Agreed: Windows, Mac OS X, and Linux/*BSD are all broken in that they allow the user to install malware -- software that does stuff no user would ever want. (Some other operating systems like gaming consoles or the iPhone's OS "fix" this by only allowing vendor approved apps.) A better solution would be default limited rights for apps. For example, if a user actually wants to make 100 connections per minute on port 25, they probably know what they are doing. Similarly for a key logger. None of Windows, OS X, and *nix make such limitations easy, if they are even possible. This should be considered a flaw in their UI or the OS itself, not the user.
Another major problem is the lack of ubiquitous undo for everything (even uninstalling applications, deleting system files, changing system settings). Windows tries to approximate this with System Restore, but I have never seen that do any good (but maybe Windows has gotten better in this area since I last used it). OS X has Time Machine, which is a real step in the right direction. Effectively unlimited trustable undo would mean the user would not be scared of messing things up because they would also feel safe that they could undo their changes.
Making operating systems robust but usable and discoverable for all users is no minor task, but it should be a goal.
Although it is tempting to say that in a perfect world without greed much more would be accomplished, one has to consider what the incentives are. In theory, capitalism encourages innovation by providing rewards for it. In a system where no one has to work, many won't. Of course, it is unclear this is a bad thing: many people unable to do what they want in a capitalistic system may decide to be artists or scientists if they do not have to worry about making money.
Just looking at the computer field, it seems like things would be better for everyone if all the software around were available for free, but it is not clear it would have been written in the first place. It is sometimes argued that a lot of the popular Linux programs and desktop environments do a lot of copying from Windows/Mac. If there were not people needing to make a living working on these UI projects, would they have ever gotten done? Microsoft and Apple had strong incentives to bring the personal computer to the masses. Would an alternative economic system offer the same incentives? Should it?
As for alternative systems, you may be interested to read about Social Credit which I first saw discussed in Robert Heinlein's series of essays^W^W^Wutopian novel For Us, The Living. Another alternative system is discussed in Marshall Brain's work Manna (available online) where abundance (and robot laborers) organized by a company founded for social good creates a utopia where each person gets a daily allowance of production credits.
On the more realistic and immediately implementable end of things, there's Why Work? ("CLAWS: Creating Livable Alternatives to Wage Slavery"), a website dedicated to not having a job [that you don't like].
It's a protocol which I believe is based on XMPP (Jabber) and makes heavy use of HTML5, so any client would have to pretty much embed a web browser in order to view messages. Google made a web client first because (1) it's Google, they like coding for web and (2) it is a quick and easy way to make an app cross-platform. If Google does not make a native client for it (like they made Google Talk), then other projects will because a lot of people (including myself) agree with you that the web is the wrong place for an IM client, especially seeing as persistence is part of the protocol so there is no real advantage to running the protocol on someone else's server as the protocol should effectively sync all the clients you use automatically. Most likely, once it is finalized, there will be multiple open source projects to create native clients.
Not sure about the Washington Post, but the NY Times is listening: they made an Adobe AIR app called the TimesReader which has the articles well formatted without ads which is available to print subscribers at no extra charge or for $14.95/month. The price is a bit higher than you are discussing, but it is certainly a step in the right direction.
Correct. Have you really never used a [memory] safe programming language? The only Java debugger I have used is the one in Eclipse which displays objects in a tree with the contents listed by field name (and uses their.toString() methods to get a String summary to display, but for many objects that is useless). Java does not really deal with memory layouts of its objects.
(Or at least trying to: at time of writing, the text link labeled "www.ItsBetterwithWindows.com" is a link to "file:///C:/DOCUME%7E1/jross/LOCALS%7E1/Temp/Temporary%20Directory%202%20for%201000HE%20Black%20New.zip/1000HE%20Black%20New/www.itsbetterwithwindows.com". That seems massively unprofessional for Amazon, not sure what is going on there. The image link is correct.)
Yes, DNS can be fixed. The basic idea is to let ICANN just be one authority among many. Put the entire current DNS under.icann (and default to appending.icann so you don't break stuff, I guess) and let anyone else run their own DNS hierarchy setting up competition in the area of properly assigning domain names (for however your users define "properly" here). The result would most likely be a Wikipedia-like distributed oversight system for who controls which domain names, hopefully with no cost for "registering" a domain which seems pretty silly.
The linked essays explain and argue for it much better than I can. In the end, the proposed system makes the root more or less powerless so it would no longer really matter who controlled it.
Why is everyone going on and on about the screen being too small or low resolution or having bad colors? He suggested including HDMI-out so it could be hooked up to a normal screen for tasks which require one.
Thanks for the link. They already have a proportional representation system which is significantly better than the US system for avoiding a two party system. It is a bit disappointing that the voters would consider STV too complicated, but it does sound like it significantly changes the political landscape, especially in that it sounds like it changes the focus of voting from parties to individuals which may make the elected individuals less predictable as they would not necessarily be kept in line by their party.
The regional lines that legislative positions are based on in the US mean that PR is not really an option. IRV is basically the same system, but it does become slightly simpler when counting for only one position.
Well, very simply, you could use range voting which is not limited by Arrow's impossibility theorem as it does not satisfy the hypotheses.
In reality, pretty much any voting system other than the current first past the post would be a massive improvement due to strategic voting. All voting systems have problems with it, but under plurality voting, it essentially forces a two-party system. By allowing voters to rank their preferences, votes would come much closer to revealing the true preferences of the electorate, and therefore the government would be closer to what the people really want (or at the very least, the politicians would be claiming to support what the people really want).
In the current system, the major parties are able to shape the debate and effectively silence any opposing viewpoints as no major party candidate holds them so they will get no votes. Admittedly, the major parties tend to absorb popular third-party positions in order to avoid getting voted out, but that is a very slow process.
The voting system does not have to be perfect, but you must keep in mind that it indirectly influences political debate and the responsiveness of the government to the people.
To be fair, the choice of a voting system (and the entire structure of the political process) is a choice of what actually matters in government. A two-party system theoretically forces moderate views and compromise instead of ending up with multiple warring parties sharing power. Also the argument can be made that plurality voting is easier to understand, especially when compared to range voting. That is part of why, despite its flaws, I suspect instant-runoff voting is the most reasonable choice for a voting system.
No. Perhaps you are failing to accept that people might just plain disagree with you on whether monitoring at the level the UK government performs it is okay. I have a friend from Scotland who is fully aware of what is going on and simply trusts the government and police to use it properly and more or less sees any assertions that they will use it improperly as conspiracy theories -- and is convinced that the government is far too disorganized to actually follow through on any complicated secret plans. She also trusts the government officials to actually believe that these systems will do good.
The common "nothing to hide" counterarguments mean nothing to her. She figures both that if someone is doing something they want kept secret, it is probably something they should not be doing anyway, and that if the surveillance picks up on something secret, most likely it will be seen by random police officer who doesn't know the people and doesn't care, so it does not really matter.
She's not stupid. She just thinks the examples of possible problems sound like rare extreme cases which are not worth preventing.
I would be interested to hear (have linked?) good arguments against monitoring the population which will not just sound like "privacy nuts want privacy so you should, too" to most people, but I have yet to see any.
The Palm Treo was originally a Palm OS device, which was switched over to Windows Mobile. They were originally designed by Handspring whose products were always a bit glitchy, so I suspect the problems may have been partially hardware related.
My issue with WoT for e-mail is that if botnets are sending the e-mail already, then the spammers have easy access to at least thousands of computers: getting access to "known good" signing keys of the web of trust would trivial. If the spammer gets reported then the signer becomes suspect. Essentially, to the common user this looks like their signing key randomly breaks. Depending on the system design, some recipients may decide they are a spammer and drop their e-mails without notice.
As long as malware and botnets are ubiquitous, a trustworthy widespread web of trust seems unlikely. On the other hand, without botnets, spam would likely not be a major issue (although there are plenty of reasons why a web of trust is a nice thing to have around anyway, mainly that it would allow most e-mail to be encrypted).
There are other refinements possible. Your whitelist can accept, not just individual signatures, but "badges" from some organization. So, anyone from Mozilla.org can attach a Mozilla.org badge to their emails, and I can allow all Mozilla.org emails through. IEEE member badge, SourceForge.net badge, Apple.com badge, go nuts. Even an organization of "I Swear I Will Never Send Out Spam". The key with the badges is that, if you get kicked out of an organization, you have to lose access to the badge. One simple way would be for the check to be live: if you attach a Mozilla.org badge, the Mozilla.org server had better agree that your identity is one known to it.
Congratulations, you just invented Facebook messages.
Seriously, I like the idea of signed e-mails, web of trust, etc., but I cannot see it being feasible for the common user. And anyway, why would they bother? Facebook allows them to set which groups may message them and handles spammers. I do not like it much -- and explicitly ask people to use e-mail if they try to contact me via Facebook message -- but there are significant social and technical barriers to what you are suggesting.
Furthermore, I am not sure how much signing e-mails would actually help if spam is already sent by botnets via random people's computers: the botnet could easily just collect a bunch of keys and contact lists (so it would know who would accept e-mails signed with that key). It would at least let you know whose computer was compromised which would (possibly) add some clear personal responsibility to keeping one's computer clean of malware.
Thanks. I enabled that. Unfortunately, it is only for Qt4 and the few Qt apps I use are still on Qt3, but that is not really Qt's fault. I can't comment on how well it handles emulating the GTK feel because the only Qt4 app I have appears to be qtconfig and it does not have that wide a variety of widgets.
From your example, a computer that can run an O(2^n) algorithm on up to n=1000 inputs in less than or equal to 1 second, must be able to execute at least 2^1000 operations per second! You couldn't do that with computer composed of 1 processor for each atom in the universe, each running in parallel.
No, Big-O notation basically means to ignore constant factors. The complexity might be 2^-1000*2^n operations for a specific implementation. Then it takes 1 operation for n=1000 and 2 operations for n=1001. (Maybe the processor implements the problem up to n=1000 in hardware and for bigger n some work needs to be done in software.) This is an extreme example, of course, but it is within the definition.
Also, n is the number of input to the algorithm. It has nothing to do with its implementation. I think the value you are thinking of is k, which is a scalar that represents computational overhead of the algorithm, and that is certainly implementation dependent. Big-O says nothing about k.
I never said it had anything to do with the implementation. The statement that a problem is O(2^n) means that there exist algorithms for the problem whose speed grows asymptotically no faster than 2^n for large n, so for each of those implementations, the maximum amount of time that implementation takes to run on a specific input size can be written as a function f(n) (may be piecewise -- in fact probably has no simple explicit formula) which is less than C*2^n for all n>N (for some C, N dependent on the implementation). For n<=N, it says nothing. C may be arbitrarily small or large. See Big O notation for the precise mathematical definition using limits.
I am not sure what you mean about k. k is often used for things other than the input size in the complexity of an algorithm like the parameter in Toom-Cook multiplication. In that case, k varies giving a family of algorithms Toom-k where Toom-3 is a specific algorithm from that family.
Complexity theory is about how quickly the time it takes to do a task for large n where "large" is very much so implementation dependent. If n=1000 counts as "large" for an O(2^n) problem (say, for all implementations that fit in this universe), then any machine will take twice as long to do a n=1001 instance as it takes to do a n=1000 instance. For "small" n, it says nothing. For example, there might be an algorithm for that O(2^n) problem that takes 1 second for all n up to 1000.
That is not considering constant factors: maybe the increase is exponential, but for every meaningful input size a modern computer can run the algorithm in under a second. Also, many hard problems have somewhat easier approximations which are good enough for many purposes.
That said, overall, you are right: there are real limits to computation which we believe will not be broken. The best we can hope for is better chips and slightly better algorithms. I think Kurzweil is focusing on Moore's law continuing or being accelerated, heading towards the fastest possible physical computers which also have limits but much further out ones.
Your requirements sound similar to the ones that led me to buy a Nokia n810: I wanted something I could carry everywhere without really thinking about it that I could use for PDA-type applications and internet access (it replaced a Sony Clie Palm OS-based PDA (yeah, yeah, Sony sucks, but they sure knew how to make a Palm OS 4 device: I would still be using it if I could have gotten Wi-Fi easily)). I did consider the iPod Touch, but between the on-screen keyboard and lack of external memory slot, I wasn't interested. On the other hand, I was looking at Wi-Fi capable handhelds instead of phones because, as a university student, pretty much everywhere I am going to be will have Wi-Fi, so there is no point in paying for cell phone internet.
I am happy with the N810, but it is a bit slow for web browsing and freezes occasionally (may be due to me opening too many browser windows / terminals sometimes...) -- it would probably greatly benefit from doubling the RAM and a slightly faster processor -- so I would recommend finding one to play around with and seeing if the speed is acceptable. There are rumors about a successor being worked on.
On the upside, it runs a Debian-based distro called Maemo and pretty much any armel Debian package is installable.
To be honest, this looks like a developing/early adopters market at the moment, and realistically your best bet for getting a good usable device in the PDA/smartphone form factor is probably to wait a few years.
Another user mentionedSingularity and covered it pretty well, so I will not repeat what he said except that using Sing# makes Singularity faster than C because it is managed. It is a bit counterintuitive because when most people think of managed code, they think of JIT/interpreter overhead, but there is no reason to not native compile it and a good compiler should get better results than with C because the compiler knows more about the structure of the program.
One idea is to limit the destructive actions the user can take. Disk space is cheap, so, for most uses, infinite undo/undelete kept around forever would not be that hard (in the worst case, undo can be handled by a copy of the original file and a timestamped list of every action done to it... it just may take a while to get back to a specific version). Of course, there are privacy issues in not being able to really delete files, so there would have to be some way to do it... but the OS could be setup to try to ensure that files could only be permanently deleted by direct user interaction. Most other issues can probably be handled by limiting permissions (so out of hand applications can't get too out of hand) and setting up systems so users would not consider running unsigned code as normal (package management goes a long way here).
I agree that user security is a hard problem, but there is certainly more research to be done in the area.
You are right. The technology available is horrible. There is no good reason that you should have to run your own server or rely on some specific company. Your computer should have a "share this photo album" button that magically creates a torrent and makes it easily accessible with a pretty interface. Just no one has written that program yet because Flickr/Picassa/Facebook photo albums work well enough. RMS is not saying there that are good solutions, just that all of the currently used ones have huge issues and they should not be considered good enough.
There are multiple rationales for copyright. You seem to believe in the natural rights of the author and their heirs to control the uses of the work. Wikipedia mentions that the cases you present are related to the concept of personality rights.
In the United States, the U.S. Constitution gives the rationale of "to promote the Progress of Science and useful Arts". The view you reference is that that means copyright is a sort of loan from the public to the creator(s) and that copyright exists purely to allow creative works to be sold for a long enough period of time to ensure their creation is sufficiently profitable for it to actually happen -- and no more. That is, copyright is far from being a natural right: it is a necessary evil that should be minimized as much as possible without damaging the creation of new works.
From that perspective, the question is not "Why is copyright bad?" but "Why is copyright good?" based on the belief that all limitations on personal liberty need to be justified.
I hope so. At the very least, then stuff like Facebook's constant interface changes would not annoy people (as you would just use your own personalized (web?) client). Really the idea of any one company's servers be a single point of failure for a service seems silly. From an internet architecture point of view, tying a service to a specific set of servers run by a select few seems silly if not plain ridiculous. Ex. Imagine if, say, AT&T ran e-mail. (Then again, some people mainly use Facebook messages instead of e-mail...)
There are some problems with such an approach, though. The main one I can think of is how to handle an equivalent of Facebook's networks (which essentially comes down to groups defined by people in the group or by controlling an e-mail address in a specific domain) across multiple servers. Jabber already seems to handle spam okay (in that I have never gotten spam on any Jabber account), but that might be due to its relatively small size and that IM spam is pretty rare anyway.
How could Twitter be done in "P2P fashion"? Do we all put our tweets in text files and add them to a big torrent?
Twitter is pretty similar to IM status messages with a few extensions like being able to see the time the status message was set and seeing previous status messages. It also does not require the user to be online, but I am not certain XMPP requires that anyway. The @ and # and any other features could probably also be done as extensions to XMPP. It would still require special software (web being a fine interface), but it would then be an open standard, and, more importantly, not rely on a centralized system.
Facebook would be more complicated to decentralize, but it could probably be done without too much trouble.
The idea is that there is no reason for these services to be centralized. In the extreme case, every user could run their own server, making it p2p, but, more likely, a situation like e-mail or XMPP would evolve where there are many public and private servers that users use and it does matter which one a user is on.
Slashdot Mirror
Agreed: Windows, Mac OS X, and Linux/*BSD are all broken in that they allow the user to install malware -- software that does stuff no user would ever want. (Some other operating systems like gaming consoles or the iPhone's OS "fix" this by only allowing vendor approved apps.) A better solution would be default limited rights for apps. For example, if a user actually wants to make 100 connections per minute on port 25, they probably know what they are doing. Similarly for a key logger. None of Windows, OS X, and *nix make such limitations easy, if they are even possible. This should be considered a flaw in their UI or the OS itself, not the user.
Another major problem is the lack of ubiquitous undo for everything (even uninstalling applications, deleting system files, changing system settings). Windows tries to approximate this with System Restore, but I have never seen that do any good (but maybe Windows has gotten better in this area since I last used it). OS X has Time Machine, which is a real step in the right direction. Effectively unlimited trustable undo would mean the user would not be scared of messing things up because they would also feel safe that they could undo their changes.
Making operating systems robust but usable and discoverable for all users is no minor task, but it should be a goal.
Although it is tempting to say that in a perfect world without greed much more would be accomplished, one has to consider what the incentives are. In theory, capitalism encourages innovation by providing rewards for it. In a system where no one has to work, many won't. Of course, it is unclear this is a bad thing: many people unable to do what they want in a capitalistic system may decide to be artists or scientists if they do not have to worry about making money.
Just looking at the computer field, it seems like things would be better for everyone if all the software around were available for free, but it is not clear it would have been written in the first place. It is sometimes argued that a lot of the popular Linux programs and desktop environments do a lot of copying from Windows/Mac. If there were not people needing to make a living working on these UI projects, would they have ever gotten done? Microsoft and Apple had strong incentives to bring the personal computer to the masses. Would an alternative economic system offer the same incentives? Should it?
As for alternative systems, you may be interested to read about Social Credit which I first saw discussed in Robert Heinlein's series of essays^W^W^Wutopian novel For Us, The Living. Another alternative system is discussed in Marshall Brain's work Manna (available online) where abundance (and robot laborers) organized by a company founded for social good creates a utopia where each person gets a daily allowance of production credits.
On the more realistic and immediately implementable end of things, there's Why Work? ("CLAWS: Creating Livable Alternatives to Wage Slavery"), a website dedicated to not having a job [that you don't like].
It's a protocol which I believe is based on XMPP (Jabber) and makes heavy use of HTML5, so any client would have to pretty much embed a web browser in order to view messages. Google made a web client first because (1) it's Google, they like coding for web and (2) it is a quick and easy way to make an app cross-platform. If Google does not make a native client for it (like they made Google Talk), then other projects will because a lot of people (including myself) agree with you that the web is the wrong place for an IM client, especially seeing as persistence is part of the protocol so there is no real advantage to running the protocol on someone else's server as the protocol should effectively sync all the clients you use automatically. Most likely, once it is finalized, there will be multiple open source projects to create native clients.
They're working on it. Well, not exactly, but Firefox and Chrome will both support the new HTML/JS extension framework being discussed.
Not sure about the Washington Post, but the NY Times is listening: they made an Adobe AIR app called the TimesReader which has the articles well formatted without ads which is available to print subscribers at no extra charge or for $14.95/month. The price is a bit higher than you are discussing, but it is certainly a step in the right direction.
Correct. Have you really never used a [memory] safe programming language? The only Java debugger I have used is the one in Eclipse which displays objects in a tree with the contents listed by field name (and uses their .toString() methods to get a String summary to display, but for many objects that is useless). Java does not really deal with memory layouts of its objects.
I agree it seems pretty sketchy, but then why is one of Amazon's Eee pages linking to it?
(Or at least trying to: at time of writing, the text link labeled "www.ItsBetterwithWindows.com" is a link to "file:///C:/DOCUME%7E1/jross/LOCALS%7E1/Temp/Temporary%20Directory%202%20for%201000HE%20Black%20New.zip/1000HE%20Black%20New/www.itsbetterwithwindows.com". That seems massively unprofessional for Amazon, not sure what is going on there. The image link is correct.)
Yes, DNS can be fixed. The basic idea is to let ICANN just be one authority among many. Put the entire current DNS under .icann (and default to appending .icann so you don't break stuff, I guess) and let anyone else run their own DNS hierarchy setting up competition in the area of properly assigning domain names (for however your users define "properly" here). The result would most likely be a Wikipedia-like distributed oversight system for who controls which domain names, hopefully with no cost for "registering" a domain which seems pretty silly.
The linked essays explain and argue for it much better than I can. In the end, the proposed system makes the root more or less powerless so it would no longer really matter who controlled it.
Why is everyone going on and on about the screen being too small or low resolution or having bad colors? He suggested including HDMI-out so it could be hooked up to a normal screen for tasks which require one.
Thanks for the link. They already have a proportional representation system which is significantly better than the US system for avoiding a two party system. It is a bit disappointing that the voters would consider STV too complicated, but it does sound like it significantly changes the political landscape, especially in that it sounds like it changes the focus of voting from parties to individuals which may make the elected individuals less predictable as they would not necessarily be kept in line by their party.
The regional lines that legislative positions are based on in the US mean that PR is not really an option. IRV is basically the same system, but it does become slightly simpler when counting for only one position.
There are a few places that use IRV, even some in the US.
Well, very simply, you could use range voting which is not limited by Arrow's impossibility theorem as it does not satisfy the hypotheses.
In reality, pretty much any voting system other than the current first past the post would be a massive improvement due to strategic voting. All voting systems have problems with it, but under plurality voting, it essentially forces a two-party system. By allowing voters to rank their preferences, votes would come much closer to revealing the true preferences of the electorate, and therefore the government would be closer to what the people really want (or at the very least, the politicians would be claiming to support what the people really want).
In the current system, the major parties are able to shape the debate and effectively silence any opposing viewpoints as no major party candidate holds them so they will get no votes. Admittedly, the major parties tend to absorb popular third-party positions in order to avoid getting voted out, but that is a very slow process.
The voting system does not have to be perfect, but you must keep in mind that it indirectly influences political debate and the responsiveness of the government to the people.
To be fair, the choice of a voting system (and the entire structure of the political process) is a choice of what actually matters in government. A two-party system theoretically forces moderate views and compromise instead of ending up with multiple warring parties sharing power. Also the argument can be made that plurality voting is easier to understand, especially when compared to range voting. That is part of why, despite its flaws, I suspect instant-runoff voting is the most reasonable choice for a voting system.
No. Perhaps you are failing to accept that people might just plain disagree with you on whether monitoring at the level the UK government performs it is okay. I have a friend from Scotland who is fully aware of what is going on and simply trusts the government and police to use it properly and more or less sees any assertions that they will use it improperly as conspiracy theories -- and is convinced that the government is far too disorganized to actually follow through on any complicated secret plans. She also trusts the government officials to actually believe that these systems will do good.
The common "nothing to hide" counterarguments mean nothing to her. She figures both that if someone is doing something they want kept secret, it is probably something they should not be doing anyway, and that if the surveillance picks up on something secret, most likely it will be seen by random police officer who doesn't know the people and doesn't care, so it does not really matter.
She's not stupid. She just thinks the examples of possible problems sound like rare extreme cases which are not worth preventing.
I would be interested to hear (have linked?) good arguments against monitoring the population which will not just sound like "privacy nuts want privacy so you should, too" to most people, but I have yet to see any.
The Palm Treo was originally a Palm OS device, which was switched over to Windows Mobile. They were originally designed by Handspring whose products were always a bit glitchy, so I suspect the problems may have been partially hardware related.
My issue with WoT for e-mail is that if botnets are sending the e-mail already, then the spammers have easy access to at least thousands of computers: getting access to "known good" signing keys of the web of trust would trivial. If the spammer gets reported then the signer becomes suspect. Essentially, to the common user this looks like their signing key randomly breaks. Depending on the system design, some recipients may decide they are a spammer and drop their e-mails without notice.
As long as malware and botnets are ubiquitous, a trustworthy widespread web of trust seems unlikely. On the other hand, without botnets, spam would likely not be a major issue (although there are plenty of reasons why a web of trust is a nice thing to have around anyway, mainly that it would allow most e-mail to be encrypted).
There are other refinements possible. Your whitelist can accept, not just individual signatures, but "badges" from some organization. So, anyone from Mozilla.org can attach a Mozilla.org badge to their emails, and I can allow all Mozilla.org emails through. IEEE member badge, SourceForge.net badge, Apple.com badge, go nuts. Even an organization of "I Swear I Will Never Send Out Spam". The key with the badges is that, if you get kicked out of an organization, you have to lose access to the badge. One simple way would be for the check to be live: if you attach a Mozilla.org badge, the Mozilla.org server had better agree that your identity is one known to it.
Congratulations, you just invented Facebook messages.
Seriously, I like the idea of signed e-mails, web of trust, etc., but I cannot see it being feasible for the common user. And anyway, why would they bother? Facebook allows them to set which groups may message them and handles spammers. I do not like it much -- and explicitly ask people to use e-mail if they try to contact me via Facebook message -- but there are significant social and technical barriers to what you are suggesting.
Furthermore, I am not sure how much signing e-mails would actually help if spam is already sent by botnets via random people's computers: the botnet could easily just collect a bunch of keys and contact lists (so it would know who would accept e-mails signed with that key). It would at least let you know whose computer was compromised which would (possibly) add some clear personal responsibility to keeping one's computer clean of malware.
Thanks. I enabled that. Unfortunately, it is only for Qt4 and the few Qt apps I use are still on Qt3, but that is not really Qt's fault. I can't comment on how well it handles emulating the GTK feel because the only Qt4 app I have appears to be qtconfig and it does not have that wide a variety of widgets.
Cool. How do I get Amarok and k3b to use GTK widgets on my computer?
From your example, a computer that can run an O(2^n) algorithm on up to n=1000 inputs in less than or equal to 1 second, must be able to execute at least 2^1000 operations per second! You couldn't do that with computer composed of 1 processor for each atom in the universe, each running in parallel.
No, Big-O notation basically means to ignore constant factors. The complexity might be 2^-1000*2^n operations for a specific implementation. Then it takes 1 operation for n=1000 and 2 operations for n=1001. (Maybe the processor implements the problem up to n=1000 in hardware and for bigger n some work needs to be done in software.) This is an extreme example, of course, but it is within the definition.
Also, n is the number of input to the algorithm. It has nothing to do with its implementation. I think the value you are thinking of is k, which is a scalar that represents computational overhead of the algorithm, and that is certainly implementation dependent. Big-O says nothing about k.
I never said it had anything to do with the implementation. The statement that a problem is O(2^n) means that there exist algorithms for the problem whose speed grows asymptotically no faster than 2^n for large n, so for each of those implementations, the maximum amount of time that implementation takes to run on a specific input size can be written as a function f(n) (may be piecewise -- in fact probably has no simple explicit formula) which is less than C*2^n for all n>N (for some C, N dependent on the implementation). For n<=N, it says nothing. C may be arbitrarily small or large. See Big O notation for the precise mathematical definition using limits.
I am not sure what you mean about k. k is often used for things other than the input size in the complexity of an algorithm like the parameter in Toom-Cook multiplication. In that case, k varies giving a family of algorithms Toom-k where Toom-3 is a specific algorithm from that family.
Complexity theory is about how quickly the time it takes to do a task for large n where "large" is very much so implementation dependent. If n=1000 counts as "large" for an O(2^n) problem (say, for all implementations that fit in this universe), then any machine will take twice as long to do a n=1001 instance as it takes to do a n=1000 instance. For "small" n, it says nothing. For example, there might be an algorithm for that O(2^n) problem that takes 1 second for all n up to 1000.
That is not considering constant factors: maybe the increase is exponential, but for every meaningful input size a modern computer can run the algorithm in under a second. Also, many hard problems have somewhat easier approximations which are good enough for many purposes.
That said, overall, you are right: there are real limits to computation which we believe will not be broken. The best we can hope for is better chips and slightly better algorithms. I think Kurzweil is focusing on Moore's law continuing or being accelerated, heading towards the fastest possible physical computers which also have limits but much further out ones.
Your requirements sound similar to the ones that led me to buy a Nokia n810: I wanted something I could carry everywhere without really thinking about it that I could use for PDA-type applications and internet access (it replaced a Sony Clie Palm OS-based PDA (yeah, yeah, Sony sucks, but they sure knew how to make a Palm OS 4 device: I would still be using it if I could have gotten Wi-Fi easily)). I did consider the iPod Touch, but between the on-screen keyboard and lack of external memory slot, I wasn't interested. On the other hand, I was looking at Wi-Fi capable handhelds instead of phones because, as a university student, pretty much everywhere I am going to be will have Wi-Fi, so there is no point in paying for cell phone internet.
I am happy with the N810, but it is a bit slow for web browsing and freezes occasionally (may be due to me opening too many browser windows / terminals sometimes...) -- it would probably greatly benefit from doubling the RAM and a slightly faster processor -- so I would recommend finding one to play around with and seeing if the speed is acceptable. There are rumors about a successor being worked on.
On the upside, it runs a Debian-based distro called Maemo and pretty much any armel Debian package is installable.
To be honest, this looks like a developing/early adopters market at the moment, and realistically your best bet for getting a good usable device in the PDA/smartphone form factor is probably to wait a few years.
Another user mentioned Singularity and covered it pretty well, so I will not repeat what he said except that using Sing# makes Singularity faster than C because it is managed. It is a bit counterintuitive because when most people think of managed code, they think of JIT/interpreter overhead, but there is no reason to not native compile it and a good compiler should get better results than with C because the compiler knows more about the structure of the program.
One idea is to limit the destructive actions the user can take. Disk space is cheap, so, for most uses, infinite undo/undelete kept around forever would not be that hard (in the worst case, undo can be handled by a copy of the original file and a timestamped list of every action done to it... it just may take a while to get back to a specific version). Of course, there are privacy issues in not being able to really delete files, so there would have to be some way to do it... but the OS could be setup to try to ensure that files could only be permanently deleted by direct user interaction. Most other issues can probably be handled by limiting permissions (so out of hand applications can't get too out of hand) and setting up systems so users would not consider running unsigned code as normal (package management goes a long way here).
I agree that user security is a hard problem, but there is certainly more research to be done in the area.
You are right. The technology available is horrible. There is no good reason that you should have to run your own server or rely on some specific company. Your computer should have a "share this photo album" button that magically creates a torrent and makes it easily accessible with a pretty interface. Just no one has written that program yet because Flickr/Picassa/Facebook photo albums work well enough. RMS is not saying there that are good solutions, just that all of the currently used ones have huge issues and they should not be considered good enough.
You should probably start here: philosophy of copyright (consequentialist theories). Actually, the whole article is probably worth a read.
There are multiple rationales for copyright. You seem to believe in the natural rights of the author and their heirs to control the uses of the work. Wikipedia mentions that the cases you present are related to the concept of personality rights.
In the United States, the U.S. Constitution gives the rationale of "to promote the Progress of Science and useful Arts". The view you reference is that that means copyright is a sort of loan from the public to the creator(s) and that copyright exists purely to allow creative works to be sold for a long enough period of time to ensure their creation is sufficiently profitable for it to actually happen -- and no more. That is, copyright is far from being a natural right: it is a necessary evil that should be minimized as much as possible without damaging the creation of new works.
From that perspective, the question is not "Why is copyright bad?" but "Why is copyright good?" based on the belief that all limitations on personal liberty need to be justified.
I hope so. At the very least, then stuff like Facebook's constant interface changes would not annoy people (as you would just use your own personalized (web?) client). Really the idea of any one company's servers be a single point of failure for a service seems silly. From an internet architecture point of view, tying a service to a specific set of servers run by a select few seems silly if not plain ridiculous. Ex. Imagine if, say, AT&T ran e-mail. (Then again, some people mainly use Facebook messages instead of e-mail...)
There are some problems with such an approach, though. The main one I can think of is how to handle an equivalent of Facebook's networks (which essentially comes down to groups defined by people in the group or by controlling an e-mail address in a specific domain) across multiple servers. Jabber already seems to handle spam okay (in that I have never gotten spam on any Jabber account), but that might be due to its relatively small size and that IM spam is pretty rare anyway.
How could Twitter be done in "P2P fashion"? Do we all put our tweets in text files and add them to a big torrent?
Twitter is pretty similar to IM status messages with a few extensions like being able to see the time the status message was set and seeing previous status messages. It also does not require the user to be online, but I am not certain XMPP requires that anyway. The @ and # and any other features could probably also be done as extensions to XMPP. It would still require special software (web being a fine interface), but it would then be an open standard, and, more importantly, not rely on a centralized system.
Facebook would be more complicated to decentralize, but it could probably be done without too much trouble.
The idea is that there is no reason for these services to be centralized. In the extreme case, every user could run their own server, making it p2p, but, more likely, a situation like e-mail or XMPP would evolve where there are many public and private servers that users use and it does matter which one a user is on.