"if it was able to record two or more (ie, any number of) shows being broadcast simultaneously!"
AFAIK, you can do that with MythTV as long as you put two or more cards in the box (cards such as the PVR250 cards that have an MPEG2 encoder chip on it, so that it barely takes any CPU to compress).
"Of course, the antenna lightning protecteor and the antenna were more expensive than the AP"
Of course, that is Murphy. Actually I'm pretty surprised the AP and everything behind it survived.
I wonder what happened to the antenna to make it die, and let the AP connected to it survive. Must have been heat only then I guess.
Was the antenna a bare metal object, or a plastic covered or otherwise shielded thing?
Using google for "lightning power": "A typical lightning bolt contains 1 Billion Volts and contains between 10,000 to 200,000 amperes of current. And 1 strike could power a 100 watt light bulb for 3 months."
Even if only 0.1% of that remains you still have at least 10amps rushing down into your router/switch... ouch.
Whoah. You didn't read the gemesis article, did you? There are two very recent artificial diamond producers only now ready to begin production. One of them is gemisis (gemisys/gemysis/whatever). Rumors are that the 'debeers' that now control the worldwide diamond supply are pretty worried about those recent development. But ironically it is not a taking over of the 'debeers' diamond markets that either of these companies is aiming for. They both are quoted to say that the current diamond market for them is just a stepping stone to expand their business and technology to produce diamond transistors. Note that silicon melts a lot quicker than diamond and that the main problem of current high-end chips is power dissipation.
Sure, the textile industry application of silk alone is huge, but the integrated transistor market is not something to laught at at all.
It is quite probably that 10 years from now, that statement of you sounds like the 640kb statement of bill gates does now.
Interesting. Do you know that from experience on an actual lightning strike, or is that based on the sales pitch of the equipment?
I think lighting protectors will reduce your chances of being fried by lightning, but the effectiveness is quite probably very dependent on the quality of the installation of the lighting protection system.
A lightning bolt has a hell of a lot of energy in it. Remember you can hear the strike from miles away and they light the sky from an equal distance, that is a hell of a lot of energy, and that is just the radiated part of the energy, there is more of it in the direct path of the bolt. The network cable is an ideal conductor into electronics that is already sensitive to static discharges, so there better be no residual at all. Even some oscillation of the ground loop of the lightning protector resulting from the inherent selfinductance of a non straight wire, in combination with crosstalk between the lightning protector wiring and the network wiring down the chimney may just create enough voltage and current to fry that network behind the AP.
I tried to read the article, but didn't get past the image.
Am I blind or does everybody have problems reading the 3x5 pixels font on the "Switching to Linux? Consider This..." picture. Even after clicking on the 'click for larger view' it hurts my eyes trying to read anything else but the title.
The whole image is friggin 500x252 pixels. What? Is that article written for people who still have their screens at 640x480?
What kind of magazine editor would let that kind of crappy layout through?
Is the content under equally well done quality control?
From the article: "An obvious strategy is to minimize the product of these two quantities."
Mr Brian Hayes is assuming equal cost for a state anywhere in the 'rw' space. That is much too simplistic to be a practical model.
Luckyly he shows that he realizes that later on by saying "Everything hinges on the assumption that rw is a proper measure of hardware complexity". And I'll make my argument below that that indeed is not a proper measure at all. You can agree with me and dismiss that 'base 3 is the most efficient' and skip the rest of this post, or read on and find out why I think that it is bogus to say that the terniary system or base-e is 'the most efficient'.
To begin with, what does he mean when he says 'An obvious strategy is...'? Sure it's a simple strategy, but that doesn't have to mean that it is the correct one for the task at hand. For a deer it's an obvious strategy to look at the headlights of the approaching car....
Actually, I have been told and have found in my own experience that if an article says that something is 'obvious' without further explanation, usually that is exactly the main flaw of the 'proof' that is presented (but to be honest, when people start to talk about rumors, or their 'experience', that usually means something similar, so please ignore this paragraph;-))
After stating what he believes is obvious, he goes on on that premise without explaining why minimizing 'rw' equals to finding the highest 'efficiency'...
I my dictionary, efficiency in this context is the cost to use the number representation in computational tasks. Then I go further assuming that the cost is mainly a function of power usage, implementation size, and circuit speed. That is what I mean in the rest of this posting when I say 'efficiency' (yes even my model is too simplistic for practical use, because I don't specify the exact relationship of power, size, and speed in the 'efficiency' equation).
In practical implementations the cost per state is never the same for all symbol sizes. Actually, each state may have its own cost due to things like current leakage (actually: not just the number of total states to represent a value, but also the number and type of state transitions need to be analyzed to determine the efficiency).
Efficiency of a number representation system is very much linked to implementation. In today's silicon with the transistors we have today, base 2 is the most efficient. 'Why?' Well, that is because adding more states to the value of a signal would make the circuits that switch them sufficiently larger, slower, and more power hungry that the end result would be a less efficient circuit.
Some materials researchers have been talking about base 4 logic, because they think they may be able to use the electon spin as well as the current so carry the information in an electrical signal. In semiconductors that could very well double the information carrying capacity of switching signals without increasing the power usage or transistor size in semiconductors. Hence, more switched information per joule, hence better efficiency. For such a chip production process, that would result in base 4 being the most efficient, still not base 3.
Besides all that, for performing computations it's not just the value storage in symbol states that is important for the efficiency, but the symbol state transition is often an even more important parameter for efficiency in computations.
When counting in a base 2 system, we have less than 2 symbol state transitions per step (the lsb changes every step, the next bit every other step, the next bit once every 4 steps, etc). Similarly, in base 3 we have less than 1 2/3 symbol state transitions per step, and in base 4 we have less than 1.5 symbol state transitions per step). Now we still need to know how much does symbol state transition costs in base 2, 3, or 4. Without that we don't know the efficiency. We can guess and say that a symbol state transition costs 1 units when
"Your 1-1000 example was kind of crap...try it again with 1,000,000,000,000 and you get 60 states for base 2 vs 39 for base 3, a pretty good savings."
Ok, I'll try it again with "1,000,000,000,000". In scientific notation, that number is 1e12 (for easier calculations).
And you think that 1e12 is 39 states for base 3?.
3^(39/3)=1594323. 1.6e6... That's hardly 1e12.
Actually, 3^26=2.54e12, hence "1,000,000,000,000" is 26*3=78 states... That is more than the 60 states for base 2.
Math, hmm?
The rest of your reasoning would equally prove that base 4 is better than base 3.
And a 'bit' implies base 2, so you can's talk about 3-value symbols and call it a 'bit'. So assuming that by 'n bit ternary' you mean 'n symbol radix 3' -> Sure, a higher radix for the symbols results in less symbols (smaller words) for the same values, but that is not less storage, because the symbols are more complicated, storing more information per symbol. Shannon's Information theorem still holds, it's just that one radix 3 symbol can store more than one bit of information.
Umm... But if you want to store numbers 0..1000, you need 32 states (2+10+10+10) for decimal systen, 20 states (10*2) for binary, and 21 for ternary (7*3).
So I say that that reasoning does not prove that base 3 is 'more efficient'.
Btw what does 'more efficient' mean in this context: less power usage? lower cost? more girls?...
Too bad, because I would have loved to see NASA extend its life by launching a tiny and cheap 'repeater' satellite to pickup and relay its distant signal...
Actually, that finding of 'No 42' blend of Earl Grey tea is no surprise at all. The book does say that Earth and everything on it (except for the descendants of "Ship A"), is part of the computation of the Question, and the Question should be output by the system just about know, or at least not too long after the house of a guy named Arthur gets demolished somewhere in the U.K. So, the fact that '42' shows up on a can of tea of a well know store, when a book that describes the answer to the Question mentions the number 42 in relation to the Question and also talks a lot about tea, is not very surprising and probably not even very improbable.
Similar to the history of the word 'cricket'. It's embedded.
Now I must go out and compensate for this posting by being normal...
According to the mass media, it's perfectly normal for Americans not to know either:
Actually, I couldn't find it back with google (imagine!), so I can't 'prove' it with a link, but I think that it was in Season 12 of The Simpsons:
The family goes to a restaurant with a 'fancy waiter' who guesses the drinks that the guests want to drink. Homer gets, of course, a beer after the waiter obviously glances at Homers beerbelly, and for Marge the waiter literally reaches up his sleeve to pull out a glass and says "And for the lady, a Long Island Ice Tea", and puts the really tall glass in front of marge. In response to which Marge giggles and says "I think Large Island Tea would be a better name". The rest of the family does not laugh but instead tries to ignore her obviously missing prior knowledge of either the drink, or the well known island that extends northeast from the Big Apple with the obvious shape...
I can't believe that I just remembered all that. I hope somebody will find a flaw in how I remember it and post an actual transscript of the scene, including what Bart and Lisa got and where Maggie was during that scene, just to oust me from this throne of feeling alone in remembering this worthless detail (or otherwise maybe somebody who can make an argument about how the drink really has nothing to do with the island).
Oh, the tricks that the mind plays on you in keeping memory of such unimportant details of not even the funniest bits of TV shows once watched in the past.
Now where did I put my coffee, and what was your name again?
Slashdot Mirror
"if it was able to record two or more (ie, any number of) shows being broadcast simultaneously!"
AFAIK, you can do that with MythTV as long as you put two or more cards in the box (cards such as the PVR250 cards that have an MPEG2 encoder chip on it, so that it barely takes any CPU to compress).
"3 one hours shows fit onto a 256M CF card."
And the battery lasts long enough to view those three shows, even though all the decoding takes all that CPU? That would be pretty amazing.
Umm... HDTV and then 3D...
"Of course, the antenna lightning protecteor and the antenna were more expensive than the AP"
Of course, that is Murphy. Actually I'm pretty surprised the AP and everything behind it survived.
I wonder what happened to the antenna to make it die, and let the AP connected to it survive. Must have been heat only then I guess.
Was the antenna a bare metal object, or a plastic covered or otherwise shielded thing?
Using google for "lightning power": "A typical lightning bolt contains 1 Billion Volts and contains between 10,000 to 200,000 amperes of current. And 1 strike could power a 100 watt light bulb for 3 months."
Even if only 0.1% of that remains you still have at least 10amps rushing down into your router/switch... ouch.
The less common 50,000 amp lightning strike (90% are at or under 8Kamps) which is typically associated with 1000 joules. That is about 36 megawatts for 28 useconds.
Only 1000 joules. Hmm. Is there maybe a capacitance effect that actually makes a lightning protector work very reliably when installed correctly?
Or this posting from a guy who worked as a fire fighter.
Too many Questions Error. Note to self: read up on lighting protection.
"(diamonds do as well, but not as many)"
Whoah. You didn't read the gemesis article, did you? There are two very recent artificial diamond producers only now ready to begin production. One of them is gemisis (gemisys/gemysis/whatever). Rumors are that the 'debeers' that now control the worldwide diamond supply are pretty worried about those recent development. But ironically it is not a taking over of the 'debeers' diamond markets that either of these companies is aiming for. They both are quoted to say that the current diamond market for them is just a stepping stone to expand their business and technology to produce diamond transistors. Note that silicon melts a lot quicker than diamond and that the main problem of current high-end chips is power dissipation.
Sure, the textile industry application of silk alone is huge, but the integrated transistor market is not something to laught at at all.
It is quite probably that 10 years from now, that statement of you sounds like the 640kb statement of bill gates does now.
Interesting. Do you know that from experience on an actual lightning strike, or is that based on the sales pitch of the equipment?
I think lighting protectors will reduce your chances of being fried by lightning, but the effectiveness is quite probably very dependent on the quality of the installation of the lighting protection system.
A lightning bolt has a hell of a lot of energy in it. Remember you can hear the strike from miles away and they light the sky from an equal distance, that is a hell of a lot of energy, and that is just the radiated part of the energy, there is more of it in the direct path of the bolt. The network cable is an ideal conductor into electronics that is already sensitive to static discharges, so there better be no residual at all. Even some oscillation of the ground loop of the lightning protector resulting from the inherent selfinductance of a non straight wire, in combination with crosstalk between the lightning protector wiring and the network wiring down the chimney may just create enough voltage and current to fry that network behind the AP.
I tried to read the article, but didn't get past the image.
Am I blind or does everybody have problems reading the 3x5 pixels font on the "Switching to Linux? Consider This..." picture. Even after clicking on the 'click for larger view' it hurts my eyes trying to read anything else but the title.
The whole image is friggin 500x252 pixels. What? Is that article written for people who still have their screens at 640x480?
What kind of magazine editor would let that kind of crappy layout through?
Is the content under equally well done quality control?
Geesh man.
"a cable goes into the fireplace up to the roof, where it is attached to an antenna."
Am I the only one who after reading that immediately thinks "what about lightning"?
From the article: "An obvious strategy is to minimize the product of these two quantities."
;-))
Mr Brian Hayes is assuming equal cost for a state anywhere in the 'rw' space. That is much too simplistic to be a practical model.
Luckyly he shows that he realizes that later on by saying "Everything hinges on the assumption that rw is a proper measure of hardware complexity". And I'll make my argument below that that indeed is not a proper measure at all. You can agree with me and dismiss that 'base 3 is the most efficient' and skip the rest of this post, or read on and find out why I think that it is bogus to say that the terniary system or base-e is 'the most efficient'.
To begin with, what does he mean when he says 'An obvious strategy is...'? Sure it's a simple strategy, but that doesn't have to mean that it is the correct one for the task at hand. For a deer it's an obvious strategy to look at the headlights of the approaching car....
Actually, I have been told and have found in my own experience that if an article says that something is 'obvious' without further explanation, usually that is exactly the main flaw of the 'proof' that is presented (but to be honest, when people start to talk about rumors, or their 'experience', that usually means something similar, so please ignore this paragraph
After stating what he believes is obvious, he goes on on that premise without explaining why minimizing 'rw' equals to finding the highest 'efficiency'...
I my dictionary, efficiency in this context is the cost to use the number representation in computational tasks. Then I go further assuming that the cost is mainly a function of power usage, implementation size, and circuit speed. That is what I mean in the rest of this posting when I say 'efficiency' (yes even my model is too simplistic for practical use, because I don't specify the exact relationship of power, size, and speed in the 'efficiency' equation).
In practical implementations the cost per state is never the same for all symbol sizes. Actually, each state may have its own cost due to things like current leakage (actually: not just the number of total states to represent a value, but also the number and type of state transitions need to be analyzed to determine the efficiency).
Efficiency of a number representation system is very much linked to implementation. In today's silicon with the transistors we have today, base 2 is the most efficient. 'Why?' Well, that is because adding more states to the value of a signal would make the circuits that switch them sufficiently larger, slower, and more power hungry that the end result would be a less efficient circuit.
Some materials researchers have been talking about base 4 logic, because they think they may be able to use the electon spin as well as the current so carry the information in an electrical signal. In semiconductors that could very well double the information carrying capacity of switching signals without increasing the power usage or transistor size in semiconductors. Hence, more switched information per joule, hence better efficiency. For such a chip production process, that would result in base 4 being the most efficient, still not base 3.
Besides all that, for performing computations it's not just the value storage in symbol states that is important for the efficiency, but the symbol state transition is often an even more important parameter for efficiency in computations.
When counting in a base 2 system, we have less than 2 symbol state transitions per step (the lsb changes every step, the next bit every other step, the next bit once every 4 steps, etc). Similarly, in base 3 we have less than 1 2/3 symbol state transitions per step, and in base 4 we have less than 1.5 symbol state transitions per step). Now we still need to know how much does symbol state transition costs in base 2, 3, or 4. Without that we don't know the efficiency. We can guess and say that a symbol state transition costs 1 units when
That was my point to begin with. What are you doing stealing my point? ;-)
"Your 1-1000 example was kind of crap...try it again with 1,000,000,000,000 and you get 60 states for base 2 vs 39 for base 3, a pretty good savings."
Ok, I'll try it again with "1,000,000,000,000". In scientific notation, that number is 1e12 (for easier calculations).
And you think that 1e12 is 39 states for base 3?.
3^(39/3)=1594323. 1.6e6... That's hardly 1e12.
Actually, 3^26=2.54e12, hence "1,000,000,000,000" is 26*3=78 states... That is more than the 60 states for base 2.
Math, hmm?
The rest of your reasoning would equally prove that base 4 is better than base 3.
And a 'bit' implies base 2, so you can's talk about 3-value symbols and call it a 'bit'. So assuming that by 'n bit ternary' you mean 'n symbol radix 3' -> Sure, a higher radix for the symbols results in less symbols (smaller words) for the same values, but that is not less storage, because the symbols are more complicated, storing more information per symbol. Shannon's Information theorem still holds, it's just that one radix 3 symbol can store more than one bit of information.
Umm... But if you want to store numbers 0..1000, you need 32 states (2+10+10+10) for decimal systen, 20 states (10*2) for binary, and 21 for ternary (7*3).
...
So I say that that reasoning does not prove that base 3 is 'more efficient'.
Btw what does 'more efficient' mean in this context: less power usage? lower cost? more girls?
Hmm. shouldn't they be making an X11 version of tvision then?
Ok, I'm open for suggestions.
What else causes redshift?
Too bad, because I would have loved to see NASA extend its life by launching a tiny and cheap 'repeater' satellite to pickup and relay its distant signal...
Well, there are many websites that vigorously claim that the email tax is a number one hoax....
So it must be true then.
Right?
RIGHT!
Uh Oh.
That was this summer 2003, Sen. Mark Dayton's idea to fight Spam...
Pfew, that was a close call: Senator Downplays E-Mail Tax Idea, Thursday, May 22, 2003.
If they tax email, then the spammers have won.
Now, if 'they' find this posting, they'll probably come up with a tax on hyperlinks...
But will anybody think of the children?
Actually, that finding of 'No 42' blend of Earl Grey tea is no surprise at all. The book does say that Earth and everything on it (except for the descendants of "Ship A"), is part of the computation of the Question, and the Question should be output by the system just about know, or at least not too long after the house of a guy named Arthur gets demolished somewhere in the U.K. So, the fact that '42' shows up on a can of tea of a well know store, when a book that describes the answer to the Question mentions the number 42 in relation to the Question and also talks a lot about tea, is not very surprising and probably not even very improbable.
Similar to the history of the word 'cricket'. It's embedded.
Now I must go out and compensate for this posting by being normal...
According to the mass media, it's perfectly normal for Americans not to know either:
Actually, I couldn't find it back with google (imagine!), so I can't 'prove' it with a link, but I think that it was in Season 12 of The Simpsons:
The family goes to a restaurant with a 'fancy waiter' who guesses the drinks that the guests want to drink. Homer gets, of course, a beer after the waiter obviously glances at Homers beerbelly, and for Marge the waiter literally reaches up his sleeve to pull out a glass and says "And for the lady, a Long Island Ice Tea", and puts the really tall glass in front of marge. In response to which Marge giggles and says "I think Large Island Tea would be a better name". The rest of the family does not laugh but instead tries to ignore her obviously missing prior knowledge of either the drink, or the well known island that extends northeast from the Big Apple with the obvious shape...
I can't believe that I just remembered all that. I hope somebody will find a flaw in how I remember it and post an actual transscript of the scene, including what Bart and Lisa got and where Maggie was during that scene, just to oust me from this throne of feeling alone in remembering this worthless detail (or otherwise maybe somebody who can make an argument about how the drink really has nothing to do with the island).
Oh, the tricks that the mind plays on you in keeping memory of such unimportant details of not even the funniest bits of TV shows once watched in the past.
Now where did I put my coffee, and what was your name again?
So by now the story is a couple of hours old.
Can anybody please confirm that this guy did the right (and safe) thing and is actually still alive?
I know, I know. And that is why I make the point that the 'new name' of 'freedom' is really not as 'unfrench' as it seems...
Umm, freedom (liberty) and France is not as far fetched as it seems recently.
...
liberte, egalite, fraternite,
So much to learn, isn't there? Such as a well known factoid about the Statue of Liberty. They didn't ask for that back, did they?
AFAIK, lvm snapshotting on Linux only stores the differences.
For $300k/year I will support it.
Well, the $350k/year he is talkin about can hire quite some 24/7 on-site support. For Debian too. Just put some good people on the payroll.
What? What is next, a robot that can eat ice cream in my place, and a robot that can have sex in my place?
Geesh. Did they have any useful 'inventions' at all?