Who else immediately thought "Ben Stiller Show"?...the one where they parody Mutual of Omaha's Wild Kingdom, and tranquilize "a typical North American Street Bum", and ``Jim'' proceeds to attach "a harmless radio collar", which is actually a huge roll of copper wire. The bum revives, and staggers off under the wight of the roll of wire "completely unharmed, and none the worse for wear".
Wish I could find a link to the video for this... it was hilarious.
A Republican in 1886 was not the same thing as a Republican in 2010. One of the most famous Democrats of the time was Jefferson Davis, the secessionist who had became President of the Confederate States of America in an effort to keep the South's economic tradition of slavery alive.
Santa Clara County v. Southern Pacific Railroad was heard in 1886. The governor of California at the time the case was raised was George Stoneman, a Democrat, and it was pushed to the California Supreme Court because their articles of Constitution specifically went out of their way to tax the railroads. The reason it got to the US Supreme Court was due to the Jurisdiction and Removal Act of 1875, which was passed by a Republican Congress and signed into law by Ulysses S. Grant (a Republican and a military officer who served in the American Civil War on the side of the Northern Abolitionists). This law was specifically created to protect former slaves from state-level judicial persecution. http://www.fjc.gov/history/home.nsf/page/landmark_11.html, and the railroad more or less abused it to get a favorable ruling... but it was California that forced the issue.
Next on "Ask Slashdot"... Building lasers to burn out traffic cameras?
"I live in a city with a population in the millions (someplace in the Middle East; the country is not important), and I am mad as hell. There are vigilantes who are going from bad to worse, and I'm sick of all the revenue collection schemes they keep hatching. So far, they've delivered thousands of tickets, and it hasn't done anything to address the accident and mortality rates. After speaking with some of my friends, we decided to take the issue into our own hands: build an infrared CO2 laser system able to fry the traffic cameras, install it on a van at our own expense, and rescue our fellow citizens from these vigilantes and the corrupt city government officials. They are currently in a mode where they are reinvesting their extortion money in increasing the size of their racketeering operation, so now is the time to stop them. We're not rich and we will not ask for our money back. So, I'm asking Slashdot: what would be a workable way to build such a system? It must have an invisible beam not observable by the extortionists or the general public, which might cause word to get back to the extortionists. Any suggestions would be appreciated. This is about technology stopping criminals -- literally."
You could even detect American vs. other passports
All you have to do is hook a RFID detector up to an explosive device in a populated tourist area. Once the RFID detector senses enough unique RFID passport codes within a certain timeframe, BOOM!
Chris demonstrated country-of-origin detection based on passport RFID values. The same thing works with military IDs.
Germany is 1/2 the size of Texas: 357,022 sq. km. vs. 678,054 sq. km., into which they've jammed a little over a quarter of the population of the U.S. 82,282,988 vs. 310,232,863.
What are you smoking that makes you believe the same transportation economics will apply in the U.S. as in Europe?
"...a threat to national security or stability..."
Assuming Google filters this out, then there ought to be a list somewhere of "stuff we filtered out". This will make it much easier to find things to threaten and/or destabilize, should someone want to do that.
Just saying... "One Stop Shopping for Anarchists" is probably not what China had in mind here...
Seems like a lot, until you realize it would all fit in a cube ~230 feet (71 meters) to a side. Assuming you were dumb and didn't reprocess it into usable fuel instead.
Apprentice & dismissing the "normalization fanboy"
-
As the head guy, it would probably be worthwhile for you to hire a skilled database person to sit down with you and pair-program over at least one design for a small project with you; this type of thing is the sort of thing you can pretty easily learn by doing/osmosis. Hiring a guy for a day is pretty cheap, as long as he understands he's there to help you get your sea legs, and that if you get really stuck, there might be a future day or so of work in it for him. In the long run, this will be better than a book.
-
When someone talks about normalization, they generally really mean "3rd normal form", i.e. without replication of non-key field data between records, or to put it in Bill Kent's words: "non-key fields in a record must provide a fact about the key, the whole key, and nothing but the key".
However, when you start talking about denormalization, you are actually probably talking about use of fourth and sometimes fifth normal form instead of third normal form. Unless you are data warehousing, you're not going to see any utility in sixth normal form.
This is still normalization, so you can't throw out the baby with the bathwater if third normal form doesn't apply to your problem.
Unless a database design book describes at least first through fifth normal forms, it's not a database book you want to buy. Scan the index of any book you might think you want to buy using Amazon or Google books, and if there isn't a section on 1NF-5NF normalization, then that book is not a candidate for you.
Since when is having revenue that is less than last year but still positive considered "losing money"?
Ever since publicly held companies have been valued in terms of earnings per share and profit per employee. The second you can do something about short term by firing people (this is what Word Perfect, Inc. did right before the acquisition by Novell, Inc., to inflate the offering price: cut all R&D and fire everyone not contributing to the short term bottom line). The first you can deal with maybe once by trading cash on hand for a stock buyback, or by doing a reverse split... after that, you are out of ammunition and in the same boat two quarters later. If your value is going down by either of those measures, then your stock price will shortly follow it down.
Or if you want a more cynical answer, ever since RIAA and MPAA started claiming that their decreased revenues were from piracy rather than market saturation and/or their products sucking, and declared they were "losing money to pirates".
Well, you got me on this one. I guess if they were actually retrenching, they wouldn't be reporting losses in revenue or be only the second largest software company in the world. So that one's a "no".
Possibly they should get off their butts, and instead of throwing the chair they were sitting on, they should actually retrench.
It's most likely that the app itself is asking for the iTunes username and password. This could nominally be for an in-game purchase, or it could be prompting claiming it was for some other reason, such as "activating the application", where people are willing to put in the information because they've already thrown money at the application. Or it could just be asking for them with no reason given.
It's really hard to avoid this kind of trojaning, if it's either time activated, or activated by the nefarious application checking for an activation token on a remote web site.
One thing that could be helped is if the request is for the nominal in-application purchase; however, the fix lacks a little convenience for the user, which would in turn be likely to reduce the number of in-application purchases: queue the request, and handle the actual purchase through an Apple-supplied application that goes through the queue and has you OK or abort the purchase on a case-by-case basis. This would also have to deal with the enqueued request records not being accessible to the enqueueing application, once enqueued, to prevent rewriting of data in other enqueued requests by a nefarious application to appear to be the request requested, but to actually contain a different payment target as payload.
Most attacks, however, will probably just pop up a request dialog and trust that most users will just be foolish and enter the information requested, and so would not be mitigated by such a (complicated for the user) prevention scheme.
They basically complain that there's still no clear litmus test for patentability because the decision was to vague on the definition of what constitutes "too abstract".
Exactly... "the resolution falls off on non-projected weapons trajectory route splines".
You get very good data for the areas in which you want to fly your birds, and lesser data for where you don't expect to do that. This is necessary to, for example, use inertial guidance rather than active TFR in a cruise missile and keep it below the enemy radar.
It's used in inertial navigation for weapons systems. Interestingly, the inertial navigation software itself is available as source code for download, but the data of the map itself is classified to prevent its use by non-U.S. aggressors. Also, for what its worth, the military data resolution is far better than the 100km between data points, as it is with GOCE, but is the resolution falls off on non-projected weapons trajectory route splines.
See also the geoid from the earlier GRACE observations (animated spinning globe) which were 322km resolution, along with a more technical discussion of GOCE:
The wikipedia article also notes video for bit banging.
I think you could technically make an argument for both sound and video being a form of serial data. We always used it to refer to any time where we externally controlled the timing of dumping stuff into I/O device registers, for example, when doing raster interrupts on a C64 to increase the virtual number of sprites the machine could support, or to reload color registers to "cheat" on the number of colors that could be simultaneously displayed.
My first use of bit-banging was on current loop interfaces, before we had RS-232C terminals available to us. We built bit-slice processors back in the second year of college in a physics of electronics class, we then did both serial and video for keyboard and video out to an oscilloscope, and both approaches were bit-banging. We also bit-banged sound out of the Commodore PET by pounding bits at the VIA (MCS6520) by hacking the output of the CB2 in the NMI handler for the clock to change the shift register contents and bit rate.
One guy got really ambitious and bit-banged a Bell 103C accoustic coupled modem (it could only handle 110 BAUD, though).
If you are intent on bit banging... the available options these days are pretty much limited to microcontrollers, unless you want to end up in huge projects or small modifications on huge projects.
Most of what you can do with these tends to be robotics projects, since there aren't a lot of 8-bit general purpose computers available out there any more.
There are a lot of web sites that provide small source code for special purpose robotics projects which you could apply much in the same way as typing in BASIC games from Compute! or Byte magazine, and then playing with them.
If your intent is to provide a project for a kid, you could do a lot worse than going some place like Weird Stuff, buying up a handful of Compute! magazines and a Commodore 64, a 1541 disk drive, and a box of 10 floppies. There are plenty of analog TV's out there still to use a monitors which are otherwise sitting unloved in peoples garages.
This is somewhat counter-intuitive, but it is nonetheless true, if you are wanting to have a very large number of timers, such as if you wanted a large number of 2 MSL timers because you want to support a huge number of TCP connections.
The problem is that given a number of buckets on the wheel, in order to cancel a timer, you have to either keep them as a doubly-linked list, or you have to run on average 50% of the buckets in the list. For timer lbolts, when the wheel turns to the next slot, you have to run the whole list and check each element to see if its expired, since you can't order the elements in the bucket in time-order to preclude the need to continue running when you hit the first non-expired timer. This is because the elements are (effectively) hashed onto the wheel slot. Even if you have buckets-by-magnitude, it's actually not good enough to represent every resolution equally.
For things like the TCP 2 MSL timer wheels that are used in BSD4.4 and later, for scalability, you are better off going back to the BSD 4.3 tailq-list implementation so that you don't have to run all the entries; as previously noted, most timers are cancelled before they expire, so if the lists are ordered by increasing expiration interval (which they can be, for fixed interval values), your overhead then doesn't go up especially badly as your number of timers outstanding goes up.
I've made this change to the TCP 2 MSL timer implementation at several companies, and got good scalability wins every time. One of those companies was Array Networks, and we sold a commercial reverse proxy cache similar to Varnish, although we did massive kernel work in order to scale it better (as in handling ~36,000 TCP connections a second).
If your school did that, then they probably violated the constitution of your state. The school is supposed to provide all those sorts of things to any student - anything else is discriminatory towards poor students
So they should also provide the answers, since anything else would be discriminatory towards stupid students.
Slashdot Mirror
Who else immediately thought "Ben Stiller Show"? ...the one where they parody Mutual of Omaha's Wild Kingdom, and tranquilize "a typical North American Street Bum", and ``Jim'' proceeds to attach "a harmless radio collar", which is actually a huge roll of copper wire. The bum revives, and staggers off under the wight of the roll of wire "completely unharmed, and none the worse for wear".
Wish I could find a link to the video for this... it was hilarious.
-- Terry
Abraham Lincoln appointed two of those judges.
A Republican in 1886 was not the same thing as a Republican in 2010. One of the most famous Democrats of the time was Jefferson Davis, the secessionist who had became President of the Confederate States of America in an effort to keep the South's economic tradition of slavery alive.
Santa Clara County v. Southern Pacific Railroad was heard in 1886. The governor of California at the time the case was raised was George Stoneman, a Democrat, and it was pushed to the California Supreme Court because their articles of Constitution specifically went out of their way to tax the railroads. The reason it got to the US Supreme Court was due to the Jurisdiction and Removal Act of 1875, which was passed by a Republican Congress and signed into law by Ulysses S. Grant (a Republican and a military officer who served in the American Civil War on the side of the Northern Abolitionists). This law was specifically created to protect former slaves from state-level judicial persecution. http://www.fjc.gov/history/home.nsf/page/landmark_11.html, and the railroad more or less abused it to get a favorable ruling... but it was California that forced the issue.
-- Terry
Next on "Ask Slashdot"... Building lasers to burn out traffic cameras?
"I live in a city with a population in the millions (someplace in the Middle East; the country is not important), and I am mad as hell. There are vigilantes who are going from bad to worse, and I'm sick of all the revenue collection schemes they keep hatching. So far, they've delivered thousands of tickets, and it hasn't done anything to address the accident and mortality rates. After speaking with some of my friends, we decided to take the issue into our own hands: build an infrared CO2 laser system able to fry the traffic cameras, install it on a van at our own expense, and rescue our fellow citizens from these vigilantes and the corrupt city government officials. They are currently in a mode where they are reinvesting their extortion money in increasing the size of their racketeering operation, so now is the time to stop them. We're not rich and we will not ask for our money back. So, I'm asking Slashdot: what would be a workable way to build such a system? It must have an invisible beam not observable by the extortionists or the general public, which might cause word to get back to the extortionists. Any suggestions would be appreciated. This is about technology stopping criminals -- literally."
-- Terry
You could even detect American vs. other passports
All you have to do is hook a RFID detector up to an explosive device in a populated tourist area. Once the RFID detector senses enough unique RFID passport codes within a certain timeframe, BOOM!
Chris demonstrated country-of-origin detection based on passport RFID values. The same thing works with military IDs.
-- Terry
Germany is 1/2 the size of Texas: 357,022 sq. km. vs. 678,054 sq. km., into which they've jammed a little over a quarter of the population of the U.S. 82,282,988 vs. 310,232,863.
What are you smoking that makes you believe the same transportation economics will apply in the U.S. as in Europe?
-- Terry
David Brin beat you by a decade
Here's his 1996 wired column, which he later expanded into a book:
http://www.wired.com/wired/archive/4.12/fftransparent.html
Here's the Wikipedia page on his book "The Transparent Society", which was published in 1998.
http://en.wikipedia.org/wiki/The_Transparent_Society
-- Terry
While Google may be the dominant information indexer, what they're doing doesn't require any special magic.
Someone needs to tell that to Bing and Yahoo... ...just saying, if there wasn't secret sauce, there wouldn't be a dominant player.
-- Terry
"...a threat to national security or stability..."
Assuming Google filters this out, then there ought to be a list somewhere of "stuff we filtered out". This will make it much easier to find things to threaten and/or destabilize, should someone want to do that.
Just saying... "One Stop Shopping for Anarchists" is probably not what China had in mind here...
-- Terry
Think of poor ceiling cat...
-- Terry
Perhaps they are just interested in the applicants ...but not for the purpose of actually hiring them.
-- Terry
100,000,000 gallons of nuclear waste in the US
Seems like a lot, until you realize it would all fit in a cube ~230 feet (71 meters) to a side. Assuming you were dumb and didn't reprocess it into usable fuel instead.
-- Terry
Apprentice & dismissing the "normalization fanboy"
-
As the head guy, it would probably be worthwhile for you to hire a skilled database person to sit down with you and pair-program over at least one design for a small project with you; this type of thing is the sort of thing you can pretty easily learn by doing/osmosis. Hiring a guy for a day is pretty cheap, as long as he understands he's there to help you get your sea legs, and that if you get really stuck, there might be a future day or so of work in it for him. In the long run, this will be better than a book.
-
When someone talks about normalization, they generally really mean "3rd normal form", i.e. without replication of non-key field data between records, or to put it in Bill Kent's words: "non-key fields in a record must provide a fact about the key, the whole key, and nothing but the key".
However, when you start talking about denormalization, you are actually probably talking about use of fourth and sometimes fifth normal form instead of third normal form. Unless you are data warehousing, you're not going to see any utility in sixth normal form.
This is still normalization, so you can't throw out the baby with the bathwater if third normal form doesn't apply to your problem.
Unless a database design book describes at least first through fifth normal forms, it's not a database book you want to buy. Scan the index of any book you might think you want to buy using Amazon or Google books, and if there isn't a section on 1NF-5NF normalization, then that book is not a candidate for you.
-- Terry
"Identification of a Brazil-Nut Allergen in Transgenic Soybeans"
Full article from 1996 New England Journal of Medicine
http://content.nejm.org/cgi/content/full/334/11/688
-- Terry
Since when is having revenue that is less than last year but still positive considered "losing money"?
Ever since publicly held companies have been valued in terms of earnings per share and profit per employee. The second you can do something about short term by firing people (this is what Word Perfect, Inc. did right before the acquisition by Novell, Inc., to inflate the offering price: cut all R&D and fire everyone not contributing to the short term bottom line). The first you can deal with maybe once by trading cash on hand for a stock buyback, or by doing a reverse split... after that, you are out of ammunition and in the same boat two quarters later. If your value is going down by either of those measures, then your stock price will shortly follow it down.
Or if you want a more cynical answer, ever since RIAA and MPAA started claiming that their decreased revenues were from piracy rather than market saturation and/or their products sucking, and declared they were "losing money to pirates".
-- Terry
Is MS losing money ?
"Microsoft reports first YoY revenue slide in company history" ...so I guess that would be a "yes".
http://www.boygeniusreport.com/2009/04/24/microsoft-reports-first-yoy-revenue-slide-in-company-history/
no longer the biggest software company in the world ?
As of close on Tuesday 6 Jul 2010:
Microsoft market cap: 208.75B
Apple market cap:226.24B
http://finance.yahoo.com/q/cq?d=v1&s=MSFT,AAPL ...so I'm guessing that one's a "yes", too...
retrenching ?
Well, you got me on this one. I guess if they were actually retrenching, they wouldn't be reporting losses in revenue or be only the second largest software company in the world. So that one's a "no".
Possibly they should get off their butts, and instead of throwing the chair they were sitting on, they should actually retrench.
-- Terry
I agree that it's probably phishing
It's most likely that the app itself is asking for the iTunes username and password. This could nominally be for an in-game purchase, or it could be prompting claiming it was for some other reason, such as "activating the application", where people are willing to put in the information because they've already thrown money at the application. Or it could just be asking for them with no reason given.
It's really hard to avoid this kind of trojaning, if it's either time activated, or activated by the nefarious application checking for an activation token on a remote web site.
One thing that could be helped is if the request is for the nominal in-application purchase; however, the fix lacks a little convenience for the user, which would in turn be likely to reduce the number of in-application purchases: queue the request, and handle the actual purchase through an Apple-supplied application that goes through the queue and has you OK or abort the purchase on a case-by-case basis. This would also have to deal with the enqueued request records not being accessible to the enqueueing application, once enqueued, to prevent rewriting of data in other enqueued requests by a nefarious application to appear to be the request requested, but to actually contain a different payment target as payload.
Most attacks, however, will probably just pop up a request dialog and trust that most users will just be foolish and enter the information requested, and so would not be mitigated by such a (complicated for the user) prevention scheme.
-- Terry
Thanks...
Guess we all just got trolled...
-- Terry
Who writes that crap blog?
Here is the actual IEEE press release:
http://www.ieeeusa.org/communications/releases/2010/062910.asp
They basically complain that there's still no clear litmus test for patentability because the decision was to vague on the definition of what constitutes "too abstract".
-- Terry
Exactly... "the resolution falls off on non-projected weapons trajectory route splines".
You get very good data for the areas in which you want to fly your birds, and lesser data for where you don't expect to do that. This is necessary to, for example, use inertial guidance rather than active TFR in a cruise missile and keep it below the enemy radar.
-- Terry
The U.S. military already has one of these
It's used in inertial navigation for weapons systems. Interestingly, the inertial navigation software itself is available as source code for download, but the data of the map itself is classified to prevent its use by non-U.S. aggressors. Also, for what its worth, the military data resolution is far better than the 100km between data points, as it is with GOCE, but is the resolution falls off on non-projected weapons trajectory route splines.
See also the geoid from the earlier GRACE observations (animated spinning globe) which were 322km resolution, along with a more technical discussion of GOCE:
http://www.scientificblogging.com/planetbye/grace_goce
-- Terry
Become a developer or set up enterprise deployment or both
Then you can put pretty much any application you want on any device you want, without needing to go through the iTunes App Store.
Developer program:
http://developer.apple.com/iphone/index.action
Enterprise deployment:
http://manuals.info.apple.com/en_US/Enterprise_Deployment_Guide.pdf
-- Terry
The wikipedia article also notes video for bit banging.
I think you could technically make an argument for both sound and video being a form of serial data. We always used it to refer to any time where we externally controlled the timing of dumping stuff into I/O device registers, for example, when doing raster interrupts on a C64 to increase the virtual number of sprites the machine could support, or to reload color registers to "cheat" on the number of colors that could be simultaneously displayed.
My first use of bit-banging was on current loop interfaces, before we had RS-232C terminals available to us. We built bit-slice processors back in the second year of college in a physics of electronics class, we then did both serial and video for keyboard and video out to an oscilloscope, and both approaches were bit-banging. We also bit-banged sound out of the Commodore PET by pounding bits at the VIA (MCS6520) by hacking the output of the CB2 in the NMI handler for the clock to change the shift register contents and bit rate.
One guy got really ambitious and bit-banged a Bell 103C accoustic coupled modem (it could only handle 110 BAUD, though).
-- Terry
If you are intent on bit banging... the available options these days are pretty much limited to microcontrollers, unless you want to end up in huge projects or small modifications on huge projects.
Most of what you can do with these tends to be robotics projects, since there aren't a lot of 8-bit general purpose computers available out there any more.
There are a lot of web sites that provide small source code for special purpose robotics projects which you could apply much in the same way as typing in BASIC games from Compute! or Byte magazine, and then playing with them.
If your intent is to provide a project for a kid, you could do a lot worse than going some place like Weird Stuff, buying up a handful of Compute! magazines and a Commodore 64, a 1541 disk drive, and a box of 10 floppies. There are plenty of analog TV's out there still to use a monitors which are otherwise sitting unloved in peoples garages.
-- Terry
Timer wheels don't scale
This is somewhat counter-intuitive, but it is nonetheless true, if you are wanting to have a very large number of timers, such as if you wanted a large number of 2 MSL timers because you want to support a huge number of TCP connections.
The problem is that given a number of buckets on the wheel, in order to cancel a timer, you have to either keep them as a doubly-linked list, or you have to run on average 50% of the buckets in the list. For timer lbolts, when the wheel turns to the next slot, you have to run the whole list and check each element to see if its expired, since you can't order the elements in the bucket in time-order to preclude the need to continue running when you hit the first non-expired timer. This is because the elements are (effectively) hashed onto the wheel slot. Even if you have buckets-by-magnitude, it's actually not good enough to represent every resolution equally.
For things like the TCP 2 MSL timer wheels that are used in BSD4.4 and later, for scalability, you are better off going back to the BSD 4.3 tailq-list implementation so that you don't have to run all the entries; as previously noted, most timers are cancelled before they expire, so if the lists are ordered by increasing expiration interval (which they can be, for fixed interval values), your overhead then doesn't go up especially badly as your number of timers outstanding goes up.
I've made this change to the TCP 2 MSL timer implementation at several companies, and got good scalability wins every time. One of those companies was Array Networks, and we sold a commercial reverse proxy cache similar to Varnish, although we did massive kernel work in order to scale it better (as in handling ~36,000 TCP connections a second).
-- Terry
They should also provide the answers
If your school did that, then they probably violated the constitution of your state. The school is supposed to provide all those sorts of things to any student - anything else is discriminatory towards poor students
So they should also provide the answers, since anything else would be discriminatory towards stupid students.
-- Terry