Even without buying a WiSpy, you can get a good feel for how much background noise exists with the free inssider app from the same guys - and it works with most wifi cards. I've used it at home to help me select which channel(s) to use.
Also, the guys at work, where we outfit large commercial operations with Wifi also recommend sticking with 802.11g. Most routers can be configured to limit themselves to g. The 802.11n stuff is really fast -when it gets through, but if you're pushing the limits of your link budget, you'll get better overall performance at the slower speeds.
(disclosure, I currently work for www.fidelity-comtech.com)
Russ
Within the last year, we had a new high end FPGA vendor visit us. They mentioned that some of their first customers were people/companies doing these fast trades. They were willing to spend $50k+ per part to increase the algorythm speed by 10 - I think going from 10's of ms to ~1ms, but I'm not sure the exact magnitudes. In some ways, this cost isn't that much, considering that the product development costs are likely $300k+.
I do wonder how widely the algorythms vary across the different players.
There are definetely some large players in this space, and industrial espionage too.
I've heard that many residential electric meters have ERT enabled meters, which use a custom wireless protocol, so that the guys don't have to get out of the truck to read your meter.
So, you'll need some sort of radio, and then hack the protocol (maybe sombody already has?).
I'm lucky to live in a place with nice roads. Back in Virginia, the choices were slim. I think it is key to live somewhere that makes it possible to exercise outside.
So, I just train like the dickens for events over the year. Things like the Triple Bypass or 24hrs of Moab.
Actually, just about any type of endurance exercise seems to be suit the geek personality pretty well. There is a lot of science and performance tracking that goes on to optimize your performance.
I totally agree. H&H (AoE) is the bible. It is the standard reference everyone uses. But, it is also way overkill if all you want is how to make something work.
There are lots of other options out there. My personal choice is "Circuit Design for Electronic Instrumentation" by Darold Wobschall, ISBN:978-0070712300. I like it because most of the time I'm interested in hooking up sensors of various sorts to a micro, for which this book is a perfect fit. It does a great job of explaining the vast majority of the basic stuff you might want to use.
The one caveat is that it is old. So it is not going to have things like spread spectrum (cell phone). Fortunately, most of the basics haven't really changed in over 30 years.
More to the point, the demographic is not for the perfect Pepsi, but for the perfect Pespsis. Malcom Gladwell had a talk at a TED conference, where he explains demographic research that has been done to show that there is usually not a perfect solution, but rather several near optimal solutions. Probably not applicable to all data-mining applications, but certainly appropriate for anything relating to subjective tastes.
Dunno, seems like some earlier research on a similar idea might be more useful - the plenoptic camera. Just one big CCD and a lot of math. CCDs are getting bigger all the time, most people don't really need all the pixels they have already, so sure, why not use some of those pixels for depth. But sticking with one CCD will probably be cheaper.
http://graphics.stanford.edu/papers/lfcamera/
Ever notice how its getting increasingly harder to get a console game where you can do multiplayer on the same box? My bet is its driven by the profits from the on-line fees earned by forcing you to play on-line for multiplayer.
Noticed how thinings are becomming more MMO oriented? Thank the income streams generated by WOW for that.
So I went through and customized the same model with Windows Vista and with Ubuntu, to the same configuration (better LCD and larger battery, but otherwise stock). The Ubuntu model was roughly $854, and the Windows Model was roughly $824 (I might be off by a few bucks). But why is the Ubuntu model more than the Vista model?
grrr.
Its all about the killer app to justify the tablet. I could see the afore mentioned movie player possibly doing the trick. But for me the killer app would be something that lets me draw boxes, circles, triangles, etc, and cleans them up. Think Visio on steroids with OCR.
I've never researched it, but I would be surprised (read: I'm pretty sure), that the major PC vendors get kickbacks from AOL, Symantec, etc for pre-installing trial software on the PC's. I also wouldn't be surprised to learn that the PC makers don't give the consumer a discount for these "bonus" materials. Personally, I hate it. It usually takes me several hours just to clean off the crap.
But the important part is that Linux and BSD distros do not have any extra, shall we say, margin enhancers. So, I doubt that we see significant efforts on the PC manufacturer's part to do much about moving towards non-Windows offerings (ie: show me the money).
Also because humans are so good at reading past our many errors we do tend to put much effort as in programming, because in programming we know that we are right when the program works. Spelling and grammar is much more difficult to find errors in.
Or as I tend to think of it - programmers get corrected every single time the programming language syntax is not correct, but that rarely happens in email.
There are tools in place to help you - spelling and gramatical checkers. Most of them even work as you type. However, fixing little red and green underlines are not required before you press send.
Also the these tools are not perfect, nor by any means ubiquitous (I don't see any red squiggly underlines in the slashdot comment form). Spell checkers are pretty good, but I still find errors in them, mostly due to words missing from the dictionary. Grammer checkers still suck. They often recommend reorgnizing phrases in ways that don't make sense, or more typically, in ways that ruin the thought process/flow of a paragraph.
This takes a few paragraphs to become relevant, so stay with me...
Recently, I've been involved with submission of an SBIR (Small Business Inovation Research) proposal, in which the Gov. gives small companies money to do research - the idea being to help create innovation - the big boys (companies over 500? people) are not elligible to get the funding.
Beauty. The interesting part is that companies who get the funding get rights to any patents as a result of the research, with the proviso that the Gov. gets to use the technology royalty free.
And as Lessig has pointed out, patents were also originally intended to help spur innovation, by giving the inventor and incentive to invent.
Also, seems to me like a lot of people end up sitting on patents - they never intend to produce the thing. I'm not sure how much of a wacked out conspiracy theory it is, but I've heard urban legends of the oil companies buying patents/rights to energy saving devices, never intending to make the device. Regardless of the truth, you can see the logic.
Back to the SBIR: So the good part is the Gov. gets rights to the results of the public money. Yea. And the underdog business gets a foot in the door. Yea.
But what if nothing is done with the patent? Or the rights are sold to one of the big boys, for a sort of denial of service attack in the patent world.
I think the patent should be reverted to the public domain (no royalties for anyone) if steps to develop the product/idea have not been taken within 1 year (in addition to the other current patent time restrictions).
Relevancy:
Why not apply the same thing to software? Gov agency's get the code for free, to diminish inter-agency rivalry. And if a private company/person/etc developed the code under a Gov. contract, they get patent style rights to it - a few years of proprietary code, which must eventually be released to the public. And if there is no active development on a commerical product (keep in mind a product aimed at being sold back to the government probably counts), then the time to release into the public domain is even sooner.
Possibly, I don't know the details of the compression technique, but if it requires random, or nearly random data, you won't be able to do significant recursion since the compression technique will more than likely impose order on the output data.
I worked on HETE (the first one) briefly. The team was small but I think the design choices may have affected cost more than team size.
Aeroastro was employing people to do stuff like design S-band radios from scratch. That's expensive. Probably much more expensive than buying some from a radio manufacturer, who does that sort of thing all the time.
This is something I saw a few times on that project, and I've seen lots of times on other projects: In the desire to do things as cheap as possible, you tend to grow/build your own solutions. I think this is especially true in an educational environment, where you have essentially free labor (grad students). However, you often end up spending more money and taking longer by doing so. Big NRE (non recurring engineering) costs should be saved for high volume products - not single production items like satellites.
As with many educationally funded projects, there were also a few design choices that were dictated by politics, rather than just functional requirements. I may have the details wrong, but I seem to remember using a piece of French equipment on the satellite (nothing inherently wrong with French stuff), but none of the engineers at AeroAstro were fluent in French (although some of the MIT personell were), and I don't think the instructions were in English. Complex communications like this can often cost a lot of time (which equals money).
Also, I'm not totally positive, but I'm pretty sure there were some Radio-Shack parts on HETE. I don't know about HETE-2, but since a lot of the flight spares from HETE were used on HETE-2, Its not outside the realm of possibility that some Radio-Shack parts flew on HETE-2. At one point, I do remember there being a lot of political hoopla about not using Radio-Shack parts though.
One other thing: while the architecture of HETE is really neat (four boards with a DSP and two transputers each, and any board could function as the satellite control or be assigned to an experiment), it still seemed overly complicated to me. Of course, not fully understanding the instrument data path needs, I'm ready to admit that I'm wrong.
>Current management at Excite won't change, either.
Maybe, maybe not. I just had an ATT cable modem installed to day.(I love talking to techs) Anyway, the tech mentioned that one of the main reasons Excite is declaring Bankruptcy, is so ATT would buy them out. Apparently, ATT refused to do so w/o Excite going CH11.
My proof of ignorance, in the form of two questions, which were sparked by my recolection of the axiom: If you can't explain it to your mom, then you don't understand it. So I have two questions on the same subject, one geared towards a technical answer, the second on how to explain it best in a non-technical fashion.
1) I'm not overly familiar with the current state of crypto software, but I am familiar with Moore's law, so is the proverbial cat really out of the bag? i.e. how easy is it for terrorist (or human rights activists) to do very simple things, such as change a #define in the code of most crypto packages, especially PGP, and recompile for a much stronger crypto package. So that, with the existing code base (until the advent of useful quantum computers) anyone could easily continue to use good crypto, despite Moore's law.
2) I'm assuming that, to some extent at least, the cat is out of the bag. So, what do you see as the best way to support this argument to the powers that be. Perhaps a better way to phrase the question is to ask if you have tips for writing your congressman, specifically on for why going back to legistlated weak encryption only limits US citizens.
Slashdot Mirror
Even without buying a WiSpy, you can get a good feel for how much background noise exists with the free inssider app from the same guys - and it works with most wifi cards. I've used it at home to help me select which channel(s) to use.
Russ
Also, the guys at work, where we outfit large commercial operations with Wifi also recommend sticking with 802.11g. Most routers can be configured to limit themselves to g. The 802.11n stuff is really fast -when it gets through, but if you're pushing the limits of your link budget, you'll get better overall performance at the slower speeds. (disclosure, I currently work for www.fidelity-comtech.com) Russ
Within the last year, we had a new high end FPGA vendor visit us. They mentioned that some of their first customers were people/companies doing these fast trades. They were willing to spend $50k+ per part to increase the algorythm speed by 10 - I think going from 10's of ms to ~1ms, but I'm not sure the exact magnitudes. In some ways, this cost isn't that much, considering that the product development costs are likely $300k+. I do wonder how widely the algorythms vary across the different players. There are definetely some large players in this space, and industrial espionage too.
I've heard that many residential electric meters have ERT enabled meters, which use a custom wireless protocol, so that the guys don't have to get out of the truck to read your meter. So, you'll need some sort of radio, and then hack the protocol (maybe sombody already has?).
I'm lucky to live in a place with nice roads. Back in Virginia, the choices were slim. I think it is key to live somewhere that makes it possible to exercise outside.
So, I just train like the dickens for events over the year. Things like the Triple Bypass or 24hrs of Moab.
Actually, just about any type of endurance exercise seems to be suit the geek personality pretty well. There is a lot of science and performance tracking that goes on to optimize your performance.
I totally agree. H&H (AoE) is the bible. It is the standard reference everyone uses. But, it is also way overkill if all you want is how to make something work. There are lots of other options out there. My personal choice is "Circuit Design for Electronic Instrumentation" by Darold Wobschall, ISBN:978-0070712300. I like it because most of the time I'm interested in hooking up sensors of various sorts to a micro, for which this book is a perfect fit. It does a great job of explaining the vast majority of the basic stuff you might want to use.
The one caveat is that it is old. So it is not going to have things like spread spectrum (cell phone). Fortunately, most of the basics haven't really changed in over 30 years.
More to the point, the demographic is not for the perfect Pepsi, but for the perfect Pespsis. Malcom Gladwell had a talk at a TED conference, where he explains demographic research that has been done to show that there is usually not a perfect solution, but rather several near optimal solutions. Probably not applicable to all data-mining applications, but certainly appropriate for anything relating to subjective tastes.
Dunno, seems like some earlier research on a similar idea might be more useful - the plenoptic camera. Just one big CCD and a lot of math. CCDs are getting bigger all the time, most people don't really need all the pixels they have already, so sure, why not use some of those pixels for depth. But sticking with one CCD will probably be cheaper. http://graphics.stanford.edu/papers/lfcamera/
Ever notice how its getting increasingly harder to get a console game where you can do multiplayer on the same box? My bet is its driven by the profits from the on-line fees earned by forcing you to play on-line for multiplayer. Noticed how thinings are becomming more MMO oriented? Thank the income streams generated by WOW for that.
So I went through and customized the same model with Windows Vista and with Ubuntu, to the same configuration (better LCD and larger battery, but otherwise stock). The Ubuntu model was roughly $854, and the Windows Model was roughly $824 (I might be off by a few bucks). But why is the Ubuntu model more than the Vista model? grrr.
Its all about the killer app to justify the tablet. I could see the afore mentioned movie player possibly doing the trick. But for me the killer app would be something that lets me draw boxes, circles, triangles, etc, and cleans them up. Think Visio on steroids with OCR.
I think there is another missing key...
I've never researched it, but I would be surprised (read: I'm pretty sure), that the major PC vendors get kickbacks from AOL, Symantec, etc for pre-installing trial software on the PC's. I also wouldn't be surprised to learn that the PC makers don't give the consumer a discount for these "bonus" materials. Personally, I hate it. It usually takes me several hours just to clean off the crap.
But the important part is that Linux and BSD distros do not have any extra, shall we say, margin enhancers. So, I doubt that we see significant efforts on the PC manufacturer's part to do much about moving towards non-Windows offerings (ie: show me the money).
Or as I tend to think of it - programmers get corrected every single time the programming language syntax is not correct, but that rarely happens in email.
There are tools in place to help you - spelling and gramatical checkers. Most of them even work as you type. However, fixing little red and green underlines are not required before you press send.
Also the these tools are not perfect, nor by any means ubiquitous (I don't see any red squiggly underlines in the slashdot comment form). Spell checkers are pretty good, but I still find errors in them, mostly due to words missing from the dictionary. Grammer checkers still suck. They often recommend reorgnizing phrases in ways that don't make sense, or more typically, in ways that ruin the thought process/flow of a paragraph.
This takes a few paragraphs to become relevant, so stay with me...
Recently, I've been involved with submission of an SBIR (Small Business Inovation Research) proposal, in which the Gov. gives small companies money to do research - the idea being to help create innovation - the big boys (companies over 500? people) are not elligible to get the funding.
Beauty. The interesting part is that companies who get the funding get rights to any patents as a result of the research, with the proviso that the Gov. gets to use the technology royalty free.
And as Lessig has pointed out, patents were also originally intended to help spur innovation, by giving the inventor and incentive to invent.
Also, seems to me like a lot of people end up sitting on patents - they never intend to produce the thing. I'm not sure how much of a wacked out conspiracy theory it is, but I've heard urban legends of the oil companies buying patents/rights to energy saving devices, never intending to make the device. Regardless of the truth, you can see the logic.
Back to the SBIR: So the good part is the Gov. gets rights to the results of the public money. Yea. And the underdog business gets a foot in the door. Yea.
But what if nothing is done with the patent? Or the rights are sold to one of the big boys, for a sort of denial of service attack in the patent world.
I think the patent should be reverted to the public domain (no royalties for anyone) if steps to develop the product/idea have not been taken within 1 year (in addition to the other current patent time restrictions).
Relevancy:
Why not apply the same thing to software? Gov agency's get the code for free, to diminish inter-agency rivalry. And if a private company/person/etc developed the code under a Gov. contract, they get patent style rights to it - a few years of proprietary code, which must eventually be released to the public. And if there is no active development on a commerical product (keep in mind a product aimed at being sold back to the government probably counts), then the time to release into the public domain is even sooner.
Possibly, I don't know the details of the compression technique, but if it requires random, or nearly random data, you won't be able to do significant recursion since the compression technique will more than likely impose order on the output data.
Just a thought.
I worked on HETE (the first one) briefly. The team was small but I think the design choices may have affected cost more than team size.
Aeroastro was employing people to do stuff like design S-band radios from scratch. That's expensive. Probably much more expensive than buying some from a radio manufacturer, who does that sort of thing all the time.
This is something I saw a few times on that project, and I've seen lots of times on other projects: In the desire to do things as cheap as possible, you tend to grow/build your own solutions. I think this is especially true in an educational environment, where you have essentially free labor (grad students). However, you often end up spending more money and taking longer by doing so. Big NRE (non recurring engineering) costs should be saved for high volume products - not single production items like satellites.
As with many educationally funded projects, there were also a few design choices that were dictated by politics, rather than just functional requirements. I may have the details wrong, but I seem to remember using a piece of French equipment on the satellite (nothing inherently wrong with French stuff), but none of the engineers at AeroAstro were fluent in French (although some of the MIT personell were), and I don't think the instructions were in English. Complex communications like this can often cost a lot of time (which equals money).
Also, I'm not totally positive, but I'm pretty sure there were some Radio-Shack parts on HETE. I don't know about HETE-2, but since a lot of the flight spares from HETE were used on HETE-2, Its not outside the realm of possibility that some Radio-Shack parts flew on HETE-2. At one point, I do remember there being a lot of political hoopla about not using Radio-Shack parts though.
One other thing: while the architecture of HETE is really neat (four boards with a DSP and two transputers each, and any board could function as the satellite control or be assigned to an experiment), it still seemed overly complicated to me. Of course, not fully understanding the instrument data path needs, I'm ready to admit that I'm wrong.
My proof of ignorance, in the form of two questions, which were sparked by my recolection of the axiom: If you can't explain it to your mom, then you don't understand it. So I have two questions on the same subject, one geared towards a technical answer, the second on how to explain it best in a non-technical fashion.
1) I'm not overly familiar with the current state of crypto software, but I am familiar with Moore's law, so is the proverbial cat really out of the bag? i.e. how easy is it for terrorist (or human rights activists) to do very simple things, such as change a #define in the code of most crypto packages, especially PGP, and recompile for a much stronger crypto package. So that, with the existing code base (until the advent of useful quantum computers) anyone could easily continue to use good crypto, despite Moore's law.
2) I'm assuming that, to some extent at least, the cat is out of the bag. So, what do you see as the best way to support this argument to the powers that be. Perhaps a better way to phrase the question is to ask if you have tips for writing your congressman, specifically on for why going back to legistlated weak encryption only limits US citizens.
Quick search on google (for IRDA serial port)comes up with: Decros IRDA Seems like fundamentally, that should get you there.
Pretty useless for hooking up to your local LAN without a ethernet port or at least a PCMCIA slot.