You may be joking... but I think rather than submitting sketches, the user would submit samples of things that already match what they want. For pornography, if a search engine were able to find images similar to those already tagged as "I like this," that would be a really sought-after search engine!
More broadly, if a search engine were able to find similar pictures, then you could narrow down to the result you wanted by submitting images that are close to what you want. For instance you may have found a thumbnail of the image you're looking for, but can't find a full-size version. Or you have a few pictures of airplanes, but want a whole bunch more.
Another way the technology could work is to present you with a series of candidate images, and then you click on the one that is "closest" to what you want. It then performs the search again, showing new candidates, and you click on the closest match again. If the search engine keeps showing you things closely matched to your last few selections, then this iterative process would quickly home-in on images of exactly what you want. In this kind of mode, the search engine could be using visual similarity as well as keywords and tags to figure out what kind of images you're trying to find.
Lastly, the idea of sketching an image might also work--at least for simplistic images. For instance I've often thought that the symbol-picker applet should, instead of listing thousands and thousands of symbols (which font should I look in?), it should have a box where the user can clumsily draw the symbol, and then display close matches (if you draw a circle it would show the degree sign, the letter 'o', the number '0', etc.). If it works, image-similarity technology like this could be a way to find the desired symbol. (It might work for clipart, too.)
This present design is a cool idea. I don't want to take anything away from the presented concept, but I thought it would be important to point out previous work on nanomechanical computers. First of all, Eric Drexler (the guy who popularized the term "nanotechnology" and who basically invented the field now known as molecular nanotechnology) has been advocating the concept of nanomechanical computers for many years now (they are described in his book Engines of Creation (1986) and detailed feasibility calculations, and rough schematics, are presented in his book Nanosystems (1992)). Drexler has been trying to get people on-board with his very foreward-looking ideas for nanotechnology: where nano-sized mechanical systems would be performing computation, and controlling chemical reactions with a precision that currently only biological systems can achieve. (It should be noted that current work in "nanotechnology" is hilariously primitive compared to what Drexler intended the term to describe.) Drexler's vision of nano-mechanical systems has been challenged by many people, most notably by Richard Smalley (the guy who discovered buckyballs).
Beyond Drexler's theoretical work, carbon nanotubes were demonstrated as nano-mechanical transistors in 2000. Basically, the nanotube was positioned over various electric pads. A current could be applied to mechanically deform the nanotube. The deformation was stable, and could be read-out by measuring current across the tube. Since the deformation was stable and reversible, the tubes could be used as persistent storage or as switching/logic elements. In fact, switching speeds of gigahertz were demonstrated. The vision was to have long nanotubes in a huge cross-bar architecture, leading to high-density persistent storage. As is often the case, scale-up was difficult.
This present work appears to pattern a nano-sized post between conducting pads (out of a gold/silicon layered system) , and to use that post as a single-electron transistor. The 'mechanical' part comes from mechanically coupling multiple pillars to use as a gain mechanism for a transistor. This is basically much closer to conventional micro-lithography, and as such, it should fit in with current lithographic infrastructure much more easily than the nanotube concept did.
It's things like this that force me to disable Password Manager altogether.... With that said, I must admit that I am having more trouble remembering all of my passwords since I acquire more accounts and each account has different password requirements.
Well my solution is to be selective about what passwords get saved. Low-priority things like slashdot and forum logins are fine for password managers. However I memorize, never write down, and never save passwords for financial sites. This keeps the number of "must-be-memorized" passwords down to a manageable level.
Password managers are not an "all-or-nothing" tool. Use them where they make sense.
what I would like to see is Firefox switch to this kind of password manager--where the passwords are all encrypted with a "master password."
To clarify (before someone points out my mistake!): I see that Firefox has a "Set Master Password" option in the Security settings. What I should have said was:
what I would like to see is Firefox switch to this kind of password manager--where the passwords are all encrypted with a "master password" in the default configuration.
It seems to me that if this program can do that, then it can't be hard for a more nefarious program on my computer to do the same.
Well, any program running with user rights can probably read the firefox passwords, since they are not hard for a user to obtain. Just go into "Options" > "Security" > "Show Passwords..." > "Show Passwords" and click "Yes" on the confirmation dialog. You'll see all the stored passwords in plaintext. This means that your passwords can be read without trouble. For instance anyone who sits down at your computer can check through those and find out what your passwords are.
This kind of password manager is not very secure... but then again the intention here is convenience. The idea is that on computers you "trust" (you are confident that they are physically secured to only trusted people sitting down at them, and you are confident they are not riddled with spyware) then you give up a little bit of security for a good amount of convenience. Of course it goes without saying that you should not be using the same passwords on these unimportant websites (which get stored in your password manager) and important things in your life (e.g. root password on an important server!).
In KDE, the KWallet application allows you to use a single master-password to store/encrypt all these lower-priority passwords. This is slightly less convenient, but is much more secure. When you start browsing in Konqueror and encounter a password field, KWallet pops up, you enter your master-password and it fills in the password fields for you. You also set a timeout if you want (the wallet "stays open" for some amount of time, so you don't have to re-enter the master password too often). Without the master password you can't unlock (or even decrypt) all the other passwords. With only one password to remember, it's reasonably convenient. Probably there similar apps available for OS X and Windows? In any case what I would like to see is Firefox switch to this kind of password manager--where the passwords are all encrypted with a "master password."
To any Slashdot users (like myself) who were part of the "University of Michigan" testing group, be aware that Discussion2 was recently silently added as an option in the user preferences. Just go "Preferences" > "Comments" > "Slashdot's New Discussion System Testing".
For those of you who don't know what I'm talking about: A random group of Slashdot users were asked to participate in a study conducted by the University of Michigan on a new comment/discussion interface. This study presumably ended at some point (and possibly the data from it helped create Discussion2). However those users who were subjects in this study had a modified preferences panel, which prevented the Slashdot "Discussion2" mode from being available to us (bug report here). Now they have apparently fixed the issue, though no official message was sent to anyone. So, if anyone else was affected by that bug, you can now test the Discussion2 system along with everyone else.
I just love the people who treat Slashdot as a single entity with a single opinion... and then proceed to point out hypocrisy in Slashdot because it holds two seemingly incompatible notions.
Interestingly, they usually describe this as "Groupthink"... which is basically the notion that everyone starts supports the popular opinion. That is, that everyone agrees with each other to be part of the crowd, and suppress dissenting views. The irony, however, is that the very hypocrisy that is being referred to is telling evidence that groupthink is not as prevalent as it is assumed to be.
The fact is that Slashdot users have a variety of backgrounds and opinions. On every issue, there is a distribution of opinions. On some subjects we all seem to agree (e.g. "technology is good"), on others we mostly agree (e.g. "Linux is cool") and on others still there is so much disagreement that you will see completely contradictory and opposing opinions both modded up to +5 (e.g. "global warming is a myth").
Your example, of disliking MS but supporting Wal-Mart, is a total strawman. The general impression I get is that there is a consistent but not universal dislike of Microsoft's business tactics, and that there is solid division of opinion on the Wal-Mart issue. I've seen insightful comments both supporting the good that Wal-Mart does as part of a thriving free market, and insightful comments about the harm that Wal-Mart does as a megacorp that only cares about money. Both sides make good points and the most reasonable stance is probably a nuanced view that takes into account all of these factors. To suggest that Slashdot has a single opinion on these subjects betrays a serious lack of perspective on your part.
Your closing sentence, "I wish I lived in the fantasy world of most Slashdotters", again is deeply rooted in the fantasy that Slashdot is a single entity with a single mind, and that any self-contradictory statements it makes represent its own insanity, rather than diversity of opinion among its constituents.
Strange, I don't remember feeling a satisfying sense of victory when Microsoft tried to undermine our software by claiming that it infringed their patents. I don't remember feeling that we had "won" when Novell signed a deal with Microsoft to protect themselves and leave the rest of the community out in the cold.
The loopholes were just that: sneaky ways to evade the intentions of most of the most important contributors in the realm of FOSS. I have no problem with businesses making money using FOSS, and many of them do it in a way that is not only compatible with the intentions of the GPL, but actively promotes the cause of free software. However, those businesses who were exploiting loopholes in the GPL knew that they were not promoting our interests, and therefore should not be surprised when the community shifts to close those loopholes. Such a shift will only alienate businesses who were not helping "the cause" anyways.
The GPLv3 is not perfect, and is not a perfect license. I don't think that every project should switch to GPLv3... for some the GPLv2 may be a better match. However GPLv3 was crafted to address a very real problem, and judging from Microsoft's reaction, it is doing a great job in that regard.
Well, given that the GPLv3 was written specifically to make those "patent protection deals" untenable, this is a huge success for the GPLv3. Microsoft is essentially admitting that, legally, the GPLv3 does what it intends to do.
So, anyone who was bothered by the MS/Novell deal (and its variants) can and should encourage usage of GPLv3. Coders who want to prevent MS from using patent threats to splinter the community should consider adopting the GPLv3.
Since a certain number of important projects have already switched to GPLv3, this means that within a year or two the MS/Novell deal (and variants) will essentially disappear. As someone who was not happy with those deals in the first place, I say good riddance.
But who is the "next highest bidder"? If you sell your vulnerability to MS and also the black-market, for instance, then you're screwing both of them... and they will notice. MS will notice the vulnerability in the wild, and if it happens repeatedly, they will probably stop trusting you.
If the black-market guys notice that MS came up with a patch surprisingly quickly after you sold them the exploit, they are going to be very angry, because you've very much decreased the value of the exploit. And I would imagine that cheating black-market guys is not a smart thing to do... if they lack ethics in the "break into computers" department they may behave similarly in the "break your legs" department.
So, at the end of the day, how are you going to sell an exploit twice? Few would try, and fewer still would get away with it.
If I make one from a 3-d printer or SLA, then what? That's a Mobius strip with no stresses and equal energy density throughout.
Sure. In principle you can generate an arbitrary shape with an arbitrary internal stress distribution (including no stress distribution).
The paper in question, however, was modeling the minimum-energy state that a Möbius strip would adopt assuming that the local energy on the strip is based on local curvature (and that stretching energies can be neglected). As they point out, this is a very good approximation for building a Möbius strip by bending common thin materials (e.g. a sheet of paper or plastic). Knowing stress distributions is of course important for things like failure mechanics.
They also note that in the field of synthesizing nano-ribbons and nano-Möbius strips (yes, it's been done!), this bending energy can be critical to understanding the behavior of the final object, and is also important in understanding how such objects can be synthesized. (The growth of anisotropic nano-crystals, including nano-ribbons, is strongly dependent on the relative energies of the various growing surfaces.)
Having said all that, I think it's pretty clear that the authors tackled this particular mathematical problem because it was fun, and because of the notoriety of the Möbius strip. Ultimately it's a neat piece of mathematics and makes for some cool-looking graphs.
The consensus here seems to be that, for some reason, Microsoft is afraid of ODF. Does anyone honestly believe that Microsoft is not capable, or believes it is not capable, of delivering the best and most able word processor producing ODF files?
Actually maybe you should direct that question to Microsoft. If Microsoft is confident that they can deliver the best and most capable Office Suite that opens/writes ODF, and the users are asking for ODF, then what's the problem? Why doesn't Microsoft just shut up, release Office 2008 with ODF as the default file format, and let the free market decide what the "best and most able word processor" really is?
The reason, of course, is that while Microsoft Office is a great product, it is also an expensive product. Many people are more than willing to buy it, because they like the interface, or need the features. However, many other people do not need all those features, or are not yet adapted to the interface of MS Office (or at least not MS Office 2007). For those people, a cheaper (or free!) product might be a better fit.
Microsoft is scared to death of the free market. In a fair competition of various products, MS would still make money, but not nearly as much as they do now, where they have the entire market captured due to file-format lock-in. This is what makes Microsoft scared. This is why they are being pulled kicking and screaming into the world of open and standards-compliant file formats.
No one here is arguing that MS could not write a decent ODF word processor. That is a strawman. The fact is that MS doesn't want to write a decent ODF word processor (or even plugin) because that would mean giving up a certain percentage of their devoted (read: captured) user base.
The good news is that if Microsoft is changing their tactics, it means that they are admitting (partial) defeat in their previous attempts. Essentially they have lost the technical argument. Many groups have weighed-in on the subject and agreed that ODF is a more open format, and actually meets the needs of a standard. OXML is not winning that particular competition.
So they have a new tactic. This tactic basically amounts to saying: "Let's just have both standards, and let people pick the one they want. Oh... did we mention that OXML will be the default in all of our products?" Moreover, they are strongly implying that ODF is a lame duck, and that OXML has "more features" and is "richer." They are trying to paint ODF as the poor-man's format, with OXML being the format you use when you're serious.
The bad news is that this tactic will probably work. If OXML is the default format (in the dominant Office suite), people will view it as being the "serious" one and anything else as being "dumb." It doesn't matter that the additional "richness" is a bunch of features that these users will never activate. It also doesn't matter that the additional "richness" won't be maintained cleanly across platforms, during filetype conversions, and possibly even across software version changes. All that matters is building mindshare that truly believes that OXML is "the real deal" and that anything else is "that weird thing that geeks use."
So the counterattack from those of us who would prefer a true standard (such as ODF) to become the default need to use ODF as much as possible, and encourage others to do the same. ODF is the one that guarantees readability into the future, and that guarantees interoperability. We need to make this clear to everyone else.
I like the concept, but what about spam-style advertising? If I scatter viagra ad beacons around the museum, what do you do to see the information that the painting is broadcasting without seeing the information that my ad beacons are broadcasting?
Spam is definitely a worry. Then again, the thing that keeps Spam under control when it comes to physical mail is the inherent cost of sending the mail. Email spam is cheap. Real mail is not.
Similarly, putting a bunch of beacons all over the place is very expensive. Even if each one is a dirt-cheap devices ($10?), it's still a very expensive way to reach a very small audience. Local businesses will be willing to spend the money because they want to draw people in to spend money. Pump-and-dump scammers probably won't bother.
Hacking of regional broadcast beacons is of course possible, but no worse than the danger of a website being hacked. Businesses will of course try to broadcast all kinds of annoying ads and whatnot, but presumably it won't be hard for the end consumer's device to filter this in some way. I mean if you turn on your PDA and see broadcast signals for various things, you'll only look at the ones that interest you. If you're hungry you'll click on the link for "Moe's Restaurant--Great Prices!" but won't bother with the "Best Deals in town on furniture!" links...
TFA overstates its case a little bit, but the basic idea is a good one (but not really revolutionary--just an extension of the Internet as we know it).
I think this kind of thing will happen as wireless net connectivity becomes more widespread and more affordable. Also, the growing number of user-contributed sites are encouraging massive amounts of tagging. Combined, this will create a new way that the net can be useful to all of us. I see data coming from a variety of sources, many of which you mentioned. Consider: 1. User will tag documents, photos, etc. with locations (GPS coordinates, or names of places, or whatever). 2. Devices start automatically encoding coordinates. For example digital cameras have GPS and tag all photos properly. Maybe your laptop will store location information in documents you work on. (So that later when you are searching for a particular revision, you can search for "edited while in Paris"...) 3. Local businesses or organizations broadcast location-specific information. (You can sit in a park and browse the menus of all the nearby restaurants, rather than walking around randomly.) 4. Web-pages (especially businesses) go to the bother of adding meta-data that provides location information. 5. Automated algorithms (e.g. Google) cross-reference data and establish location-relevancy for data that wasn't originally tagged with a location. (e.g. pictures of 'mona lisa' become geo-tagged to the coordinates of the Louvre)
When combined, this would provide a ton of useful data. Any web-search could be geographically limited. Google Earth would have thousands of pictures linked to the proper physical locations (we're already seeing the start of this). So when you're sitting at your desk, you can be planning a trip and search for "italian restaurant" and use a GPS coordinate as a search parameter. You can even move the pushpin around on a map and see the search change.
When using a PDA, it could default to use your current location in the search, providing contextual information. Nearby businesses is an obvious use. Traffic, historical information, blogging commentary, news events... everything could be filtered based on location of the subject area of interest.
I can also imagine turning the tagging information backwards. Rather than searching for "protests" near your current location, you could mine the combined data... for instance generate a map of the earth color-coded by search relevancy to the word "protest." This would give new kinds of information that we have not really thought about to date.
None of this is revolutionary... but I think it is an exciting extension of the net as we know it... and I'm looking forward to many of these applications. (But not looking forward to dealing with novel types of spam.)
You know what makes the video demo in TFA more impressive that the Microsoft Surface demos I've seen?
This demo uses real applications! It's easy for MS (or whoever) to throw together a video of someone using a neat interface. You see all kinds of slick animations of photo-libraries and data being automatically uploaded to cellphones. The problem is it's probably all fake--the visual equivalent of a mockup. Basically they are showing you the way they *hope* it will look at work. If you look at some of the older Vista demos (before it was released) you'll see alot of mockup video that was never realized into actual code.
In this demo, they actually start by using Google Earth and scrolling through webpages. The fact that they are using real applications is much more impressive. It makes me believe that they may have something functional in a reasonable amount of time. It also shows that they are thinking about it as an extensible platform that can run generic software, rather than something locked-down that will only run approved code (i.e. just a really big PDA interface, rather than a novel way to interface with existing computer hardware and software).
It's even more complicated than that. Most of the Creative Commons licenses explicitly forbid adding DRM to the files. (See FAQ here.) So, adding DRM to CC files would be a license violation.
It's unclear who is actually doing the violating, though. If I transfer a file to you, and our devices conspire to add DRM to the file, who is at fault? Is it me? Is it you? (We should have known how the devices operate, and it is our responsibility from ever using them in conjunction with CC files?) Or is the device manufacturer liable? Normally I would say that the device manufacturer cannot be held accountable for copyright violations on the part of the users... however in this case if the user has no way to turn off the 'feature' then the device is not letting them comply with copyright. In effect, the device is enforcing copyright violations, by not allowing users to respect CC licenses, even if they wanted to.
Okay, the logic is a little contorted, but I think you get the point. Devices that mandatorily add DRM are incompatible with a great many legitimate uses, and in conflict with many distribution licenses.
But there is yet another twist: free software projects (as the Linux kernel, for example) have been constantly developed for years, i.e. new code added and removed all the time. How can you realistically sort out the code that fell into the public domain from the GPL'ed code?
That part is fairly easy. If you have a snapshot of the codebase on a particular date, you know all the code was added before that date. So if you download a copy of the code, and wait 14 years (or whatever the term of copyright is), you know all the code in your copy is now public domain. Most version control systems maintain date stamps, so that would be another way to figure out the 'age' of code.
Nowadays such questions are meaningless, because by the time the copyright expires, there won't be any computers around that can even run the code... but if the copyright terms were shorter, no doubt people would pay closer attention to the exact dates various bits of code were written.
Actually I heard Richard Stallman give a talk, and this came up during the question period. He basically said "Yes, if copyright terms were, say, 7 years, then anyone could take gcc and create a closed-source version of it. Of course, they would only be able to do that with a 7-year-old version of gcc! That's an eternity, and I don't see a problem with it." (Heavily paraphrasing from memory.)
Basically, RMS doesn't view this as a problem. The true value of software is in the current version, in the constant community improvements. As long as the most up-to-date version is free, that's what really matters. Having older versions of code in the public domain doesn't restrict the freedoms of the users of the software. Having proprietary forks of ancient code wouldn't be a big deal. (Especially in a world where you could be freely distributing old proprietary binaries.)
Okay, you suggest that the author of this work is massaging the equations and numbers to obtain the result he wants. You also point out some ways in which the paper is flawed: e.g. assuming single values for inputs rather than discussing the range of possibilities and error bars.
I agree with your criticisms, though not with your implication that he is massaging the numbers to get the results he wants. Regardless, I would like to point out something that, I think, is crucial about the approach he has taken: Because he expresses his logic and results in rigorous, mathematical form, it is possible for us to analyze and improve on them rationally.
Most debates in public policy are just rhetoric: trying to convince people by appeals to emotion and "common sense" (or contorted logic). There is no way to improve upon the debate other than to throw your own rhetoric into the mix. Here, we have a mathematical analysis. If you think there are flaws in the math, you can easily point them out. If you think he should have done an error analysis, you can do this error analysis. If you think graphing the range of possibilities is more fair, you can go ahead and do that. His work can be built upon, objectively criticized, and improved. This is less like rhetoric and now more like science.
So, far from being the "final word" on the optimal length on copyright, I view this as a step towards logical analysis (finally!). I hope that others pick upon on this work and come up with more reliable input numbers. That's how progress is made.
Yes, statistics can be contorted to "prove" alot of things. But the more rigorously and mathematically you frame your argument, the easier it is to point out mistakes and fallacies. I think this is a step in the right direction for this debate.
Having said all that, I will have to read it a few more times to determine whether I agree with the logic and math. However I think it would be premature to dismiss this without due consideration.
The paper is clearly not a paper about mathematics. No 'new math' is being invented. But to say "there is no mathematics in the manuscript, only economics" is not at all right. The paper has plenty of math in it, used to analyze an economics question. That is like opening up a physics journal, looking at all the equations, and concluding "there is no mathematics in these manuscripts, only physics". Physics requires mathematics. Economics requires mathematics.
no theorems are formulasted or proven
Well, actually the paper has 13 theorems presented and proved. Again, these are not pure-mathematical theorems, they are economic theorems being proved using mathematical techniques.
closer to accounting than to real mathematics
I'm not sure what you mean by "real math." Accounting uses "real math." Engineering uses "real math." Analyzing the economics of copyright using rigorous equations and logical mathematical arguments is, in my estimation, "real math." He is using math as a tool, yes, but that doesn't make it "fake math." Moreover I have trouble believing that accounting typically involves setting out abstract theorems and proving them.
The next release will be interesting to see. Being a LTS version...
Small correction: The next release will be 7.10 (Gusty Gibbon, October 2007). However the next "Long Term Support" (LTS) release, according to this page, will be "Feisty+2", or the release after Gusty. This release will probably be in April 2008.
I agree with everything you said, however. I use the LTS edition for servers that need to be stable, and use the latest version for desktops. The Long Term Support is long enough that you can be confident with it (and easily upgrade to the next LTS when it comes along). Upgrading Ubuntu (e.g. from Edgy to Feisty) has always been painless in my experience. (Yes, YMMV.)
I'm very pleased with the speed (and predictability) of the Ubuntu release schedule, and with the quality of what gets put out.
Businesses actively work to prevent other sites from scraping content. They certainly aren't going to spend extra effort to support it!
True enough. But one of the main points of "Web 2.0" is user-generated content and participatory media. Although businesses make contributions to the usefulness of the web, user-generated content is becoming more and more useful and powerful. Just look at the impact of Wikipedia, web-forums, free software, creative commons, etc.
These user-driven efforts are where the tagging and semantic web will probably start. If Wikipedia contributors care to take the time to write good articles, then surely they will also be willing to semantically tag articles. (In fact Wikipedia already has alot of semantic tagging.) Similarly creative commons artists are actively tagging their works with machine-readable creative-commons tags. Social sites like Flickr are also doing alot of useful tagging.
So businesses may resist it... but as long as users care about it (and are given easy to tools to make it happen--like wikis), then this semantic web can be created. Once it expands, businesses will have to play along or risk being left behind and ignored by the web-users who come to depend upon the power of the Semantic Web. So, whether they like it or not, businesses will have to connect to the semantic web and add to its descriptive power, or else they will lose all their customers.
And, yes, I'm keenly aware of the flip-side, which is that businesses will then try to commoditize and monetize these technologies, sometimes in bad ways, like Spam. It will be interesting to see how it plays out. But I don't think businesses will be able to stop it.
Users care about presentation. Looks are everything.
I disagree. Or rather, I think that describes only some users. There are plenty of users who are care about content. (Wikipedia and free software are examples of the resultant projects.) So even if many (or most) users don't care about the semantic web, as long as some dedicated group does care, then it will expand and everyone (including users who don't care about the underlying implementation details) will benefit.
As a tangent to your thoughts on the subject... I think one of the things that I (and others) worry about when it comes to "the law" (police, etc.) using technology to make their jobs more efficient is that there is never a "restoring force" that modifies the laws along with.
Allow me to explain. The current laws, like it or not, are not entirely idealistic. They were written within a certain social and technological environment. Using technology to more perfectly enforce a law can turn a reasonable law into an unreasonable one.
A stereotypical example is speeding. Most reasonable people agree that there should be speed limits. The current speed limits, however, were in some sense set with the knowledge that people would "cheat a little bit," so the posted limit turns out to be below the limit most safe drivers actually drive at. This works out okay in the end. The cops stop the people who are speeding alot but tend not to bother with people that speed by 10% or whatever. However if you use technology to enforce this law perfectly, it becomes unfair in a hurry. Or, if you use technology to perfectly enforce a law like "stopping at a stop-sign" then the law becomes unfair (remember that your bumper is supposed to be behind some arbitrarily line and you must be stopped for X seconds, etc.). Even the safest of drivers will not follow these rules to the letter; nor should they: the laws are written with very little leniency in their wording because they are meant to be used to stop people from egregious abuses of the law. They were never meant to punish everyone for doing normal daily things.
Another example would be copyright. I don't want to get into this debate too deeply, since it is a "hot topic" on Slashdot. Suffice it to say that many aspects of copyright seem reasonable enough, but when copyright is enforced perfectly, or worse when technology makes compliance mandatory (e.g. DRM) then a reasonable law gets transformed into an unreasonable law in a hurry. Many of the "well obviously *this* should be allowed" things that were not formally written into the law disappear.
Laws that make the everyday, normal activities of socially-responsible people illegal are not good laws. So the problem is that if law enforcement uses new technologies to allow them to do their jobs "more efficiently" but there is no corresponding rewriting of laws (to make them *more lax* or even repeal them), then our society will tend towards being less free.
That is one of the worries. So the solution is either to limit the implementation of technologies by law enforcement in some cases, or to have the laws modified. (Or a combination.)
Any technology that COULD count successfully all the Linux boxes out there would be a bit scary - many people probably don't WANT anyone to be able to know what they are running.
Well there is a very old and very well-tested "technology" that could determine the number of Linux users, and all without invading privacy, or installing software on people's computers. It's called: "statistical surveys."
Yes, surveys are imperfect. They have error bars. However if the sample size is big enough, they give a reasonably correct result. Moreover the error bars can be estimated based on sample size and analyzing the stats. You don't have to count every single Linux box to get a statistically significant result. So, really, it would be quite easy to get a reasonable estimate of the number of Linux users: do various rigorous surveys (by "rigorous" I mean truly randomized studies, conducted by trusted sources, not a voluntary web-poll), and check the results.
Obtaining this information is not impossible. It would, however, cost some money, and to date no one has felt it worth the effort to spend the money to find out. But it would certainly be possible to do.
(P.S.: Does anyone have a rough idea how much it would cost to fund such a survey? Some of us are keen to be able to go to hardware vendors and say "if you do not support this OS you are ignoring X % of the computer market", using accurate stats. How much donated money would be required to fund such a survey? Perhaps some university could be coaxed into carrying out a national survey?)
Slashdot Mirror
You may be joking... but I think rather than submitting sketches, the user would submit samples of things that already match what they want. For pornography, if a search engine were able to find images similar to those already tagged as "I like this," that would be a really sought-after search engine!
More broadly, if a search engine were able to find similar pictures, then you could narrow down to the result you wanted by submitting images that are close to what you want. For instance you may have found a thumbnail of the image you're looking for, but can't find a full-size version. Or you have a few pictures of airplanes, but want a whole bunch more.
Another way the technology could work is to present you with a series of candidate images, and then you click on the one that is "closest" to what you want. It then performs the search again, showing new candidates, and you click on the closest match again. If the search engine keeps showing you things closely matched to your last few selections, then this iterative process would quickly home-in on images of exactly what you want. In this kind of mode, the search engine could be using visual similarity as well as keywords and tags to figure out what kind of images you're trying to find.
Lastly, the idea of sketching an image might also work--at least for simplistic images. For instance I've often thought that the symbol-picker applet should, instead of listing thousands and thousands of symbols (which font should I look in?), it should have a box where the user can clumsily draw the symbol, and then display close matches (if you draw a circle it would show the degree sign, the letter 'o', the number '0', etc.). If it works, image-similarity technology like this could be a way to find the desired symbol. (It might work for clipart, too.)
This present design is a cool idea. I don't want to take anything away from the presented concept, but I thought it would be important to point out previous work on nanomechanical computers. First of all, Eric Drexler (the guy who popularized the term "nanotechnology" and who basically invented the field now known as molecular nanotechnology) has been advocating the concept of nanomechanical computers for many years now (they are described in his book Engines of Creation (1986) and detailed feasibility calculations, and rough schematics, are presented in his book Nanosystems (1992)). Drexler has been trying to get people on-board with his very foreward-looking ideas for nanotechnology: where nano-sized mechanical systems would be performing computation, and controlling chemical reactions with a precision that currently only biological systems can achieve. (It should be noted that current work in "nanotechnology" is hilariously primitive compared to what Drexler intended the term to describe.) Drexler's vision of nano-mechanical systems has been challenged by many people, most notably by Richard Smalley (the guy who discovered buckyballs).
Beyond Drexler's theoretical work, carbon nanotubes were demonstrated as nano-mechanical transistors in 2000. Basically, the nanotube was positioned over various electric pads. A current could be applied to mechanically deform the nanotube. The deformation was stable, and could be read-out by measuring current across the tube. Since the deformation was stable and reversible, the tubes could be used as persistent storage or as switching/logic elements. In fact, switching speeds of gigahertz were demonstrated. The vision was to have long nanotubes in a huge cross-bar architecture, leading to high-density persistent storage. As is often the case, scale-up was difficult.
This present work appears to pattern a nano-sized post between conducting pads (out of a gold/silicon layered system) , and to use that post as a single-electron transistor. The 'mechanical' part comes from mechanically coupling multiple pillars to use as a gain mechanism for a transistor. This is basically much closer to conventional micro-lithography, and as such, it should fit in with current lithographic infrastructure much more easily than the nanotube concept did.
Password managers are not an "all-or-nothing" tool. Use them where they make sense.
This kind of password manager is not very secure... but then again the intention here is convenience. The idea is that on computers you "trust" (you are confident that they are physically secured to only trusted people sitting down at them, and you are confident they are not riddled with spyware) then you give up a little bit of security for a good amount of convenience. Of course it goes without saying that you should not be using the same passwords on these unimportant websites (which get stored in your password manager) and important things in your life (e.g. root password on an important server!).
In KDE, the KWallet application allows you to use a single master-password to store/encrypt all these lower-priority passwords. This is slightly less convenient, but is much more secure. When you start browsing in Konqueror and encounter a password field, KWallet pops up, you enter your master-password and it fills in the password fields for you. You also set a timeout if you want (the wallet "stays open" for some amount of time, so you don't have to re-enter the master password too often). Without the master password you can't unlock (or even decrypt) all the other passwords. With only one password to remember, it's reasonably convenient. Probably there similar apps available for OS X and Windows? In any case what I would like to see is Firefox switch to this kind of password manager--where the passwords are all encrypted with a "master password."
To any Slashdot users (like myself) who were part of the "University of Michigan" testing group, be aware that Discussion2 was recently silently added as an option in the user preferences. Just go "Preferences" > "Comments" > "Slashdot's New Discussion System Testing".
For those of you who don't know what I'm talking about: A random group of Slashdot users were asked to participate in a study conducted by the University of Michigan on a new comment/discussion interface. This study presumably ended at some point (and possibly the data from it helped create Discussion2). However those users who were subjects in this study had a modified preferences panel, which prevented the Slashdot "Discussion2" mode from being available to us (bug report here). Now they have apparently fixed the issue, though no official message was sent to anyone. So, if anyone else was affected by that bug, you can now test the Discussion2 system along with everyone else.
Thanks to whoever coded the workaround.
I just love the people who treat Slashdot as a single entity with a single opinion... and then proceed to point out hypocrisy in Slashdot because it holds two seemingly incompatible notions.
... which is basically the notion that everyone starts supports the popular opinion. That is, that everyone agrees with each other to be part of the crowd, and suppress dissenting views. The irony, however, is that the very hypocrisy that is being referred to is telling evidence that groupthink is not as prevalent as it is assumed to be.
Interestingly, they usually describe this as "Groupthink"
The fact is that Slashdot users have a variety of backgrounds and opinions. On every issue, there is a distribution of opinions. On some subjects we all seem to agree (e.g. "technology is good"), on others we mostly agree (e.g. "Linux is cool") and on others still there is so much disagreement that you will see completely contradictory and opposing opinions both modded up to +5 (e.g. "global warming is a myth").
Your example, of disliking MS but supporting Wal-Mart, is a total strawman. The general impression I get is that there is a consistent but not universal dislike of Microsoft's business tactics, and that there is solid division of opinion on the Wal-Mart issue. I've seen insightful comments both supporting the good that Wal-Mart does as part of a thriving free market, and insightful comments about the harm that Wal-Mart does as a megacorp that only cares about money. Both sides make good points and the most reasonable stance is probably a nuanced view that takes into account all of these factors. To suggest that Slashdot has a single opinion on these subjects betrays a serious lack of perspective on your part.
Your closing sentence, "I wish I lived in the fantasy world of most Slashdotters", again is deeply rooted in the fantasy that Slashdot is a single entity with a single mind, and that any self-contradictory statements it makes represent its own insanity, rather than diversity of opinion among its constituents.
The loopholes were just that: sneaky ways to evade the intentions of most of the most important contributors in the realm of FOSS. I have no problem with businesses making money using FOSS, and many of them do it in a way that is not only compatible with the intentions of the GPL, but actively promotes the cause of free software. However, those businesses who were exploiting loopholes in the GPL knew that they were not promoting our interests, and therefore should not be surprised when the community shifts to close those loopholes. Such a shift will only alienate businesses who were not helping "the cause" anyways.
The GPLv3 is not perfect, and is not a perfect license. I don't think that every project should switch to GPLv3... for some the GPLv2 may be a better match. However GPLv3 was crafted to address a very real problem, and judging from Microsoft's reaction, it is doing a great job in that regard.
Well, given that the GPLv3 was written specifically to make those "patent protection deals" untenable, this is a huge success for the GPLv3. Microsoft is essentially admitting that, legally, the GPLv3 does what it intends to do.
So, anyone who was bothered by the MS/Novell deal (and its variants) can and should encourage usage of GPLv3. Coders who want to prevent MS from using patent threats to splinter the community should consider adopting the GPLv3.
Since a certain number of important projects have already switched to GPLv3, this means that within a year or two the MS/Novell deal (and variants) will essentially disappear. As someone who was not happy with those deals in the first place, I say good riddance.
But who is the "next highest bidder"? If you sell your vulnerability to MS and also the black-market, for instance, then you're screwing both of them... and they will notice. MS will notice the vulnerability in the wild, and if it happens repeatedly, they will probably stop trusting you.
If the black-market guys notice that MS came up with a patch surprisingly quickly after you sold them the exploit, they are going to be very angry, because you've very much decreased the value of the exploit. And I would imagine that cheating black-market guys is not a smart thing to do... if they lack ethics in the "break into computers" department they may behave similarly in the "break your legs" department.
So, at the end of the day, how are you going to sell an exploit twice? Few would try, and fewer still would get away with it.
The paper in question, however, was modeling the minimum-energy state that a Möbius strip would adopt assuming that the local energy on the strip is based on local curvature (and that stretching energies can be neglected). As they point out, this is a very good approximation for building a Möbius strip by bending common thin materials (e.g. a sheet of paper or plastic). Knowing stress distributions is of course important for things like failure mechanics.
They also note that in the field of synthesizing nano-ribbons and nano-Möbius strips (yes, it's been done!), this bending energy can be critical to understanding the behavior of the final object, and is also important in understanding how such objects can be synthesized. (The growth of anisotropic nano-crystals, including nano-ribbons, is strongly dependent on the relative energies of the various growing surfaces.)
Having said all that, I think it's pretty clear that the authors tackled this particular mathematical problem because it was fun, and because of the notoriety of the Möbius strip. Ultimately it's a neat piece of mathematics and makes for some cool-looking graphs.
The reason, of course, is that while Microsoft Office is a great product, it is also an expensive product. Many people are more than willing to buy it, because they like the interface, or need the features. However, many other people do not need all those features, or are not yet adapted to the interface of MS Office (or at least not MS Office 2007). For those people, a cheaper (or free!) product might be a better fit.
Microsoft is scared to death of the free market. In a fair competition of various products, MS would still make money, but not nearly as much as they do now, where they have the entire market captured due to file-format lock-in. This is what makes Microsoft scared. This is why they are being pulled kicking and screaming into the world of open and standards-compliant file formats.
No one here is arguing that MS could not write a decent ODF word processor. That is a strawman. The fact is that MS doesn't want to write a decent ODF word processor (or even plugin) because that would mean giving up a certain percentage of their devoted (read: captured) user base.
The good news is that if Microsoft is changing their tactics, it means that they are admitting (partial) defeat in their previous attempts. Essentially they have lost the technical argument. Many groups have weighed-in on the subject and agreed that ODF is a more open format, and actually meets the needs of a standard. OXML is not winning that particular competition.
So they have a new tactic. This tactic basically amounts to saying: "Let's just have both standards, and let people pick the one they want. Oh... did we mention that OXML will be the default in all of our products?" Moreover, they are strongly implying that ODF is a lame duck, and that OXML has "more features" and is "richer." They are trying to paint ODF as the poor-man's format, with OXML being the format you use when you're serious.
The bad news is that this tactic will probably work. If OXML is the default format (in the dominant Office suite), people will view it as being the "serious" one and anything else as being "dumb." It doesn't matter that the additional "richness" is a bunch of features that these users will never activate. It also doesn't matter that the additional "richness" won't be maintained cleanly across platforms, during filetype conversions, and possibly even across software version changes. All that matters is building mindshare that truly believes that OXML is "the real deal" and that anything else is "that weird thing that geeks use."
So the counterattack from those of us who would prefer a true standard (such as ODF) to become the default need to use ODF as much as possible, and encourage others to do the same. ODF is the one that guarantees readability into the future, and that guarantees interoperability. We need to make this clear to everyone else.
Similarly, putting a bunch of beacons all over the place is very expensive. Even if each one is a dirt-cheap devices ($10?), it's still a very expensive way to reach a very small audience. Local businesses will be willing to spend the money because they want to draw people in to spend money. Pump-and-dump scammers probably won't bother.
Hacking of regional broadcast beacons is of course possible, but no worse than the danger of a website being hacked. Businesses will of course try to broadcast all kinds of annoying ads and whatnot, but presumably it won't be hard for the end consumer's device to filter this in some way. I mean if you turn on your PDA and see broadcast signals for various things, you'll only look at the ones that interest you. If you're hungry you'll click on the link for "Moe's Restaurant--Great Prices!" but won't bother with the "Best Deals in town on furniture!" links...
Why not all three?
TFA overstates its case a little bit, but the basic idea is a good one (but not really revolutionary--just an extension of the Internet as we know it).
I think this kind of thing will happen as wireless net connectivity becomes more widespread and more affordable. Also, the growing number of user-contributed sites are encouraging massive amounts of tagging. Combined, this will create a new way that the net can be useful to all of us. I see data coming from a variety of sources, many of which you mentioned. Consider:
1. User will tag documents, photos, etc. with locations (GPS coordinates, or names of places, or whatever).
2. Devices start automatically encoding coordinates. For example digital cameras have GPS and tag all photos properly. Maybe your laptop will store location information in documents you work on. (So that later when you are searching for a particular revision, you can search for "edited while in Paris"...)
3. Local businesses or organizations broadcast location-specific information. (You can sit in a park and browse the menus of all the nearby restaurants, rather than walking around randomly.)
4. Web-pages (especially businesses) go to the bother of adding meta-data that provides location information.
5. Automated algorithms (e.g. Google) cross-reference data and establish location-relevancy for data that wasn't originally tagged with a location. (e.g. pictures of 'mona lisa' become geo-tagged to the coordinates of the Louvre)
When combined, this would provide a ton of useful data. Any web-search could be geographically limited. Google Earth would have thousands of pictures linked to the proper physical locations (we're already seeing the start of this). So when you're sitting at your desk, you can be planning a trip and search for "italian restaurant" and use a GPS coordinate as a search parameter. You can even move the pushpin around on a map and see the search change.
When using a PDA, it could default to use your current location in the search, providing contextual information. Nearby businesses is an obvious use. Traffic, historical information, blogging commentary, news events... everything could be filtered based on location of the subject area of interest.
I can also imagine turning the tagging information backwards. Rather than searching for "protests" near your current location, you could mine the combined data... for instance generate a map of the earth color-coded by search relevancy to the word "protest." This would give new kinds of information that we have not really thought about to date.
None of this is revolutionary... but I think it is an exciting extension of the net as we know it... and I'm looking forward to many of these applications. (But not looking forward to dealing with novel types of spam.)
You know what makes the video demo in TFA more impressive that the Microsoft Surface demos I've seen?
This demo uses real applications! It's easy for MS (or whoever) to throw together a video of someone using a neat interface. You see all kinds of slick animations of photo-libraries and data being automatically uploaded to cellphones. The problem is it's probably all fake--the visual equivalent of a mockup. Basically they are showing you the way they *hope* it will look at work. If you look at some of the older Vista demos (before it was released) you'll see alot of mockup video that was never realized into actual code.
In this demo, they actually start by using Google Earth and scrolling through webpages. The fact that they are using real applications is much more impressive. It makes me believe that they may have something functional in a reasonable amount of time. It also shows that they are thinking about it as an extensible platform that can run generic software, rather than something locked-down that will only run approved code (i.e. just a really big PDA interface, rather than a novel way to interface with existing computer hardware and software).
It's even more complicated than that. Most of the Creative Commons licenses explicitly forbid adding DRM to the files. (See FAQ here.) So, adding DRM to CC files would be a license violation.
It's unclear who is actually doing the violating, though. If I transfer a file to you, and our devices conspire to add DRM to the file, who is at fault? Is it me? Is it you? (We should have known how the devices operate, and it is our responsibility from ever using them in conjunction with CC files?) Or is the device manufacturer liable? Normally I would say that the device manufacturer cannot be held accountable for copyright violations on the part of the users... however in this case if the user has no way to turn off the 'feature' then the device is not letting them comply with copyright. In effect, the device is enforcing copyright violations, by not allowing users to respect CC licenses, even if they wanted to.
Okay, the logic is a little contorted, but I think you get the point. Devices that mandatorily add DRM are incompatible with a great many legitimate uses, and in conflict with many distribution licenses.
Nowadays such questions are meaningless, because by the time the copyright expires, there won't be any computers around that can even run the code... but if the copyright terms were shorter, no doubt people would pay closer attention to the exact dates various bits of code were written.
Actually I heard Richard Stallman give a talk, and this came up during the question period. He basically said "Yes, if copyright terms were, say, 7 years, then anyone could take gcc and create a closed-source version of it. Of course, they would only be able to do that with a 7-year-old version of gcc! That's an eternity, and I don't see a problem with it." (Heavily paraphrasing from memory.)
Basically, RMS doesn't view this as a problem. The true value of software is in the current version, in the constant community improvements. As long as the most up-to-date version is free, that's what really matters. Having older versions of code in the public domain doesn't restrict the freedoms of the users of the software. Having proprietary forks of ancient code wouldn't be a big deal. (Especially in a world where you could be freely distributing old proprietary binaries.)
Okay, you suggest that the author of this work is massaging the equations and numbers to obtain the result he wants. You also point out some ways in which the paper is flawed: e.g. assuming single values for inputs rather than discussing the range of possibilities and error bars.
I agree with your criticisms, though not with your implication that he is massaging the numbers to get the results he wants. Regardless, I would like to point out something that, I think, is crucial about the approach he has taken: Because he expresses his logic and results in rigorous, mathematical form, it is possible for us to analyze and improve on them rationally.
Most debates in public policy are just rhetoric: trying to convince people by appeals to emotion and "common sense" (or contorted logic). There is no way to improve upon the debate other than to throw your own rhetoric into the mix. Here, we have a mathematical analysis. If you think there are flaws in the math, you can easily point them out. If you think he should have done an error analysis, you can do this error analysis. If you think graphing the range of possibilities is more fair, you can go ahead and do that. His work can be built upon, objectively criticized, and improved. This is less like rhetoric and now more like science.
So, far from being the "final word" on the optimal length on copyright, I view this as a step towards logical analysis (finally!). I hope that others pick upon on this work and come up with more reliable input numbers. That's how progress is made.
Yes, statistics can be contorted to "prove" alot of things. But the more rigorously and mathematically you frame your argument, the easier it is to point out mistakes and fallacies. I think this is a step in the right direction for this debate.
Having said all that, I will have to read it a few more times to determine whether I agree with the logic and math. However I think it would be premature to dismiss this without due consideration.
The paper is clearly not a paper about mathematics. No 'new math' is being invented. But to say "there is no mathematics in the manuscript, only economics" is not at all right. The paper has plenty of math in it, used to analyze an economics question. That is like opening up a physics journal, looking at all the equations, and concluding "there is no mathematics in these manuscripts, only physics". Physics requires mathematics. Economics requires mathematics. Well, actually the paper has 13 theorems presented and proved. Again, these are not pure-mathematical theorems, they are economic theorems being proved using mathematical techniques. I'm not sure what you mean by "real math." Accounting uses "real math." Engineering uses "real math." Analyzing the economics of copyright using rigorous equations and logical mathematical arguments is, in my estimation, "real math." He is using math as a tool, yes, but that doesn't make it "fake math." Moreover I have trouble believing that accounting typically involves setting out abstract theorems and proving them.
I agree with everything you said, however. I use the LTS edition for servers that need to be stable, and use the latest version for desktops. The Long Term Support is long enough that you can be confident with it (and easily upgrade to the next LTS when it comes along). Upgrading Ubuntu (e.g. from Edgy to Feisty) has always been painless in my experience. (Yes, YMMV.)
I'm very pleased with the speed (and predictability) of the Ubuntu release schedule, and with the quality of what gets put out.
These user-driven efforts are where the tagging and semantic web will probably start. If Wikipedia contributors care to take the time to write good articles, then surely they will also be willing to semantically tag articles. (In fact Wikipedia already has alot of semantic tagging.) Similarly creative commons artists are actively tagging their works with machine-readable creative-commons tags. Social sites like Flickr are also doing alot of useful tagging.
So businesses may resist it... but as long as users care about it (and are given easy to tools to make it happen--like wikis), then this semantic web can be created. Once it expands, businesses will have to play along or risk being left behind and ignored by the web-users who come to depend upon the power of the Semantic Web. So, whether they like it or not, businesses will have to connect to the semantic web and add to its descriptive power, or else they will lose all their customers.
And, yes, I'm keenly aware of the flip-side, which is that businesses will then try to commoditize and monetize these technologies, sometimes in bad ways, like Spam. It will be interesting to see how it plays out. But I don't think businesses will be able to stop it.
I disagree. Or rather, I think that describes only some users. There are plenty of users who are care about content. (Wikipedia and free software are examples of the resultant projects.) So even if many (or most) users don't care about the semantic web, as long as some dedicated group does care, then it will expand and everyone (including users who don't care about the underlying implementation details) will benefit.
As a tangent to your thoughts on the subject... I think one of the things that I (and others) worry about when it comes to "the law" (police, etc.) using technology to make their jobs more efficient is that there is never a "restoring force" that modifies the laws along with.
Allow me to explain. The current laws, like it or not, are not entirely idealistic. They were written within a certain social and technological environment. Using technology to more perfectly enforce a law can turn a reasonable law into an unreasonable one.
A stereotypical example is speeding. Most reasonable people agree that there should be speed limits. The current speed limits, however, were in some sense set with the knowledge that people would "cheat a little bit," so the posted limit turns out to be below the limit most safe drivers actually drive at. This works out okay in the end. The cops stop the people who are speeding alot but tend not to bother with people that speed by 10% or whatever. However if you use technology to enforce this law perfectly, it becomes unfair in a hurry. Or, if you use technology to perfectly enforce a law like "stopping at a stop-sign" then the law becomes unfair (remember that your bumper is supposed to be behind some arbitrarily line and you must be stopped for X seconds, etc.). Even the safest of drivers will not follow these rules to the letter; nor should they: the laws are written with very little leniency in their wording because they are meant to be used to stop people from egregious abuses of the law. They were never meant to punish everyone for doing normal daily things.
Another example would be copyright. I don't want to get into this debate too deeply, since it is a "hot topic" on Slashdot. Suffice it to say that many aspects of copyright seem reasonable enough, but when copyright is enforced perfectly, or worse when technology makes compliance mandatory (e.g. DRM) then a reasonable law gets transformed into an unreasonable law in a hurry. Many of the "well obviously *this* should be allowed" things that were not formally written into the law disappear.
Laws that make the everyday, normal activities of socially-responsible people illegal are not good laws. So the problem is that if law enforcement uses new technologies to allow them to do their jobs "more efficiently" but there is no corresponding rewriting of laws (to make them *more lax* or even repeal them), then our society will tend towards being less free.
That is one of the worries. So the solution is either to limit the implementation of technologies by law enforcement in some cases, or to have the laws modified. (Or a combination.)
Yes, surveys are imperfect. They have error bars. However if the sample size is big enough, they give a reasonably correct result. Moreover the error bars can be estimated based on sample size and analyzing the stats. You don't have to count every single Linux box to get a statistically significant result. So, really, it would be quite easy to get a reasonable estimate of the number of Linux users: do various rigorous surveys (by "rigorous" I mean truly randomized studies, conducted by trusted sources, not a voluntary web-poll), and check the results.
Obtaining this information is not impossible. It would, however, cost some money, and to date no one has felt it worth the effort to spend the money to find out. But it would certainly be possible to do.
(P.S.: Does anyone have a rough idea how much it would cost to fund such a survey? Some of us are keen to be able to go to hardware vendors and say "if you do not support this OS you are ignoring X % of the computer market", using accurate stats. How much donated money would be required to fund such a survey? Perhaps some university could be coaxed into carrying out a national survey?)