Unless the camera adds a secure signature to the image (which, of course, means that the original image as the camera produced it has to be available, no cropping or contrast adjustments or color balancing or re-compressing it.
Of course, then you could make a "validated" fake just by creating a fake and taking a picture of it with a secure camera. Ultimately, the only way an image should be admissible is for the person who took it (or who set up the equipment and removed the images, say in the case of a surveillance camera) to testify that they took the picture, it is an accurate representation of reality, and/or testify to the chain of custody and that no one could have tampered with it.
Why does your thermostat need to talk to my refrigerator? Why do I (at the office) even need to talk directly to my refrigerator (or your thermostat)? I should connect to "the house server" (and let it authenticate me) and tell it to "check refrigerator" or "set thermostat", I don't need to know what IP address they are. Almost nothing needs a routable address, much less a static address that can be connected to.
Using a number that is used for routing traffic as an identifier is nuts. It isn't even secure.
Now, I'm not saying that people shouldn't have servers, but that anything that isn't a server doesn't need a routable address. For temporary connections (say, I want to transfer a file to you), an intermediary server solution would work fine. I've always wanted an addition to the TCP protocol for the process of two processes that want to talk to each other, without having to have either one choose a role. Call it "meet" (instead of "listen" or "connect"). Such a protocol could be built in to router/NAT box - all I do is ask it for an identifier, communicate that identifier to the other side (through the server), receive the other side's identifier, tell the router who I'm trying to connect with, and let them establish things. That identifier would presumably be something like an IP address and a port, but I don't care anymore, I just use the connection I established with the router.
Even a server really only needs one port (not address, port). Instead of the port # triggering a specific process/protocol, explicitly identify it, let the software route it after that. And no, HTTP is not the protocol I'd choose to do that (despite everything these days getting layered over HTTP).
That current protocols are based on the current infrastructure is understandable, but not a reason to avoid alternative solutions.
Of course, there are some easy ways to allow for servers, even when going through a NAT box. Getting away from "well known ports" and "well known IP addresses" (which is really what makes for a server), you could go to a pure "presence" solution. Add a well-defined and supported protocol (as opposed to UPnP) with some security (server program has to be authorized) to allow listening on a random port, add a service name lookup (along the lines of RPC), and advertise your service through a global hierarchical name service (along the lines of dynamic DNS updating). Each level of NAT can contact the next level up if necessary, until you get to a backbone level that either uses static addresses or a dynamically configured architecture (like Rendezvous/OpenTalk/Zeroconf) with a hierarchical public key authentication system for those servers.
A good first pass at finding patents that should be investigated more closely would be to find ones with more than 1 standard deviation greater than the mean number of claims. When I see patents with 200-300 claims, I have to wonder if there's anything there at all.
If it says it conforms to the DVI standard, then yes, it must support single-link devices properly. It doesn't necessarily have to support any particular resolutions, but it seems unlikely they'd deliberately cripple it.
The DVI spec says that if a particular resolution CAN be done on a single-link, the adapter MUST use single-link. It can only shift up to dual-link if the bandwidth is too high at the chosen resolution, refresh rate and pixel color depth.
Another way of doing the same thing, and one that I think would be interesting to try out, is to use "big endian" storage (which is more natural for people, or at least people who read strings from left to right and read numbers with the most significant digits on the left), but the address of a value is the last (least significant) byte.
One problem with either scheme, that isn't shared by big endian references, is that accessing a field using the wrong sized reference is more immediately obvious in more cases.
The current exemption for 2-seat trainers is extended until 2008, but you're right that after that, such aircraft will need to be registered/certified under the new rules (as will their operators). However, the new rules also allow things currently prohibited, such as receiving compensation for rental, towing, flight training, and carrying passengers.
254 pounds (not including floats, parachutes, safety equipment) is a lot more than a "lawn chair strapped to a wing".
I'd say that keeping yourself oriented to and following the various instrument procedures required while flying IFR (other than "radar vectors for the ILS") is easily more difficult than landings. Actually flying on instruments isn't that difficult, but I'll bet your workload goes up way more than for a landing when ATC says "Radar contact lost" and you now have to figure out ETA to your next fix and the next reporting point after that. Adjusting a hold so you can exit at the proper time. Flying outbound (i.e. backcourse) on the ILS, just after takeoff, putting on the hood, while ATC gives you clearance to hold on the ILS (inbound) at the VOR at the altitude you're about to overshoot, plus clearances for your next three approaches, on an instrument checkride, and you're not even a licensed pilot yet. Tell me landing is more difficult than that!
You'll also probably find that flying on tow in a glider is more difficult than landing (a glider or an airplane). Being a CFI-Glider, I've seen a fair number of transition pilots, and takeoff and tow is much more difficult than landing for them. Most of them have it down pretty well within 3-5 flights, but its still an area that they're uncomfortable with for a while.
I practice landing all the time as well. I get at least one in for each takeoff I do.
So then why do (some) BSD advocates cry foul when BSD code is mixed with GPL code and the whole thing is released under the GPL? The original BSD code is "just as free", and to suggest that the person who wrote additional code is in any way wrong to not want their code to be subject to appropriation into non-open-source products is -- well -- sort of non-free, isn't it?
GPL advocates are consistent - they have an avowed goal of at least not helping the closed-source cause, which is consistent with advocating that others not help either (by avoiding licenses such as BSD that can be used by closed-source developers). BSD advocates of the sort who cry that GPL is "less free" (and thus, presumably, shouldn't be used in favor of a BSD license?) seem to be a bit hypocritical; if "free" means you're free to do what you want with your code, why should it bother them in the slightest that someone wants to license their code under the GPL, call it more free, and advocate that everyone use the GPL instead of the BSD? Isn't that the kind of freedom that they're advocating for?
The great thing about hypocrisy is that once you've accepted it in yourself, you can still criticize it in others!
I can transfer copyright in code I've written, even if I've released it under the GPL. I can re-license it under more or less restrictive terms. I can make derivative works and release them as binary-only. The only thing I can't do is incorporate other's work into mine and do with it as I please without permission, which is true regardless of whether I've released it under the GPL. Releasing it under the GPL does give me that permission for a large body of code.
The whole argument about what "freedom" means is stupid. One person's freedom is another's non-freedom. North Korea is the freest place on earth, as long as you're Kim Jong Il. It only makes sense in context. RMS meant "freedom" in the context of the end-users being able to have control over software they obtain. As part of that, developers also have the freedom to modify and distribute such software, provided they don't take away the fundamental freedom RMS intended. That's what the GPL is about, that's what "freedom" in the context of the GPL means.
In the context of the BSD zealots, "freedom" means absolute freedom for the developer, not the end user. The developer is free to take away "GPL freedom" from the end user (and, incidentally, from other developers). Other developers are no longer free to further modify code that has been made proprietary (because the BSD license allows it). Sure, you can still modify the original, but the original isn't the code that's controlling your PVR, DVD player, home security system, and that's the code you as either an end-user or developer wants to modify. The original code is absolutely useless in that situation.
There is no such thing as absolute freedom.
Freedom's just another word for nothing left to hack
So if I can't lock you up and sell you to work the fields, I'm not free? Damn those strings attached.
BSD license also has restrictions. If not, then why isn't it just released as Public Domain, disclaiming all copyright or licenses? Aren't we down to the "we know what you are, now we're just haggling over the price" stage?
One way to think about the GPL is that it isn't about "freedom of the programmer", but "freedom of the code". The code itself can not be made a slave and locked away. That gives "freedom to the user", but not "freedom to the slaveowner" (someone who would modify the code and keep the source hidden). It doesn't require anything of the user of the code, only of someone who would modify and distribute it. Seems pretty damned free to me.
DEC's IA-32 emulation under Windows was flawless and quite fast. The reason the Alpha didn't become a popular architecture was (a) Digital didn't know how to sell it; (b) Microsoft dropped it; (c) people keep buying what they're used to buying. If they'd put as much money into increasing the speed and decreasing the power, say by moving to a modern process size, as Intel put into developing the IA-64 architecture, Alpha chips would be twice as fast as they are and using half the power (that is to say, totally blowing away anything else that's currently available)
When Intel first started talking Merced, and how it would have hardware support for emulating IA-32, my first thought was that it would be a total failure unless it could run x86 programs at least as fast as mid-range 32-bit processors available at that time. Otherwise, switching to Alpha, with its emulation, would be just as good, with the 64-bit aspects already old and stable technology. Of course, that was before Compaq bought DEC, then blew any possible synergy and sold out to HP, which had an active reason to kill Alpha, and that was that.
As far as 64-bitness being faster, there are various techniques that can take advantage of 64-bit integer operations for "ordinary" programming tasks. Alpha libraries do an amazingly good job of doing character manipulation even though the Alpha has no 8-bit instructions, for example, or using bitmaps for block allocation. String searching and sorting, doing unordered data sieves on multiple bit fields - many techniques dating back to the 60-bit Cyber architecture. Most programmers don't use them because doing them using "long long" on a 32-bit processor slows you down.
So, no, 64-bit processors being able to address more virtual/real memory is not the only advantage.
So make copyright last a minimum of 25 years, a maximum of (the lesser of) 75 years or the life of the author, and 50 years for a corporate owner (with copyright reverting to the original author, in the case of a transferred work, after 50 years if that author is still alive; he gets it for up to 25 more years). An author kicks off after only a few years, his heirs can still receive income on it up to the 25 year minimum.
Allowing the terms on old works to change implies that the length could also be reduced. Isn't that less stable than saying "you get the term you started with"? Your incentive for creating and distributing a work is whatever the current term of copyright is. If we change the bargain to "oh, and we'll change the length of the copyright to be longer, but never shorter, and it will apply to your work as well", then not extending copyright in the future will be taken as "stealing" from those poor disenfranchised authors from 250 years ago (i.e. now) who were assured that copyright would be extended indefinitely.
If legislation is passed to require that the length of copyright for a particular work is limited (and guaranteed) to the length at time of creation, I hope we at least rescind the extended terms that have been granted the last few rounds first. Like that would happen.
Demand is low, so price is high? How is that "simple supply and demand theory", especially when the supply of Linux is infinite (because of no per-copy license charge, your only expense is labor to install it, which is no different from installing Windows)?
Even when you have an excellent case, it might never be heard. DEC, I think, had a claim that they "never" lost a patent case. If true, it would be because whenever they were about to lose, they'd turn around and settle and avoid having their patent invalidated.
I was involved in a case where DEC was suing a peripheral manufacturer for using a patented method of communicating with the operating system. Given that this was the only way to communicate with the operating system, this was a significant barrier to making compatible peripherals. The patent involved "ring buffers", and the patented improvement involved storing a pointer to a buffer in the queue rather than having the buffer actually be part of the queue. This allowed removing buffers from the queue for processing, allowing out-of-order processing, etc. Time scale of this patent was in the late '70s.
Our system used the exact same method, and we could document that not only had we been doing it earlier, the reasoning by the programmer who did the work to change the queue scheme was identical to the description in the patent. What was different from many patent cases is that we had kept the on-line discussions where people announced what they had done for the other developers.
We printed out the source code, including time stamps showing when the code had originally been written, and the discussion of the advantages of the method, and gave that to the company being sued. We heard later that they had "won", yet the patent remained in effect until it expired naturally.
Another case I was involved in helping out had to do with Bingo. Bingo, it turns out, is a huge industry, and there are an amazing number of patents. This patent was for a way of using a physical bingo card, with an overlay and some electrical contacts, to insert into a machine. The machine thus knew what the card layout was, and could call a bingo when it came up. It also used a centralized server to call the numbers and send them out to the various bingo machines. The physical card was simply so that the user could see which numbers were on the card. Somehow, this patent got expanded to include virtually any means of playing bingo on a computer!
When we provided "prior art" in the form of an on-line interactive bingo game, whose internal logic was almost identical to the patented game (one player was chosen to be the caller - his instance of the program generated the numbers, and distributed them through a shared memory queue with the other instances of the program). The only difference was that (a) our program didn't automatically "call" bingo - you had to press a key to claim it, and the program verified it. Isn't that the point of playing Bingo? And (b) our program was implemented on a time-shared system, not a distributed system with a server and a bunch of game machines.
I heard on that one that it eventually was overturned. However, our "prior art" was rejected by the judge because of the distributed processing issue, because the actual decision as to whether it was a bingo wasn't made in the individual "game machine". However, the patent was being used against systems that did all the logic in the server machine, where the boards were generated and the bingos were checked. In other words, the interpretation of the limitations of the patent were different when being used against a supposedly infringing implementation than when determining whether the prior art was valid.
Unfortunately, that patent was overturned after the company that was being sued went bankrupt.
In two out of two experiences I've had with trying to bust an obvious and non-novel patent, with clear prior art, in neither one did the system work well. The Bingo case in particular should have been one of obviousness - they essentially started out with "let's patent playing bingo on a computer", came up with a patentable idea (the thing with the cards, which I think was clearly patentable) and expanded it to cover all
You could also do a consumption (sales/VAT), with a flat deduction. Make the flat deduction in the form of a tax rebate check each year to everyone, without exception, even to people with no income. Take the savings on processing income tax forms and use it to pay all the tax lawyers and accountants for a few years until they can get useful jobs.
The hard drives I bought about 15 years ago for the relatively constant sum of around $200 ("standard sized" hard drives have sold for about $200, just as "the really cool computer you wish you could afford" have sold for about $3000, since somewhere between 1985 and 1990, sort of amazing to observe actually) would cost about $0.06-.25 today if figured as a portion of the current $200 hard drive. I don't see any reason not to believe that 250GB in 15 years will be worth between $0.06-.25 as well, as part of your $200 200-900TB disk drive equivalent. The time to transfer all of the data on such a drive has also remained relatively constant. I can copy the entire contents of my first 10MB hard drive in about a second or two. Mainframe system backups that took 6-7 hours can now be copied to CD in 10 minutes. Don't limit yourself to the very short amount of time you've been familiar with technology.
One system I used had about 4.5GB of data, built up over a period of 15 years with thousands of people. Regarding the earlier claim, the only way an individual can generate "80-100GB" of data is if its video, you're a professional or very dedicated photographer, or you're simply archiving other people's work (in which case, it isn't really your data, it's just a cache - still important to allow you to access it even if it becomes unavailable elsewhere, but hardly something you need with you at all times, or a major disaster if it is wiped out (since you do back it up at least once after you've acquired it, right? And since it isn't changing, it doesn't need to be backed up every time). Even if you're producing audio, 100GB is about 150 hours (more than 6 days!) worth of uncompressed audio at CD sampling, or around 36 straight days compressed to 256kbps. 100GB is 250,000 400KB images; that's 70 hours to look at them all, at 1 per second. As text, it would take about 12.5 years to read it all, at a reading speed of 1500 wpm. Without sleeping!
A more realistic view is that "most people" use hardly any of their hard drives, these days. What is there they don't remember where it came from, and if they do know its "on there somewhere", they're not really sure where. If they do accumulate a GB or two, they have no idea how to back it all up, and the few MB of stuff that's really theirs that they'd be upset at losing becomes literally lost.
I have a 20GB partition that contains disk images of previous systems, some of which contain disk images of previous systems themselves, nesting to 3 or 4 levels back to 1985. It has accumulated junk, images, programs, stuff I've written, saved newsgroup articles, all my e-mail, my girlfriend's accumulated saved images off the web from several years of browsing, an audio CD I ripped but forgot to delete, installers, system updates, backups of backups, etc. I still have 11GB left on that partition. I have a 130GB partition on a 250GB drive. After copying ALL of my backup CDs (mostly identical, and mostly containing variations on the same stuff as the previously mentioned disk images, with stuff copied around between systems being duplicated multiple times), I still have 110GB left. That's going back about 4 years. Copying all of my backup diskettes will probably take up several hundred more megabytes.
20 years ago or so, 14.4Kbps modems were cutting edge, slightly before that 2400bps, a few more years back 1200, 300 or 110 bps. 3Mbps is fairly standard now, why wouldn't we have 500Mbps or more standard in 10-20 years?
However, the idea of using a (communal) computer to do things on, giving it access to your entire digital existance, is absurd, and the reason is one of trust. You have to trust the owner of the machine to have not bugged it, and to have kept it secure so no one else can have bugged it. Much more likely is a ubiquitous wireless network, wired ports for even higher speed available as telephones are now (complimentary or for a small fee, depending on context), and small wearable inconspicuous computers with highly portable I/O - either voice response (e.g. using sub-vocal nerve impulse detection), virtual or projected keyboards, unobtrusive data gloves, 3D overlay displays beamed into your eyes (using eyeball and head tracking; put on a pair of sunglasses if you want to be able to block out the rest of the world and see only your virtual display). You never have to configure your computer becauuse it is always with you. Plug into a network or a local system, but trust your data integrity by using secure protocols between you and your home data store when you want to access data you're not carrying in the 1-terabyte chunk you carry around with you.
Of course, computers in the home will be completely different as well. The "home server" will be always-on, taking messages, monitoring security, controlling environment, providing wireless and wired network connections, and providing data storage. There probably would be screens and control stations throughout the house (so you don't necessarily have to wear your wearable portable system at home).
No, you recover as much as you can from the current session, roll back one day to before the virus did any damage, recover your data from then, roll it back to before the virus was even there (or disinfect your system now that you know the virus is there and what it does), and restore all your data.
One use for all the excess computational power we have now and will have even more of in the future (think 3GHz to read e-mail, surf the web, and wait for a 5 key/minute typer to enter a shopping list is bad, wait until we're at 10 or 50GHz, 5GB RAM, 1.5TB disk) would be to run everything in virtual, even emulated sandbox environments. Expect Java, for instance, to become more popular as performance issues are blown away by raw horsepower.
Usually, you can get this fixed by calling the right place and being insistent.
I bought something at Best Buy that had three rebates. Two said they required the original UPC, the third didn't require it. So I called the Best Buy rebate center to enquire, and they told me that one of them could, in fact, be a photocopy.
6 weeks later, I got a postcard for that one saying my rebate was invalid. I called, explained what had happened. She said "we can't take photocopies." Insistent. Never. Take. Photocopies. Never never never. Absolutely impossible. Went back and forth, asked for a supervisor, complained about false advertising, she finally said "One moment." 2 minutes later she came back and said "I guess we do take photocopies." then told me to re-submit everything, with the photocopies I had made. Fortunately, I had them.
What's odd is, they always say "photocopy all your materials for proof of what you sent in." So if I just claimed that I did so send in the original UPC, and I have the photocopies to prove it, what are they going to do?
There was another one where they sent me back a rejection, claiming that the purchase date was incorrect. Looking closely at the form, the data entry person had entered the receipt date as the purchase date (which should have been noticed since it was AFTER the postmark date they recorded, and besides how would I get the store-printed rebate form?). That one was actually handled pretty well when I called to complain, and they authorized processing without my having to send anything else in. I think that might have been Western Digital.
I also had one where I got my letter returned, with "P.O. Box closed". I sent it the day after I bought it, which was the last purchase date of the offer, and it said postmark by 1 week later. "Somehow", it took 1 week, 2 days to get there, and they had closed the box by then. Never did get that one resolved.
Doing things by e-mail is almost always impossible. I think they specially train customer-service people to not understand what you said, give you advice you already said you tried, and if you asked more than one question, to only answer the one that they can misunderstand the most. Amazon is really good at all of this, for instance. Their phone customer support used to be wonderful, then they changed and made it almost impossible to talk to them on the phone.
I don't understand how it could be made so large it couldn't be an internal drive. No one is going to go for a new format that is larger than a laserdisc, or even an 8" floppy disk. Both of them could be easily loaded into a desktop machine, or vertically in a tower. Maybe really small towers, iMacs and laptops would have problems, but it wouldn't really stop anyone.
External SCSI CD-ROM drives were quite common for Macs for a while. Maybe people in the PC world didn't notice, as they had no decent external high speed standard.
Apartments above stores are common all over the place, it's not bizarre at all. Often you never notice because you're not looking up. They're in big cities, little cities, rural towns, United States and Europe, and probably everywhere else.
Reminds me of Charles Duell's fateful statement: "Everything that can be invented, has been invented."
I think the current situation is actually the exact opposite. Too many things are being invented, with too low a threshold of what counts as "invention". This creates an ever-increasing tarpit that makes it more difficult to create true innovation, with (to mix metaphors) large minefields for new products that are found to have infringed on bogus patents by courts that misinterpret what the patent covers, what prior art actually exists, and what "novel" and "obvious" should mean.
Whereas the body of technology grows hugely every moment, the inflow of new patents is (to some extent) offset by the expiration of old patents.
"to some extent" is overstating it, I think. The number of unexpired patents right now is much larger than the number of unexpired patents 20 or 50 years ago. Patents are harmful when it is difficult to determine if what you are producing might infringe on a patent. You either spend most of your effort on patent searches, or you risk the uncertainty of being sued later. Patents should help you produce new products - find a solution to what you're trying to do so you can license it for (some amount) less than it would cost you to come up with the same solution. When it is used strictly to enforce a monopoly, preventing others from innovating, it costs more to society than the good a patent is supposed to be doing by encouraging innovation.
In addition, with bogus patents, the public never actually gets the reward it is supposed to receive after the patent expires, as the patent never actually reveals anything that wasn't already known before. Unless the gains to society are greater than the costs, each particular patent shouldn't have been granted. An inventor isn't owed protection; society chooses to grant it for the greater good of all.
The one good thing about all this is that it forces exploration of entirely new fields. Unfortunately, that means that those fields will soon be clogged with overreaching patents which will stifle innovation for 20 years.
In an earlier post, you mentioned RSA as being an example of a good software patent. The problem with that is, while innovative and non-obvious, it was also too broad. It prevented many other schemes which had true inventiveness from being used, and had the effect of killing a lot of development of public key cryptography until the patent ran out. Without the RSA patent, we might be living in a much more secure world, communications-wise (although US munitions export laws didn't help either).
Slashdot Mirror
Unless the camera adds a secure signature to the image (which, of course, means that the original image as the camera produced it has to be available, no cropping or contrast adjustments or color balancing or re-compressing it.
Of course, then you could make a "validated" fake just by creating a fake and taking a picture of it with a secure camera. Ultimately, the only way an image should be admissible is for the person who took it (or who set up the equipment and removed the images, say in the case of a surveillance camera) to testify that they took the picture, it is an accurate representation of reality, and/or testify to the chain of custody and that no one could have tampered with it.
Why does your thermostat need to talk to my refrigerator? Why do I (at the office) even need to talk directly to my refrigerator (or your thermostat)? I should connect to "the house server" (and let it authenticate me) and tell it to "check refrigerator" or "set thermostat", I don't need to know what IP address they are. Almost nothing needs a routable address, much less a static address that can be connected to.
Using a number that is used for routing traffic as an identifier is nuts. It isn't even secure.
Now, I'm not saying that people shouldn't have servers, but that anything that isn't a server doesn't need a routable address. For temporary connections (say, I want to transfer a file to you), an intermediary server solution would work fine. I've always wanted an addition to the TCP protocol for the process of two processes that want to talk to each other, without having to have either one choose a role. Call it "meet" (instead of "listen" or "connect"). Such a protocol could be built in to router/NAT box - all I do is ask it for an identifier, communicate that identifier to the other side (through the server), receive the other side's identifier, tell the router who I'm trying to connect with, and let them establish things. That identifier would presumably be something like an IP address and a port, but I don't care anymore, I just use the connection I established with the router.
Even a server really only needs one port (not address, port). Instead of the port # triggering a specific process/protocol, explicitly identify it, let the software route it after that. And no, HTTP is not the protocol I'd choose to do that (despite everything these days getting layered over HTTP).
That current protocols are based on the current infrastructure is understandable, but not a reason to avoid alternative solutions.
Of course, there are some easy ways to allow for servers, even when going through a NAT box. Getting away from "well known ports" and "well known IP addresses" (which is really what makes for a server), you could go to a pure "presence" solution. Add a well-defined and supported protocol (as opposed to UPnP) with some security (server program has to be authorized) to allow listening on a random port, add a service name lookup (along the lines of RPC), and advertise your service through a global hierarchical name service (along the lines of dynamic DNS updating). Each level of NAT can contact the next level up if necessary, until you get to a backbone level that either uses static addresses or a dynamically configured architecture (like Rendezvous/OpenTalk/Zeroconf) with a hierarchical public key authentication system for those servers.
See, I told you it would be easy!
A good first pass at finding patents that should be investigated more closely would be to find ones with more than 1 standard deviation greater than the mean number of claims. When I see patents with 200-300 claims, I have to wonder if there's anything there at all.
If it says it conforms to the DVI standard, then yes, it must support single-link devices properly. It doesn't necessarily have to support any particular resolutions, but it seems unlikely they'd deliberately cripple it.
The DVI spec says that if a particular resolution CAN be done on a single-link, the adapter MUST use single-link. It can only shift up to dual-link if the bandwidth is too high at the chosen resolution, refresh rate and pixel color depth.
Another way of doing the same thing, and one that I think would be interesting to try out, is to use "big endian" storage (which is more natural for people, or at least people who read strings from left to right and read numbers with the most significant digits on the left), but the address of a value is the last (least significant) byte.
One problem with either scheme, that isn't shared by big endian references, is that accessing a field using the wrong sized reference is more immediately obvious in more cases.
The current exemption for 2-seat trainers is extended until 2008, but you're right that after that, such aircraft will need to be registered/certified under the new rules (as will their operators). However, the new rules also allow things currently prohibited, such as receiving compensation for rental, towing, flight training, and carrying passengers.
254 pounds (not including floats, parachutes, safety equipment) is a lot more than a "lawn chair strapped to a wing".
I'd say that keeping yourself oriented to and following the various instrument procedures required while flying IFR (other than "radar vectors for the ILS") is easily more difficult than landings. Actually flying on instruments isn't that difficult, but I'll bet your workload goes up way more than for a landing when ATC says "Radar contact lost" and you now have to figure out ETA to your next fix and the next reporting point after that. Adjusting a hold so you can exit at the proper time. Flying outbound (i.e. backcourse) on the ILS, just after takeoff, putting on the hood, while ATC gives you clearance to hold on the ILS (inbound) at the VOR at the altitude you're about to overshoot, plus clearances for your next three approaches, on an instrument checkride, and you're not even a licensed pilot yet. Tell me landing is more difficult than that!
You'll also probably find that flying on tow in a glider is more difficult than landing (a glider or an airplane). Being a CFI-Glider, I've seen a fair number of transition pilots, and takeoff and tow is much more difficult than landing for them. Most of them have it down pretty well within 3-5 flights, but its still an area that they're uncomfortable with for a while.
I practice landing all the time as well. I get at least one in for each takeoff I do.
So then why do (some) BSD advocates cry foul when BSD code is mixed with GPL code and the whole thing is released under the GPL? The original BSD code is "just as free", and to suggest that the person who wrote additional code is in any way wrong to not want their code to be subject to appropriation into non-open-source products is -- well -- sort of non-free, isn't it?
GPL advocates are consistent - they have an avowed goal of at least not helping the closed-source cause, which is consistent with advocating that others not help either (by avoiding licenses such as BSD that can be used by closed-source developers). BSD advocates of the sort who cry that GPL is "less free" (and thus, presumably, shouldn't be used in favor of a BSD license?) seem to be a bit hypocritical; if "free" means you're free to do what you want with your code, why should it bother them in the slightest that someone wants to license their code under the GPL, call it more free, and advocate that everyone use the GPL instead of the BSD? Isn't that the kind of freedom that they're advocating for?
The great thing about hypocrisy is that once you've accepted it in yourself, you can still criticize it in others!
I can transfer copyright in code I've written, even if I've released it under the GPL. I can re-license it under more or less restrictive terms. I can make derivative works and release them as binary-only. The only thing I can't do is incorporate other's work into mine and do with it as I please without permission, which is true regardless of whether I've released it under the GPL. Releasing it under the GPL does give me that permission for a large body of code.
The whole argument about what "freedom" means is stupid. One person's freedom is another's non-freedom. North Korea is the freest place on earth, as long as you're Kim Jong Il. It only makes sense in context. RMS meant "freedom" in the context of the end-users being able to have control over software they obtain. As part of that, developers also have the freedom to modify and distribute such software, provided they don't take away the fundamental freedom RMS intended. That's what the GPL is about, that's what "freedom" in the context of the GPL means.
In the context of the BSD zealots, "freedom" means absolute freedom for the developer, not the end user. The developer is free to take away "GPL freedom" from the end user (and, incidentally, from other developers). Other developers are no longer free to further modify code that has been made proprietary (because the BSD license allows it). Sure, you can still modify the original, but the original isn't the code that's controlling your PVR, DVD player, home security system, and that's the code you as either an end-user or developer wants to modify. The original code is absolutely useless in that situation.
There is no such thing as absolute freedom.
Freedom's just another word for nothing left to hack
So if I can't lock you up and sell you to work the fields, I'm not free? Damn those strings attached.
BSD license also has restrictions. If not, then why isn't it just released as Public Domain, disclaiming all copyright or licenses? Aren't we down to the "we know what you are, now we're just haggling over the price" stage?
One way to think about the GPL is that it isn't about "freedom of the programmer", but "freedom of the code". The code itself can not be made a slave and locked away. That gives "freedom to the user", but not "freedom to the slaveowner" (someone who would modify the code and keep the source hidden). It doesn't require anything of the user of the code, only of someone who would modify and distribute it. Seems pretty damned free to me.
DEC's IA-32 emulation under Windows was flawless and quite fast. The reason the Alpha didn't become a popular architecture was (a) Digital didn't know how to sell it; (b) Microsoft dropped it; (c) people keep buying what they're used to buying. If they'd put as much money into increasing the speed and decreasing the power, say by moving to a modern process size, as Intel put into developing the IA-64 architecture, Alpha chips would be twice as fast as they are and using half the power (that is to say, totally blowing away anything else that's currently available)
When Intel first started talking Merced, and how it would have hardware support for emulating IA-32, my first thought was that it would be a total failure unless it could run x86 programs at least as fast as mid-range 32-bit processors available at that time. Otherwise, switching to Alpha, with its emulation, would be just as good, with the 64-bit aspects already old and stable technology. Of course, that was before Compaq bought DEC, then blew any possible synergy and sold out to HP, which had an active reason to kill Alpha, and that was that.
As far as 64-bitness being faster, there are various techniques that can take advantage of 64-bit integer operations for "ordinary" programming tasks. Alpha libraries do an amazingly good job of doing character manipulation even though the Alpha has no 8-bit instructions, for example, or using bitmaps for block allocation. String searching and sorting, doing unordered data sieves on multiple bit fields - many techniques dating back to the 60-bit Cyber architecture. Most programmers don't use them because doing them using "long long" on a 32-bit processor slows you down.
So, no, 64-bit processors being able to address more virtual/real memory is not the only advantage.
Isn't saying "English is the lingua franca of ..." sort of like saying "Mercedes-Benz is the Cadillac of cars"?
So make copyright last a minimum of 25 years, a maximum of (the lesser of) 75 years or the life of the author, and 50 years for a corporate owner (with copyright reverting to the original author, in the case of a transferred work, after 50 years if that author is still alive; he gets it for up to 25 more years). An author kicks off after only a few years, his heirs can still receive income on it up to the 25 year minimum.
Allowing the terms on old works to change implies that the length could also be reduced. Isn't that less stable than saying "you get the term you started with"? Your incentive for creating and distributing a work is whatever the current term of copyright is. If we change the bargain to "oh, and we'll change the length of the copyright to be longer, but never shorter, and it will apply to your work as well", then not extending copyright in the future will be taken as "stealing" from those poor disenfranchised authors from 250 years ago (i.e. now) who were assured that copyright would be extended indefinitely.
If legislation is passed to require that the length of copyright for a particular work is limited (and guaranteed) to the length at time of creation, I hope we at least rescind the extended terms that have been granted the last few rounds first. Like that would happen.
Demand is low, so price is high? How is that "simple supply and demand theory", especially when the supply of Linux is infinite (because of no per-copy license charge, your only expense is labor to install it, which is no different from installing Windows)?
Adam Smith is so misunderstood.
Even when you have an excellent case, it might never be heard. DEC, I think, had a claim that they "never" lost a patent case. If true, it would be because whenever they were about to lose, they'd turn around and settle and avoid having their patent invalidated.
I was involved in a case where DEC was suing a peripheral manufacturer for using a patented method of communicating with the operating system. Given that this was the only way to communicate with the operating system, this was a significant barrier to making compatible peripherals. The patent involved "ring buffers", and the patented improvement involved storing a pointer to a buffer in the queue rather than having the buffer actually be part of the queue. This allowed removing buffers from the queue for processing, allowing out-of-order processing, etc. Time scale of this patent was in the late '70s.
Our system used the exact same method, and we could document that not only had we been doing it earlier, the reasoning by the programmer who did the work to change the queue scheme was identical to the description in the patent. What was different from many patent cases is that we had kept the on-line discussions where people announced what they had done for the other developers.
We printed out the source code, including time stamps showing when the code had originally been written, and the discussion of the advantages of the method, and gave that to the company being sued. We heard later that they had "won", yet the patent remained in effect until it expired naturally.
Another case I was involved in helping out had to do with Bingo. Bingo, it turns out, is a huge industry, and there are an amazing number of patents. This patent was for a way of using a physical bingo card, with an overlay and some electrical contacts, to insert into a machine. The machine thus knew what the card layout was, and could call a bingo when it came up. It also used a centralized server to call the numbers and send them out to the various bingo machines. The physical card was simply so that the user could see which numbers were on the card. Somehow, this patent got expanded to include virtually any means of playing bingo on a computer!
When we provided "prior art" in the form of an on-line interactive bingo game, whose internal logic was almost identical to the patented game (one player was chosen to be the caller - his instance of the program generated the numbers, and distributed them through a shared memory queue with the other instances of the program). The only difference was that (a) our program didn't automatically "call" bingo - you had to press a key to claim it, and the program verified it. Isn't that the point of playing Bingo? And (b) our program was implemented on a time-shared system, not a distributed system with a server and a bunch of game machines.
I heard on that one that it eventually was overturned. However, our "prior art" was rejected by the judge because of the distributed processing issue, because the actual decision as to whether it was a bingo wasn't made in the individual "game machine". However, the patent was being used against systems that did all the logic in the server machine, where the boards were generated and the bingos were checked. In other words, the interpretation of the limitations of the patent were different when being used against a supposedly infringing implementation than when determining whether the prior art was valid.
Unfortunately, that patent was overturned after the company that was being sued went bankrupt.
In two out of two experiences I've had with trying to bust an obvious and non-novel patent, with clear prior art, in neither one did the system work well. The Bingo case in particular should have been one of obviousness - they essentially started out with "let's patent playing bingo on a computer", came up with a patentable idea (the thing with the cards, which I think was clearly patentable) and expanded it to cover all
You could also do a consumption (sales/VAT), with a flat deduction. Make the flat deduction in the form of a tax rebate check each year to everyone, without exception, even to people with no income. Take the savings on processing income tax forms and use it to pay all the tax lawyers and accountants for a few years until they can get useful jobs.
The hard drives I bought about 15 years ago for the relatively constant sum of around $200 ("standard sized" hard drives have sold for about $200, just as "the really cool computer you wish you could afford" have sold for about $3000, since somewhere between 1985 and 1990, sort of amazing to observe actually) would cost about $0.06-.25 today if figured as a portion of the current $200 hard drive. I don't see any reason not to believe that 250GB in 15 years will be worth between $0.06-.25 as well, as part of your $200 200-900TB disk drive equivalent. The time to transfer all of the data on such a drive has also remained relatively constant. I can copy the entire contents of my first 10MB hard drive in about a second or two. Mainframe system backups that took 6-7 hours can now be copied to CD in 10 minutes. Don't limit yourself to the very short amount of time you've been familiar with technology.
One system I used had about 4.5GB of data, built up over a period of 15 years with thousands of people. Regarding the earlier claim, the only way an individual can generate "80-100GB" of data is if its video, you're a professional or very dedicated photographer, or you're simply archiving other people's work (in which case, it isn't really your data, it's just a cache - still important to allow you to access it even if it becomes unavailable elsewhere, but hardly something you need with you at all times, or a major disaster if it is wiped out (since you do back it up at least once after you've acquired it, right? And since it isn't changing, it doesn't need to be backed up every time). Even if you're producing audio, 100GB is about 150 hours (more than 6 days!) worth of uncompressed audio at CD sampling, or around 36 straight days compressed to 256kbps. 100GB is 250,000 400KB images; that's 70 hours to look at them all, at 1 per second. As text, it would take about 12.5 years to read it all, at a reading speed of 1500 wpm. Without sleeping!
A more realistic view is that "most people" use hardly any of their hard drives, these days. What is there they don't remember where it came from, and if they do know its "on there somewhere", they're not really sure where. If they do accumulate a GB or two, they have no idea how to back it all up, and the few MB of stuff that's really theirs that they'd be upset at losing becomes literally lost.
I have a 20GB partition that contains disk images of previous systems, some of which contain disk images of previous systems themselves, nesting to 3 or 4 levels back to 1985. It has accumulated junk, images, programs, stuff I've written, saved newsgroup articles, all my e-mail, my girlfriend's accumulated saved images off the web from several years of browsing, an audio CD I ripped but forgot to delete, installers, system updates, backups of backups, etc. I still have 11GB left on that partition. I have a 130GB partition on a 250GB drive. After copying ALL of my backup CDs (mostly identical, and mostly containing variations on the same stuff as the previously mentioned disk images, with stuff copied around between systems being duplicated multiple times), I still have 110GB left. That's going back about 4 years. Copying all of my backup diskettes will probably take up several hundred more megabytes.
20 years ago or so, 14.4Kbps modems were cutting edge, slightly before that 2400bps, a few more years back 1200, 300 or 110 bps. 3Mbps is fairly standard now, why wouldn't we have 500Mbps or more standard in 10-20 years?
However, the idea of using a (communal) computer to do things on, giving it access to your entire digital existance, is absurd, and the reason is one of trust. You have to trust the owner of the machine to have not bugged it, and to have kept it secure so no one else can have bugged it. Much more likely is a ubiquitous wireless network, wired ports for even higher speed available as telephones are now (complimentary or for a small fee, depending on context), and small wearable inconspicuous computers with highly portable I/O - either voice response (e.g. using sub-vocal nerve impulse detection), virtual or projected keyboards, unobtrusive data gloves, 3D overlay displays beamed into your eyes (using eyeball and head tracking; put on a pair of sunglasses if you want to be able to block out the rest of the world and see only your virtual display). You never have to configure your computer becauuse it is always with you. Plug into a network or a local system, but trust your data integrity by using secure protocols between you and your home data store when you want to access data you're not carrying in the 1-terabyte chunk you carry around with you.
Of course, computers in the home will be completely different as well. The "home server" will be always-on, taking messages, monitoring security, controlling environment, providing wireless and wired network connections, and providing data storage. There probably would be screens and control stations throughout the house (so you don't necessarily have to wear your wearable portable system at home).
No, you recover as much as you can from the current session, roll back one day to before the virus did any damage, recover your data from then, roll it back to before the virus was even there (or disinfect your system now that you know the virus is there and what it does), and restore all your data.
One use for all the excess computational power we have now and will have even more of in the future (think 3GHz to read e-mail, surf the web, and wait for a 5 key/minute typer to enter a shopping list is bad, wait until we're at 10 or 50GHz, 5GB RAM, 1.5TB disk) would be to run everything in virtual, even emulated sandbox environments. Expect Java, for instance, to become more popular as performance issues are blown away by raw horsepower.
Usually, you can get this fixed by calling the right place and being insistent.
I bought something at Best Buy that had three rebates. Two said they required the original UPC, the third didn't require it. So I called the Best Buy rebate center to enquire, and they told me that one of them could, in fact, be a photocopy.
6 weeks later, I got a postcard for that one saying my rebate was invalid. I called, explained what had happened. She said "we can't take photocopies." Insistent. Never. Take. Photocopies. Never never never. Absolutely impossible. Went back and forth, asked for a supervisor, complained about false advertising, she finally said "One moment." 2 minutes later she came back and said "I guess we do take photocopies." then told me to re-submit everything, with the photocopies I had made. Fortunately, I had them.
What's odd is, they always say "photocopy all your materials for proof of what you sent in." So if I just claimed that I did so send in the original UPC, and I have the photocopies to prove it, what are they going to do?
There was another one where they sent me back a rejection, claiming that the purchase date was incorrect. Looking closely at the form, the data entry person had entered the receipt date as the purchase date (which should have been noticed since it was AFTER the postmark date they recorded, and besides how would I get the store-printed rebate form?). That one was actually handled pretty well when I called to complain, and they authorized processing without my having to send anything else in. I think that might have been Western Digital.
I also had one where I got my letter returned, with "P.O. Box closed". I sent it the day after I bought it, which was the last purchase date of the offer, and it said postmark by 1 week later. "Somehow", it took 1 week, 2 days to get there, and they had closed the box by then. Never did get that one resolved.
Doing things by e-mail is almost always impossible. I think they specially train customer-service people to not understand what you said, give you advice you already said you tried, and if you asked more than one question, to only answer the one that they can misunderstand the most. Amazon is really good at all of this, for instance. Their phone customer support used to be wonderful, then they changed and made it almost impossible to talk to them on the phone.
I don't understand how it could be made so large it couldn't be an internal drive. No one is going to go for a new format that is larger than a laserdisc, or even an 8" floppy disk. Both of them could be easily loaded into a desktop machine, or vertically in a tower. Maybe really small towers, iMacs and laptops would have problems, but it wouldn't really stop anyone.
External SCSI CD-ROM drives were quite common for Macs for a while. Maybe people in the PC world didn't notice, as they had no decent external high speed standard.
Apartments above stores are common all over the place, it's not bizarre at all. Often you never notice because you're not looking up. They're in big cities, little cities, rural towns, United States and Europe, and probably everywhere else.
I think the current situation is actually the exact opposite. Too many things are being invented, with too low a threshold of what counts as "invention". This creates an ever-increasing tarpit that makes it more difficult to create true innovation, with (to mix metaphors) large minefields for new products that are found to have infringed on bogus patents by courts that misinterpret what the patent covers, what prior art actually exists, and what "novel" and "obvious" should mean.
"to some extent" is overstating it, I think. The number of unexpired patents right now is much larger than the number of unexpired patents 20 or 50 years ago. Patents are harmful when it is difficult to determine if what you are producing might infringe on a patent. You either spend most of your effort on patent searches, or you risk the uncertainty of being sued later. Patents should help you produce new products - find a solution to what you're trying to do so you can license it for (some amount) less than it would cost you to come up with the same solution. When it is used strictly to enforce a monopoly, preventing others from innovating, it costs more to society than the good a patent is supposed to be doing by encouraging innovation.
In addition, with bogus patents, the public never actually gets the reward it is supposed to receive after the patent expires, as the patent never actually reveals anything that wasn't already known before. Unless the gains to society are greater than the costs, each particular patent shouldn't have been granted. An inventor isn't owed protection; society chooses to grant it for the greater good of all.
The one good thing about all this is that it forces exploration of entirely new fields. Unfortunately, that means that those fields will soon be clogged with overreaching patents which will stifle innovation for 20 years.
In an earlier post, you mentioned RSA as being an example of a good software patent. The problem with that is, while innovative and non-obvious, it was also too broad. It prevented many other schemes which had true inventiveness from being used, and had the effect of killing a lot of development of public key cryptography until the patent ran out. Without the RSA patent, we might be living in a much more secure world, communications-wise (although US munitions export laws didn't help either).