I don't think it is even pinging your phone - it is pinging *you*. As a conductive object, you reflect RF - including the RF generated by mobile phone masts. As you move, you change the pattern of reflections. The pervasive mobile phone masts create a kind of universal radar transmitter receiver, so the only thing that the snooper needs to carry to spy on you is a receiver.
OTOH, all they will see is that a person is moving hither, thither and yon. They woundt see what you are doing or hear what you are saying.
So, from the Civil Liberties point of view, this is no worse (but no better than) universal CCTV surveillance. There will be nowhere you can go - above ground, out of doors - that they can't watch you. I am skeptical about the "through walls" bit - through some walls, some of the time, but my mobile often loses signal indoors - and if I don't get enough signal to recieve, I am surely not reflecting much.
The signal is unlikely to be detailed enough to identify you, so all that they can tell is that a human is moving. This could be useful in two ways. As the article says, monitoring "no humans allowed" areas like security barriers round military and nuclear installations. And tracking someone once they have been identifier - e.g. tracking the kidnappers as they run off with the ransom money. But there would be a *lot* of ways of shaking such a tail an an urban area - if you knew it was happening.
I use Opera gestures - and love them. But they don't make sense for all applications. The problem is a dsicontinuity when you switch between keyboard and mouse - either way. Editing, and most programming operations, is fundamentally a keyboard based operation, and hot keys are far more sensible than mouse gestures for this. But for me, browsing the web is a mouse-based operation. I have to point to links to follow them, so my hand is on the mouse. I have a wheelmouse, so scrolling is also under my fingers. The only gestures I use regularly are back and forward, and they have become so automatic I use them (uselessly0 wherever the model applies - i.e. in all "browser" type applications, such as Konqueror or Windows Explorer.
One thing we want to do is to try and get people to standardise. It will be a *real* pain if one piece of software used a gesture for minimise and another for quit.
I tried that - it didn't work. He really was PROM - write once. It wan't a question of looking good/bad, saving money or face.: his world model was formed and he wasn't changing it. Like arguing with a Jehovah's Witness - logic bounces off cast iron certainty.
CDMA only causes a lot of slowdown when you have a lot of collisions - i.e. when you have a lot of small packets fighting for the bandwidth. Which is another of the consequences of TCP/IP - flow control packets etc., and obviously occurs on a "workstation" environ,ment with many independant systems talking to servers. In a SAN type network using direct Ethernet packets, you would notrmally send bulk data in large packets which would swamp the bus for a while, then back off. But, like a scsi bus, you would tend to get one logical transaction loading the net for a while, then backing off while it seeks to the next logical block. And SAN type applications tend to have a few disk users, rather than hundreds of light users.
Gbit Ethernet can sustain more than 500Mbit/sec, but usually doesn't get the chance to do so. The problem is running TCP/IP protocol over it. TCP/IP rins, over dodg links, the breadth of the world. In order to achieve this, it has things like internal checksums, time-to-live etc which are required to achieve this - but require significant intervention at the driver level to manage. And then you probably have timeout and retry systems to manage as well.
Firewire, Fibre Channel, Scsi etc. don't have this flexibility. They assume that the device you are talking to is is on this bus/ring and that messages will, give or take parity errors (which can be reported) get through reliably. Therefore they can have much ligher protocols. You can usee what, in Ethernet terms, is a MAC address not an IP address.
The firewire suggestion at the head is perfectly snesible. But if you were to go into the ethernet driver at the packet level, not at the OP level, I reckon you could get 90% of theoretical bandwidth out of an ethernet-based connection. We currently get over 500Mbit/sec using UDP connections. Drop a layer of protocol and it could be even faster.
I would agree with this in most cases. Management have (or should have) more info. They should, as Jon says, explain their thinking, and give your views a good hearing. If this happens, follow Jon's advice.
But it can happen that you are working for an idiot. It happened to me once in my career. In that case there is only one thing to do. GET OUT. If your manager is an incurable idiot, you will *never* be happy working there, and staying will only make you miserable. But give him/her plenty of rope to hang him/herself; a real idiot will soon show up unambiguously. If the idiocy is not tranparently obvious, then maybe it isn't idiocy.
(My idiot manager had a PROM brain. On day two, after looking over the job I was to do, I gave him a sketch of how I intended to attack the problem. On day about five, I found out that wasn't the right way - but it was too late; his proms were burned. For the next six months I had to shoehorn every weekly progress report into the headings from that initial handwaving guess at a solution.)
This is a classic complaint that things aren't as good as they were in the Golden Age - which almost always turns out to be the complainer's college days, or possibly his first job, when he was young and burning with enthusiasm. We have always built new tools on top of the tools our predecessors have built, and abandoned the tools they used in favour of the better ones - which, maybe, they built for us.
George Stevenson, the railway builder, judged the underlying quality of an engineer by his ability to file things out of raw metal, at a push. "Real Engineers", in his opinion, could file a threaded bolt our of bar stock if they really had to - not that they had to, because mass produced bolts were available, but a "real engineer" kept in touch with basics. How many of us would pass that test?
So what if some of the tools we use are software? While marketing may label the whizzy program "Engineer in a Box", it won't be - it'll be a new set of tools, better than the last but no panacea (anybody remeber "The Last One"). And many engineers will use those tools in stolid, unimaginative, ways, just as they do now. Don't tell me that every engineer in you plant ia imaginative and original. Most engineers, most of the time, are pulling on a few design patterns and a few tools to do essentially predictable jobs. Some of them, some of the time, do get original. And the more powerful the tools they have, the more creative they can get when the lightning of inspiration strikes.
This sort of thing has been said so many times that there is a Monty Python sketch - which I won't quote. In the early days of COBOL, it was supposed to de-skill programming. The manager would write down the business logic, a typist-class drudge would translate it into COBOL, and the compiler would do the rest. Assembler programmers complained that their craft was being destroyed before it was ever born. Is that the world you are living in?
Except you couldn't use it raw for backups. You would have to mount it in something reasonably shockproof, with insertable power and data connectors. This would add significantly to the cost.
Also, current OSs would probably not want you to put it onto your "main" IDE busses = they woudl assume that "fixed" drives on those busses are system drives and mount them only at startup. So you want it on another interface, with a card to connect to that interface and cabling to get out from the case to the removable unit, wherever you have put it.
Having said that, you can probably get to $100 fixed cost, $400 per exchangeable unit - which is not that much premium for very high speed and true random access.
To seque a little, how should (or can ) one decide objectively/mathimatically between short-term and long-term benefits?
The economists answer to this is to deflate future benefits by whatever the "zero risk" interest rate may be. Thus long term returns have to pay more, when they do pay, than the same amount of money left in a bank account.
The trouble with this is that it says that, for example, it is not economically worth saving the whales. According the this theory, we should simply slaughter the wales now, then invest the returns in something "useful".
Now, I cannot prove it mathematically, but to me this is wrong. But that is an emotional response.
So my real response is that you cannot objectively decide between short and long tem benefits. By all means do the economic calculations; look at what you are forgoeing on one plan for.benefits in the other. But allow non-financial factors to affect your decision. As well as, not instead of, financial considerations.
One of the things I like about the whole open source movement is that someone like Larry, who honestly (and probably correctly) says "I haven't an executive bone in my body" can get some of the kudos he undoubtedly deserves.
Whatever the theological arguments about perl vs python vs Java vs...., Larry has created a tool that literally thousands of people use - and value - every day. Respect, man.
I workd for a companyy in hte UK which, you may have gathered, is pretty humid. We have anti-static wriststraps etc., but they get pretty well ignored, and we don't lose much kit to inexplicable (i.e. might be ESD) failures. But sometimes we take kit to exhibitions in Las Vegas which, as you might also have gathered, is pretty non-humid. And if we don't do it just right there, equipment dies by the rackful (particulalry before they close the big doors and turn on the aircon). And for one product I worked on, one customer in Phoenix, AZ, has almost as many failures as the whole of the rest of the world untils somebody clicked and installed the full gamut of ESD precautions.
So, it depends where you are.
But, as been sais, ESD precaustionsa re cheap, and kit is expensive. And, I have beeen told by our in-house ESD/RFI expert, a lot of ESD failures don't show up in instant faulures but in shortened lives - so a board will die after six months rather than five years. Doesn't sound to me like a good risk to take.
I knew a couple in a very similar situation - they both had good thech/manegerial jobs, but she was well up the career ladder in a very big corporation, while he was probably more liike most slashdotters - a while here, a while there, sometimes team leader sometimes not. So he went in for training for his tech discipline. It is much easier to get in and out of, because you are not involved in projects with timescales of months and a need to build a working team, but week-at-a-time courses. Seemed to work for them.
My brothere works in marketing, particularly consumer goods. He says that for candy bars - a field in which he has worked - if they stop advertising, sales die off exponentially with a half-life of two years. Where products are nearly interchangeable (there were probably 10 candy bars you might have bought) continuous flow of advertising is necessary to keep a particular one in the forefront.
Same is true of, say, washing powders. For any particular position in the market, there are usually at least three essentially interchangeable products. Advertising continuously ensures that shoppers pick *yours* rather than its identical competitors.
Digital images have a very specific size: 1024x768x32 = 3 MegaBytes
This is simply not true. HDTV is already running up to 1920 by 1500. For print work, very high quality magazines will scan bu to 10H by 8K. And 32 bits/pixel is not true either. Moset pictures don't need an alpha plane, so if you are doing 8 bit RGB you need 24 bits/pixel. Even if you want 210 bit video (and you are not going to get 10 good bits out of a CCD at that speed) it is only 30 bits - and the difference between 30 and 32 matters at these rates. Actually, for broascast purposes, YUV encodung with half bandwidth chrominance is used, reducing it to 16 bits/pixel at 8 bit sampling. And this example is probably monochrome anyway.
You are right that film has higher resolution than common-or-garden CCDs. Consumer 35mm is probably abput 4k*3k, with professional plate cameras, as used by some fachion photographers, being much better.
No point in going to a cluster - beowulf or otherwise. You just want to stream straight to disk. If, say, Infiniband takes off, you could stream striaght out of the camera to lots of disks. Don't shove the data into a CPU until you need to process it.
You may be right about which grandfather it was; the point is that the idea was far from new by the time Charles Darwin voyaged on the Beagle.
If he had not written about Natural Selection, you are correct that Wallace would have done so. However, while Wallace undoubtedly had the right idea (as Darwin acknowledged), he had not piled up the enormous pile of data, arguments, and counter arguments that Darwin had. If he had published, the idea would probably have taken twenty or thirty years to "catch". Darwin's book was so convincing and with such a mass of data that large numbers of scientists were immediately convinced ("How stupid of me not to think of that", as one of them remarked). This, of course, increased the heat of the argument of those who found the idea incompatible with their faith.
Arging "what-if", it might even have been better in the long term if Wallace had published, because the idea could have crept in insidiously instead of appearing as a massive frontal attack on (some people's) religion.
Darwin certainly did not claim to have discovered evolution. The evidence for evolution of some sort was accepted by a large number (though far from all) scientists and interested people for some time before Darwin - amongst them, Darwin's Grandfather, Josiah Wedgewood, so the idea was far from new.
What Darwin did was find an explanation for evolution - a mechanism by which it occurred. Undoubtedly Lyell believed in, and pointed out to Darwin, the operation of evolution. And the ornithologist certainly pointed that all the finches he had brought back (and carelessly jumbled up) appeared to be descended from a singel ancestor. His book is titled "On The Origin of Species by Means of Natural Selection, or The Preservation of Favoured Races in the Struggle for Life", and it is the "Means of natural Selection" bit that is original.
To assert that Darwin claimed to have discovered Evolution is like claiming that Columbus discovered the Atlantic. Columbnus dicovered how to cross the atlantic, and Darwin discovered hopw to explain Evolution.
I don't think you should see it as typing, more as steering a course through "word space". Like driving, your hands are handling "here" while your eyes are yards ahead checking for problems and planning the course.
Dasher is inherently 2-D with very simple graphics. How about using a Doom-style rendering engine to develop a high-quality 3-D version of Dasher? Use Doom/Quake style firing - pick off the next three letters you need, or finish off the word with a burst of automatic fire. Burn down an entire cliche in one shot. Install objects to represent frequently used phrases and type your manager's name by blasting him off the map.
(Not that I want to give you a Socially Responsible reason for going out and buying the latest, greatest graphics card, you understand).
Noone seems to want to answer the second part of the post...
For programming, it would have to be integrated with your IDE in an editor which is both syntax and semanitcally sensitive. Dasher uses some sort of a dictionary - usually in English - to predict what the next letter is most likely to be. This is obviously not a traditional dictionary, because it manages to predict across words.
For coding use, the dictionary, instead of being static, should be dynamic. Instead of having all the words in the English language, it should have only words (and symbols) which are semantically valid at the point the cursor is positioned in the edit buffer. Furthermore, it should weight them. Local variables are very likely, method variables are less likely, strange packages even less likely. In fact, it might subcategorise these into pseudo-letters, so that when writing (e.g.) Java, packages would appear as a single pseudo-letter. It could also add localisation-type information - entities referred to within a few lines are more likely to be accessed again soon. This doesn't change the sorting order used by Dasher, which is always alphabetic, but it does change "visual space" allocated to each letter in the search area.
You probably want an auto-beautifier i.e. new lines and indents/outdents are added automatically as needed.
This has some interesting side-effects. For example, you can only enter semantically valid programs. It has implicit auto-completion - once the following letters are unambigous, the remaining characters occupy the whole namespace.
The brackets case is also nice. The only close bracket that is ever possible is the close of the most recently opened, and that often has a very high probablity. Which means that inserting an open bracket implicity "arms" the system with the matching close bracket. The same applies for closing strings - the close string charager is always high probability.
For case significant languages there are some interesting effects. Obviously, in principle both upper and lower case must be present, though it may well be that a small minority of letters are accessible at any given instant. At a guesss, it would be better to split into Upper case and Lower case rather than interleave upper and lower - but that is something to experiment with.
Punctuation-type characters show minor problems - we all know alphabetic order, but does ";" come before or after "+"? I don't know. But I expect we could learn this - there aren't that many symbols.
One could also add a special pseudo-section for language-defined keywords, so you just chose a single prefix zone and then go straight to the set of all known keywords. Usuallly no more than 50 or so, and not usually all semantically valid, so you might get quite quick access to them.
Of course, there is a tradeoff with all these special zones. One of the points about Dasher is that you don't need any more specailised knowledge than having learned the alphabet to operate it. Adding language-sensitive zones and so on adds extra operator learning time. But since you have to learn the language anyway, I don't think it is that much of a burden.
I would expect this sort of strategy to (at least) double the input speed for Dasher for a particular programming language. Forget using your eyes - for the able at least - but it might make mouse-driven program-writing a lot faster. In fact, it might overtake typing for the special case of program input in a "known" language. Though I think the most valuable feature would actually be the inability to input a semantically erroneous program. Which means you "only" have to worry about logic bugs and not typing bugs.
A good place to try this would be to create a jEdit plugin. JEdit already has plugin Java browsers and beautifiers, so a lot of the code ought to be there already. A Java-style Dasher window would be a very interesting project. If anybody feels upt to doing this, I would like to help (I don't think I have got the time to lead such a project).
I checked through some of the links to try to find out why some of these sites don't want to be linked to. On thing that came out is that there is a mass of confused thinking and motivation out there. So don't expect a clean solution to this problem. A solution which will satisfy one set of paranoid suits will not satisfy the others.
One of the reasons is that they fear that the appearans of a link from you to them implies some sort of reciprocal approval i.e. that they know of and might be assumed to approve of you. Now, to anyone here, this is absolutely dumb, but corporate zecks and AOLers might not know better.
So here is an idea of how to deal with them. When they post court papers (which are surely public documents), post a reciprocal set of papers requiring them to remove your name, addresss, URL etc. from their papers because they imply they you endorse them etc. Use wording as close as possible to theirs and petition that your case be heard first.
One of two things happen: either the court is sensible and throws out your petition as riduculous, in which case you return with that rejection as a precedent, set in the same court, to justify your linking. Or the court grants your loony case, in which case (by the court's own loguic) they have to withdraw their case against you.
This project seems to me to be attaching what is the root of the problem with all renewables: concentration. When you come down to it, all renewable energy sources are solar power, though many are at second hand. Solar energy drives the weather for hydro, wind and wave, and grows the plants for biomass.
The problem with solar power is that, while there is masses of it, is is relatively diffuse, drops a lot on cloudy days, and comes and goes on a 24 hour cycle. All effective renewable power supplies depend upon some form of a concentrator to make them economically viable - at least for bulk domestic power, rather than specialised uses which can afford the premium cost of solar cells. They also need a load balancer for the 24 hour cycle.
Hydro achieves this by using the landscape - focussing the rainfall into a vally, a lake and a dam. Wind power uses an area (dimension ^2) of land/sea to generate wind captured by a line (dimension ^1) of windmills. These all concentrate solar power - but in a very lossy fashion. Hydro, for ecample, wasts much of the potential energy available as the rain hits the mountaintops "focussing" it into lakes at least half way down the mountain. Wave powere rewuires wint to drive the waves, then the waves to rool ove hunderds of miles to accumulate, before hitting the generator.
So the interesing part of this story is not really the tower, but the acres of greenhoses below and around it. This is a superb enegy concentrator device - and one, probably, with a lot of latent heat in the ground, so would run well after sunset. It is almost direct solar power - sun->hot air->power. If they can generate a lot of hot air, the tower is one way of converting it to electricity, but there may be others. But the point is that greenhouses can be economically constructed out of plastic film at a cost, I guess, 1% of the same area of solar cells. If they are only 10% as efficient at trapping the energy (and solar cells only run 5-15% efficeint), they are still winning by a large margin.
This idea is not actually very new. There was an SF story way back which proposed this sort of thing. Rather than build a tower, you just build a double walled cylinder out of the same polythene and inflate it with the hot air, (of which you have plenty) so that it lifts itself off. If it punctures, it will collapse very slowly (if at all, with many cells) and do little damage (because it willfall onto the greenhouses). Since the the tube is tranparent, a ring of lights round the base shining up into it will cause the whole thing to glow so that any pilot who flies into it must be truly blind.
There might be ecological consequneces - which need not be for the worse. The original SF story had the tubes puncturing the temperature inversion which causes AL's amoga and bringing a cooling draft to LA in the height of summer.
Excuse me - please can you explain why you considered my post a troll? It was purely a technical response to a point made someone else.
The only thing I can think is that you are taking it as approval of Star Wars, or Star Wars 2. Actually, I do not - I think it a damned bad idea; basically the military-industrical complex is taking the US taxpayer as suckers with a bottomless wallet. Buit that doesn't stop me making reasoned technical comments about the proposed system.
If we are to discuss such things, we should do so without knee-jerk reactions.
Slashdot Mirror
I love way the text has been translated - classic Japlish with a charmin folksy style.
Like slashdot.org, huh?
For all except megacorps, the www is redundant. Though most people used it instead of http://. So if it starts www, of course it is http, dummy.
I don't think it is even pinging your phone - it is pinging *you*. As a conductive object, you reflect RF - including the RF generated by mobile phone masts. As you move, you change the pattern of reflections. The pervasive mobile phone masts create a kind of universal radar transmitter receiver, so the only thing that the snooper needs to carry to spy on you is a receiver.
OTOH, all they will see is that a person is moving hither, thither and yon. They woundt see what you are doing or hear what you are saying.
So, from the Civil Liberties point of view, this is no worse (but no better than) universal CCTV surveillance. There will be nowhere you can go - above ground, out of doors - that they can't watch you. I am skeptical about the "through walls" bit - through some walls, some of the time, but my mobile often loses signal indoors - and if I don't get enough signal to recieve, I am surely not reflecting much.
The signal is unlikely to be detailed enough to identify you, so all that they can tell is that a human is moving. This could be useful in two ways. As the article says, monitoring "no humans allowed" areas like security barriers round military and nuclear installations. And tracking someone once they have been identifier - e.g. tracking the kidnappers as they run off with the ransom money. But there would be a *lot* of ways of shaking such a tail an an urban area - if you knew it was happening.
I use Opera gestures - and love them. But they don't make sense for all applications. The problem is a dsicontinuity when you switch between keyboard and mouse - either way. Editing, and most programming operations, is fundamentally a keyboard based operation, and hot keys are far more sensible than mouse gestures for this. But for me, browsing the web is a mouse-based operation. I have to point to links to follow them, so my hand is on the mouse. I have a wheelmouse, so scrolling is also under my fingers. The only gestures I use regularly are back and forward, and they have become so automatic I use them (uselessly0 wherever the model applies - i.e. in all "browser" type applications, such as Konqueror or Windows Explorer.
One thing we want to do is to try and get people to standardise. It will be a *real* pain if one piece of software used a gesture for minimise and another for quit.
I tried that - it didn't work. He really was PROM - write once. It wan't a question of looking good/bad, saving money or face.: his world model was formed and he wasn't changing it. Like arguing with a Jehovah's Witness - logic bounces off cast iron certainty.
CDMA only causes a lot of slowdown when you have a lot of collisions - i.e. when you have a lot of small packets fighting for the bandwidth. Which is another of the consequences of TCP/IP - flow control packets etc., and obviously occurs on a "workstation" environ,ment with many independant systems talking to servers. In a SAN type network using direct Ethernet packets, you would notrmally send bulk data in large packets which would swamp the bus for a while, then back off. But, like a scsi bus, you would tend to get one logical transaction loading the net for a while, then backing off while it seeks to the next logical block. And SAN type applications tend to have a few disk users, rather than hundreds of light users.
Firewire, Fibre Channel, Scsi etc. don't have this flexibility. They assume that the device you are talking to is is on this bus/ring and that messages will, give or take parity errors (which can be reported) get through reliably. Therefore they can have much ligher protocols. You can usee what, in Ethernet terms, is a MAC address not an IP address.
The firewire suggestion at the head is perfectly snesible. But if you were to go into the ethernet driver at the packet level, not at the OP level, I reckon you could get 90% of theoretical bandwidth out of an ethernet-based connection. We currently get over 500Mbit/sec using UDP connections. Drop a layer of protocol and it could be even faster.
I would agree with this in most cases. Management have (or should have) more info. They should, as Jon says, explain their thinking, and give your views a good hearing. If this happens, follow Jon's advice.
But it can happen that you are working for an idiot. It happened to me once in my career. In that case there is only one thing to do. GET OUT. If your manager is an incurable idiot, you will *never* be happy working there, and staying will only make you miserable. But give him/her plenty of rope to hang him/herself; a real idiot will soon show up unambiguously. If the idiocy is not tranparently obvious, then maybe it isn't idiocy.
(My idiot manager had a PROM brain. On day two, after looking over the job I was to do, I gave him a sketch of how I intended to attack the problem. On day about five, I found out that wasn't the right way - but it was too late; his proms were burned. For the next six months I had to shoehorn every weekly progress report into the headings from that initial handwaving guess at a solution.)
This is a classic complaint that things aren't as good as they were in the Golden Age - which almost always turns out to be the complainer's college days, or possibly his first job, when he was young and burning with enthusiasm. We have always built new tools on top of the tools our predecessors have built, and abandoned the tools they used in favour of the better ones - which, maybe, they built for us.
George Stevenson, the railway builder, judged the underlying quality of an engineer by his ability to file things out of raw metal, at a push. "Real Engineers", in his opinion, could file a threaded bolt our of bar stock if they really had to - not that they had to, because mass produced bolts were available, but a "real engineer" kept in touch with basics. How many of us would pass that test?
So what if some of the tools we use are software? While marketing may label the whizzy program "Engineer in a Box", it won't be - it'll be a new set of tools, better than the last but no panacea (anybody remeber "The Last One"). And many engineers will use those tools in stolid, unimaginative, ways, just as they do now. Don't tell me that every engineer in you plant ia imaginative and original. Most engineers, most of the time, are pulling on a few design patterns and a few tools to do essentially predictable jobs. Some of them, some of the time, do get original. And the more powerful the tools they have, the more creative they can get when the lightning of inspiration strikes.
This sort of thing has been said so many times that there is a Monty Python sketch - which I won't quote. In the early days of COBOL, it was supposed to de-skill programming. The manager would write down the business logic, a typist-class drudge would translate it into COBOL, and the compiler would do the rest. Assembler programmers complained that their craft was being destroyed before it was ever born. Is that the world you are living in?
Except you couldn't use it raw for backups. You would have to mount it in something reasonably shockproof, with insertable power and data connectors. This would add significantly to the cost.
Also, current OSs would probably not want you to put it onto your "main" IDE busses = they woudl assume that "fixed" drives on those busses are system drives and mount them only at startup. So you want it on another interface, with a card to connect to that interface and cabling to get out from the case to the removable unit, wherever you have put it.
Having said that, you can probably get to $100 fixed cost, $400 per exchangeable unit - which is not that much premium for very high speed and true random access.
The economists answer to this is to deflate future benefits by whatever the "zero risk" interest rate may be. Thus long term returns have to pay more, when they do pay, than the same amount of money left in a bank account.
The trouble with this is that it says that, for example, it is not economically worth saving the whales. According the this theory, we should simply slaughter the wales now, then invest the returns in something "useful".
Now, I cannot prove it mathematically, but to me this is wrong. But that is an emotional response.
So my real response is that you cannot objectively decide between short and long tem benefits. By all means do the economic calculations; look at what you are forgoeing on one plan for .benefits in the other. But allow non-financial factors to affect your decision. As well as, not instead of, financial considerations.
One of the things I like about the whole open source movement is that someone like Larry, who honestly (and probably correctly) says "I haven't an executive bone in my body" can get some of the kudos he undoubtedly deserves.
...., Larry has created a tool that literally thousands of people use - and value - every day. Respect, man.
Whatever the theological arguments about perl vs python vs Java vs
I workd for a companyy in hte UK which, you may have gathered, is pretty humid. We have anti-static wriststraps etc., but they get pretty well ignored, and we don't lose much kit to inexplicable (i.e. might be ESD) failures. But sometimes we take kit to exhibitions in Las Vegas which, as you might also have gathered, is pretty non-humid. And if we don't do it just right there, equipment dies by the rackful (particulalry before they close the big doors and turn on the aircon). And for one product I worked on, one customer in Phoenix, AZ, has almost as many failures as the whole of the rest of the world untils somebody clicked and installed the full gamut of ESD precautions.
So, it depends where you are.
But, as been sais, ESD precaustionsa re cheap, and kit is expensive. And, I have beeen told by our in-house ESD/RFI expert, a lot of ESD failures don't show up in instant faulures but in shortened lives - so a board will die after six months rather than five years. Doesn't sound to me like a good risk to take.
I knew a couple in a very similar situation - they both had good thech/manegerial jobs, but she was well up the career ladder in a very big corporation, while he was probably more liike most slashdotters - a while here, a while there, sometimes team leader sometimes not. So he went in for training for his tech discipline. It is much easier to get in and out of, because you are not involved in projects with timescales of months and a need to build a working team, but week-at-a-time courses. Seemed to work for them.
My brothere works in marketing, particularly consumer goods. He says that for candy bars - a field in which he has worked - if they stop advertising, sales die off exponentially with a half-life of two years. Where products are nearly interchangeable (there were probably 10 candy bars you might have bought) continuous flow of advertising is necessary to keep a particular one in the forefront.
Same is true of, say, washing powders. For any particular position in the market, there are usually at least three essentially interchangeable products. Advertising continuously ensures that shoppers pick *yours* rather than its identical competitors.
This is simply not true. HDTV is already running up to 1920 by 1500. For print work, very high quality magazines will scan bu to 10H by 8K. And 32 bits/pixel is not true either. Moset pictures don't need an alpha plane, so if you are doing 8 bit RGB you need 24 bits/pixel. Even if you want 210 bit video (and you are not going to get 10 good bits out of a CCD at that speed) it is only 30 bits - and the difference between 30 and 32 matters at these rates. Actually, for broascast purposes, YUV encodung with half bandwidth chrominance is used, reducing it to 16 bits/pixel at 8 bit sampling. And this example is probably monochrome anyway.
You are right that film has higher resolution than common-or-garden CCDs. Consumer 35mm is probably abput 4k*3k, with professional plate cameras, as used by some fachion photographers, being much better.
No point in going to a cluster - beowulf or otherwise. You just want to stream straight to disk. If, say, Infiniband takes off, you could stream striaght out of the camera to lots of disks. Don't shove the data into a CPU until you need to process it.
You may be right about which grandfather it was; the point is that the idea was far from new by the time Charles Darwin voyaged on the Beagle.
If he had not written about Natural Selection, you are correct that Wallace would have done so. However, while Wallace undoubtedly had the right idea (as Darwin acknowledged), he had not piled up the enormous pile of data, arguments, and counter arguments that Darwin had. If he had published, the idea would probably have taken twenty or thirty years to "catch". Darwin's book was so convincing and with such a mass of data that large numbers of scientists were immediately convinced ("How stupid of me not to think of that", as one of them remarked). This, of course, increased the heat of the argument of those who found the idea incompatible with their faith.
Arging "what-if", it might even have been better in the long term if Wallace had published, because the idea could have crept in insidiously instead of appearing as a massive frontal attack on (some people's) religion.
There goes another £29. Why do you do this to me? The pile of Books To Read is already a structural hazard.
Darwin certainly did not claim to have discovered evolution. The evidence for evolution of some sort was accepted by a large number (though far from all) scientists and interested people for some time before Darwin - amongst them, Darwin's Grandfather, Josiah Wedgewood, so the idea was far from new.
What Darwin did was find an explanation for evolution - a mechanism by which it occurred. Undoubtedly Lyell believed in, and pointed out to Darwin, the operation of evolution. And the ornithologist certainly pointed that all the finches he had brought back (and carelessly jumbled up) appeared to be descended from a singel ancestor. His book is titled "On The Origin of Species by Means of Natural Selection, or The Preservation of Favoured Races in the Struggle for Life", and it is the "Means of natural Selection" bit that is original.
To assert that Darwin claimed to have discovered Evolution is like claiming that Columbus discovered the Atlantic. Columbnus dicovered how to cross the atlantic, and Darwin discovered hopw to explain Evolution.
I don't think you should see it as typing, more as steering a course through "word space". Like driving, your hands are handling "here" while your eyes are yards ahead checking for problems and planning the course.
Dasher is inherently 2-D with very simple graphics. How about using a Doom-style rendering engine to develop a high-quality 3-D version of
Dasher? Use Doom/Quake style firing - pick off the next three letters you need, or finish off the word with a burst of automatic fire. Burn down an entire cliche in one shot. Install objects to represent frequently used phrases and type your manager's name by blasting him off the map.
(Not that I want to give you a Socially Responsible reason for going out and buying the latest, greatest graphics card, you understand).
Noone seems to want to answer the second part of the post...
For programming, it would have to be integrated with your IDE in an editor which is both syntax and semanitcally sensitive. Dasher uses some sort of a dictionary - usually in English - to predict what the next letter is most likely to be. This is obviously not a traditional dictionary, because it manages to predict across words.
For coding use, the dictionary, instead of being static, should be dynamic. Instead of having all the words in the English language, it should have only words (and symbols) which are semantically valid at the point the cursor is positioned in the edit buffer. Furthermore, it should weight them. Local variables are very likely, method variables are less likely, strange packages even less likely. In fact, it might subcategorise these into pseudo-letters, so that when writing (e.g.) Java, packages would appear as a single pseudo-letter. It could also add localisation-type information - entities referred to within a few lines are more likely to be accessed again soon. This doesn't change the sorting order used by Dasher, which is always alphabetic, but it does change "visual space" allocated to each letter in the search area.
You probably want an auto-beautifier i.e. new lines and indents/outdents are added automatically as needed.
This has some interesting side-effects. For example, you can only enter semantically valid programs. It has implicit auto-completion - once the following letters are unambigous, the remaining characters occupy the whole namespace.
The brackets case is also nice. The only close bracket that is ever possible is the close of the most recently opened, and that often has a very high probablity. Which means that inserting an open bracket implicity "arms" the system with the matching close bracket. The same applies for closing strings - the close string charager is always high probability.
For case significant languages there are some interesting effects. Obviously, in principle both upper and lower case must be present, though it may well be that a small minority of letters are accessible at any given instant. At a guesss, it would be better to split into Upper case and Lower case rather than interleave upper and lower - but that is something to experiment with.
Punctuation-type characters show minor problems - we all know alphabetic order, but does ";" come before or after "+"? I don't know. But I expect we could learn this - there aren't that many symbols.
One could also add a special pseudo-section for language-defined keywords, so you just chose a single prefix zone and then go straight to the set of all known keywords. Usuallly no more than 50 or so, and not usually all semantically valid, so you might get quite quick access to them.
Of course, there is a tradeoff with all these special zones. One of the points about Dasher is that you don't need any more specailised knowledge than having learned the alphabet to operate it. Adding language-sensitive zones and so on adds extra operator learning time. But since you have to learn the language anyway, I don't think it is that much of a burden.
I would expect this sort of strategy to (at least) double the input speed for Dasher for a particular programming language. Forget using your eyes - for the able at least - but it might make mouse-driven program-writing a lot faster. In fact, it might overtake typing for the special case of program input in a "known" language. Though I think the most valuable feature would actually be the inability to input a semantically erroneous program. Which means you "only" have to worry about logic bugs and not typing bugs.
A good place to try this would be to create a jEdit plugin. JEdit already has plugin Java browsers and beautifiers, so a lot of the code ought to be there already. A Java-style Dasher window would be a very interesting project. If anybody feels upt to doing this, I would like to help (I don't think I have got the time to lead such a project).
I checked through some of the links to try to find out why some of these sites don't want to be linked to. On thing that came out is that there is a mass of confused thinking and motivation out there. So don't expect a clean solution to this problem. A solution which will satisfy one set of paranoid suits will not satisfy the others.
One of the reasons is that they fear that the appearans of a link from you to them implies some sort of reciprocal approval i.e. that they know of and might be assumed to approve of you. Now, to anyone here, this is absolutely dumb, but corporate zecks and AOLers might not know better.
So here is an idea of how to deal with them. When they post court papers (which are surely public documents), post a reciprocal set of papers requiring them to remove your name, addresss, URL etc. from their papers because they imply they you endorse them etc. Use wording as close as possible to theirs and petition that your case be heard first.
One of two things happen: either the court is sensible and throws out your petition as riduculous, in which case you return with that rejection as a precedent, set in the same court, to justify your linking. Or the court grants your loony case, in which case (by the court's own loguic) they have to withdraw their case against you.
This project seems to me to be attaching what is the root of the problem with all renewables: concentration. When you come down to it, all renewable energy sources are solar power, though many are at second hand. Solar energy drives the weather for hydro, wind and wave, and grows the plants for biomass.
The problem with solar power is that, while there is masses of it, is is relatively diffuse, drops a lot on cloudy days, and comes and goes on a 24 hour cycle. All effective renewable power supplies depend upon some form of a concentrator to make them economically viable - at least for bulk domestic power, rather than specialised uses which can afford the premium cost of solar cells. They also need a load balancer for the 24 hour cycle.
Hydro achieves this by using the landscape - focussing the rainfall into a vally, a lake and a dam. Wind power uses an area (dimension ^2) of land/sea to generate wind captured by a line (dimension ^1) of windmills. These all concentrate solar power - but in a very lossy fashion. Hydro, for ecample, wasts much of the potential energy available as the rain hits the mountaintops "focussing" it into lakes at least half way down the mountain. Wave powere rewuires wint to drive the waves, then the waves to rool ove hunderds of miles to accumulate, before hitting the generator.
So the interesing part of this story is not really the tower, but the acres of greenhoses below and around it. This is a superb enegy concentrator device - and one, probably, with a lot of latent heat in the ground, so would run well after sunset. It is almost direct solar power - sun->hot air->power. If they can generate a lot of hot air, the tower is one way of converting it to electricity, but there may be others. But the point is that greenhouses can be economically constructed out of plastic film at a cost, I guess, 1% of the same area of solar cells. If they are only 10% as efficient at trapping the energy (and solar cells only run 5-15% efficeint), they are still winning by a large margin.
This idea is not actually very new. There was an SF story way back which proposed this sort of thing. Rather than build a tower, you just build a double walled cylinder out of the same polythene and inflate it with the hot air, (of which you have plenty) so that it lifts itself off. If it punctures, it will collapse very slowly (if at all, with many cells) and do little damage (because it willfall onto the greenhouses). Since the the tube is tranparent, a ring of lights round the base shining up into it will cause the whole thing to glow so that any pilot who flies into it must be truly blind.
There might be ecological consequneces - which need not be for the worse. The original SF story had the tubes puncturing the temperature inversion which causes AL's amoga and bringing a cooling draft to LA in the height of summer.
Excuse me - please can you explain why you considered my post a troll? It was purely a technical response to a point made someone else.
The only thing I can think is that you are taking it as approval of Star Wars, or Star Wars 2. Actually, I do not - I think it a damned bad idea; basically the military-industrical complex is taking the US taxpayer as suckers with a bottomless wallet. Buit that doesn't stop me making reasoned technical comments about the proposed system.
If we are to discuss such things, we should do so without knee-jerk reactions.