I would be very surprised if they can successfully and consistently measure the movements that result in diphthongs, as the muscle movements involved are extremely minimal.
Diphthongs, by the way, are why interfaces that attempt to "read lips" without the benefit of a phonetic dictionary of some kind (and preferably a context one as well) always fail miserably, to the eternal chagrin of the CIA.
Actually, the values returned by main() are not laid out by the C specification. There is an arbitrary practice among many programmers that calls for the use of a return value of zero to indicate successful execution, but it is by no means mandatory.
And, actually, enclosing the filename for #include statements in quotation marks is often a better choice than enclosing it in angle brackets, since it makes it trivial to move from system-wide headers to project-specific headers, simply by copying files. Quotes search the local directory, then the common director-y/ies, whereas angle brackets only do the latter
Dual OC-3's. Our NOC is a NICE place to be, especially since I bought five 24x CD burners and put them in there. Can we say, "RedHat ISOs in twenty seconds or less"?
Inform me, in that case, why it is non-trivial to simply increase the size of the work sent to the fullservers?
One possibility I could see is that many needed (unprocessed) keys are non-contiguous. If so, have fullservers keep track of what keys they have sent but not received in the same way the keymaster does now. In order for this to work, clients need to stick to the same fullserver. With that in mind, at install-time, fetch a list of fullservers and stick them in an.ini, and rank them according to ping times. Now the only way keys get repeated is if somebody manually edits that.ini file [note 1], or if a fullserver goes down. If a fullserver goes down, the keys "allocated" to it do not get allocated to the other fullservers for a long, long time to ensure that the fullserver is REALLY down.
Advantages:
Keymaster could be run on a 56k modem, bandwidth-wise. (Although it might be interesting to harness the logic chips of a 56k modem to build a keymaster... hmmmm)
There's a very high probability that statsbox-iii would only have to query only a few of the fullservers in order to generate stats for any given user, reducing bandwidth usage, especially if there exists a terse "No data received from that user" reply.
Because keys that are out of linear order are ignored, there's a high probability that chunky data (1010101101, where 1 is returned, 0 is not), which is likely produced by an unreliable client, is assigned multiple times. The threshold for reporting could be played with, but I would suggest five as a starting point (fullservers only report streaks of five or more keys within their assigned keyspace that are as yet unchecked).
Just a humble little thought, Jouster
Note 1: Stats only show and only count the last submission of any given key (i.e., delete old entries for that key if you, as the fullserver, received two replies, and programmatically ignore the earliest of multiple replies during statistics aggregation), so it would be pretty pointless to switch fullservers by hand.
In actuality, gambling regulations wouldn't apply to that scenario. Because d.net uses a 100% non-random criterion to select a winner (namely, whether they found The Key), it's not a contest, it's a competition.
That said, IMHO, would most people pay to participate in d.net? No.
I find this an excercise in futility; if the protocols used to transmit the data are not available to/.ers, we cannot suggest a scheme that would be meaningful. If the blocks are indexed, and all that's returned is an "index <X> complete" message, then a system of proxies sending message like "indexes 1217-1250 completed by my subnodes" to the main server once every hour makes sense. If, on the other hand, the bulk of the data is used to verify that processing actually occured, and that it occured with the official client (which I suspect is the case), we would need to know details of the data being passed back and forth in order to help.
I know that I, for one, have boxen and bandwidth to pull off 3 Mb/s of CPU-intensive network traffic 24/7, but I'm not about to devote my precious resources to something that I don't understand, especially when I haven't even had the chance to ascertain that a solution that utilized my donated resources was, in fact, the best one.
Hmm, I wonder what the implications of this are for a minor buying--ahem, licensing a Microsoft product. Since they are minors they can disaffirm the "contract" entirely... Perhaps I'll have to have kids buy my software in the future for me.:grin:
Under those circumstances, to express it in C: Dad->kids[0]->identity == Dad->identify In other words, if you're ordering your kids to violate the law, and they do, YOU violated the law that they violated.
If you are under eighteen, you can't legally agree to the terms at all. In effect, even if you click "agree", or otherwise indicate your agreement, you didn't. As invalid as it would be if your ten-year-old tried to sign a mortgage contract.
Summary: If you are under eighteen and use the software, since you can't agree to the license, you are using the software without a license. As such, you are in violation of the license, but, since you are a minor, the charges are fairly pointless. Of course, if an adult was standing nearby and said to use the software, they're contributing to the delinquency....
And, might I add, in the data:
{ 1, 2, 6, 7, 8 },
over 50% of the elements are "above average."
So even if your misreading had been correct, you would still have been wrong.
--J
Napster has the shittiest user interface I've ever seen. Its protocols are from the mid-1980s. And its chat rooms are pointless. I love it. Napster is the single most powerful concept brought to bear yet in the crowded mp3 retrieval world. As far as RIAA is concerned, they already fought this battle once and lost. The courts ruled the Diamond Rio was not illegal; simply because something could be used for illegal purposes doesn't make it illegal. In actuality, many (though definately not most) of the files exchanged over Napster have been authorized for electronic distribution. Thus, it has legal uses, and cannot thus be declared illegal. Just a thought. Jouster
Slashdot Mirror
I would be very surprised if they can successfully and consistently measure the movements that result in diphthongs, as the muscle movements involved are extremely minimal.
Diphthongs, by the way, are why interfaces that attempt to "read lips" without the benefit of a phonetic dictionary of some kind (and preferably a context one as well) always fail miserably, to the eternal chagrin of the CIA.
Jouster
Thank you. I stand corrected.
Jouster
's a quick link.
Jouster
Actually, the values returned by main() are not laid out by the C specification. There is an arbitrary practice among many programmers that calls for the use of a return value of zero to indicate successful execution, but it is by no means mandatory.
And, actually, enclosing the filename for #include statements in quotation marks is often a better choice than enclosing it in angle brackets, since it makes it trivial to move from system-wide headers to project-specific headers, simply by copying files. Quotes search the local directory, then the common director-y/ies, whereas angle brackets only do the latter
Jouster
ROTFL, dude, I lost it when I read that!
For those who are lost, he's referring to NASA's recent failure to convert between metric and Imperial units in the loss of the Mars Climate Orbiter.
Jouster
Dual OC-3's. Our NOC is a NICE place to be, especially since I bought five 24x CD burners and put them in there. Can we say, "RedHat ISOs in twenty seconds or less"?
*grin*
Jouster
Yes, the intent was humorous, albeit apparently completely lost of everyone but me.
Sex with turtles causes chafing in the oddest places.
Jouster
One possibility I could see is that many needed (unprocessed) keys are non-contiguous. If so, have fullservers keep track of what keys they have sent but not received in the same way the keymaster does now. In order for this to work, clients need to stick to the same fullserver. With that in mind, at install-time, fetch a list of fullservers and stick them in an
Advantages:
Just a humble little thought,
Jouster
Note 1: Stats only show and only count the last submission of any given key (i.e., delete old entries for that key if you, as the fullserver, received two replies, and programmatically ignore the earliest of multiple replies during statistics aggregation), so it would be pretty pointless to switch fullservers by hand.
In actuality, gambling regulations wouldn't apply to that scenario. Because d.net uses a 100% non-random criterion to select a winner (namely, whether they found The Key), it's not a contest, it's a competition.
That said, IMHO, would most people pay to participate in d.net? No.
Jouster
But is it 9.80 m/s/s or 32 ft/s/s in our heads?
Jouster
I find this an excercise in futility; if the protocols used to transmit the data are not available to /.ers, we cannot suggest a scheme that would be meaningful. If the blocks are indexed, and all that's returned is an "index <X> complete" message, then a system of proxies sending message like "indexes 1217-1250 completed by my subnodes" to the main server once every hour makes sense. If, on the other hand, the bulk of the data is used to verify that processing actually occured, and that it occured with the official client (which I suspect is the case), we would need to know details of the data being passed back and forth in order to help.
I know that I, for one, have boxen and bandwidth to pull off 3 Mb/s of CPU-intensive network traffic 24/7, but I'm not about to devote my precious resources to something that I don't understand, especially when I haven't even had the chance to ascertain that a solution that utilized my donated resources was, in fact, the best one.
Jouster
Could they just move the project over to SourceForge?
Jouster
Not trolling, just informing...
For those who are lacking context, Dobbie is a Terry Goodkind reference.
Jouster
Under those circumstances, to express it in C:
Dad->kids[0]->identity == Dad->identify
In other words, if you're ordering your kids to violate the law, and they do, YOU violated the law that they violated.
caveat lector, IANAL,
Jouster
If you are under eighteen, you can't legally agree to the terms at all. In effect, even if you click "agree", or otherwise indicate your agreement, you didn't. As invalid as it would be if your ten-year-old tried to sign a mortgage contract.
Summary: If you are under eighteen and use the software, since you can't agree to the license, you are using the software without a license. As such, you are in violation of the license, but, since you are a minor, the charges are fairly pointless. Of course, if an adult was standing nearby and said to use the software, they're contributing to the delinquency....
Just a thought, IANAL,
Jouster
And, might I add, in the data: { 1, 2, 6, 7, 8 }, over 50% of the elements are "above average." So even if your misreading had been correct, you would still have been wrong. --J
Napster has the shittiest user interface I've ever seen. Its protocols are from the mid-1980s. And its chat rooms are pointless. I love it. Napster is the single most powerful concept brought to bear yet in the crowded mp3 retrieval world. As far as RIAA is concerned, they already fought this battle once and lost. The courts ruled the Diamond Rio was not illegal; simply because something could be used for illegal purposes doesn't make it illegal. In actuality, many (though definately not most) of the files exchanged over Napster have been authorized for electronic distribution. Thus, it has legal uses, and cannot thus be declared illegal. Just a thought. Jouster