture or destruction of personal property.
It is possible that some Japanese at the time *deserved* to be attacked by bombs, but that the vast majority of those who suffered in the ways described deserved it is unsupportable.
The math shows that A^B is irrelevant was my point.
It may be equally valuable, but not *more* valuable, and requires (in the end) either more network resources or more CPU resources, or at the very least (in a perfectly efficient system), equal resources on both counts to doing things the current way.
What we *need* is good algorithms for determining which blocks are most likely to go out of visibility soon.
I'd love to see a "I'm shutting this off in a few minutes" button on BitTorrent clients so that peers know to prioritize blocks that client has that others don't.
Redundant, by definition, means unnecessary -- its *additional* data.
It is *possible* that you could download the *exact* blocks needed with the *exact* PAR data needed to reconstruct the missing original data and have downloaded the *exact* same number of bits as the original file would have been, yes (if that's what you're getting at). It is, however, highly unlikely to happen.
Not only is it unlikely, but the original file publisher (and all seeds thereafter) have to share *more* data to the average person than they would if they just shared the original file, see?
If I make 2.5GB of FC4 available by BitTorrent, and 1GB of PAR2 data as well, I'm sharing 3.5GB of data needlessly. Why not just share the 2.5GB of actual data?
Well, because you're afraid I'll log out before you're done. That's great, except that as you stated yourself (I think), the best possible outcome for you is to download 2.5GB of data, whether of original data or some mix of data and PAR redundancy. So why not just target the original data?
See my other reply to someone else in this thread for the rest of this rant.
Very nice comment, great idea, and completely false.
If I send you A, and I send someone else B, then you can get B from someone else while they get A from you and I can go on to sending C to someone else again.
This is how BitTorrent works in super-seeder mode.
All those bits are available on the network already, with the same number of bits being transmitted. Your comment only looks good because you haven't counted the bits yet -- how many total bits do I need to send to get the result?
If I send A and A^B to you, I could've just sent A and B instead.
If I send A^B to you instead of A or B, I could've sent B since you already have A, or vice versa.
Remember that the BitTorrent protocol tends to request the least available chunks first, so that on average, all the chunks are equally available to the overall group.
The A^B issue doesn't help. Do the math and prove it, or leave it.
I agree, and I'm a die-hard Linux fan. I'm also a die-hard OpenBSD fan, but for different things.
I love Linux, and as such, I feel pain every time someone says "just fix it yourself". I'm tired of people who know the code inside and out and can't be bothered to document it or help people do something with it or just do it themselves instead of playing with new features nobody cares much about.
Its just selfish, is what it is.
Did you write your software for you and then give it away? Great. Now its in use by a million people, do you care about their opinions? If not, you're a jerk. That's all there is to it. If you don't want the opinions, don't publish it. Period.
We offer colocated corporate E-mail accounts to clients. We routinely reject blacklisted SMTP servers from a couple sources as well as our own internal spam source collection system.
All of them thank us for how little spam they get.
Its mathematically impossible to do this with less data than an original already-compressed stream.
PAR data is additional redundant data to allow reconstruction of files for which not all the original blocks are any longer available.
This is a *real* problem in some cases, mind you, but it requires sending *more* data, not less.
The additional data is either padded onto each block (as they describe it) or as additional blocks (the way RAID5 or PAR works). Either way, you're talking about having *more* data on average.
If no seeds become available *and* all the available peers do not combined have all of the blocks you each need *and* the blocks that are present are sufficient to reconstruct (from their redundant bits) the missing blocks, this becomes useful.
For some reason I don't believe YAKWMHTTCOAPBACST will be catching on as an acronym.
Sorry.
I'd like to point out that here in Canada, the regulations permit the redistribution of OTA (over the air) signals, as long as they are not altered.
That is to say, you can't strip commercials or replace them with your own without paying for the right to do so.
Its not like MLB sells DVDs of the season ... yet.
When an ISP is receiving and storing terabytes of E-mail a day of SPAM and wasting their resources doing so when almost none of their users want it, automatically blocking those messages and adding a statement to that effect to their SLA is perfectly legit.
Get over yourself.
Your parents said you were going to the moon because they didn't like you. It had nothing to do with reality :-)
http://people.google.com/
A friendster replacement.
Windows 98SE is still my favorite Windows desktop OS for non-domain networks.
In a domain configuration, I prefer Win2k.
I hate using Win2k3 or WinXP.
My basic preference is always Linux.
What is actually entertaining is to see software authors claiming its the fault of the person who used -O2 instead of -O on the compiler, and not their own fault for not testing their software well.
I'll admit that there are complexities to certain pieces of code (notably the Linux kernel) that require certain compiler versions and/or bugs to be assumed.
That said, any "ordinary" piece of software should be able to compile with most reasonable optimizations turned on and still function properly. If they don't, either the compiler's broken, or the software is. Its not the Gentoo user's fault for turning on the flag.
Yes, in some cases the flag is silly, in some cases it wouldn't make any speed difference and the user doesn't understand that, but in those cases, the author should bother to document "tested with -O, -O2 and -funit-at-a-time, -O3 made no difference, use other flags at your own discretion" or something.
I'm not a Gentoo fan, but you're obviously not capable of having an actual discussion with someone either -- he did make real arguments for Gentoo and they're valid from his perspective. You'll be right about your views when you actually *understand* his, and still disagree.
Until then, you're just blind.
It doesn't get colder -- it freezes at a higher temperature.
There's a *huge* effect difference.
High-temperature ice is still high temperature.
This is a common mistake -- although you *could* outsource to some other country for your cotton, the result would be to make that country spontaneously richer than it had been, resulting in massive inflation and then wage hikes and then a loss of savings on the purchase of said goods from said country.
Countries the size of the US *can* have this much of an effect on a foreign country that is significantly smaller. China is notable here -- although *huge* even compared to the USA, China's economy has grown considerably with textile outsourcing, etc. even with state limitations imposed on personal revenues.
In a completely chaotic free market 3rd world country, try to picture the effect of suddenly putting *that* much demand out on the market for product.
Lobbyists are also a problem however, since they prevent actual thinking and debate from happening -- think about the sugar industry. The USA is losing many major candy companies to Canada where sugar is about 1/4 the price it is in the protectionistic USA.
(And my history tells me Canada was always blamed for being protectionists)