If you want to avoid phishing (especially the economic fraud that you're talking about here), the weak security of the SPF protocol is about enough to inspire a false sense of security without actually doing anything. At *least* use a pubkey-based signing system if you are trying to pull this off.
Hell, if you wanted to use SPF's poor domain-level-granularity and security-depending-upon-nobody-spoofing-DNS, you could just have a big flag for a domain that says "all legitimate mail originating from this domain is S/MIME signed".
Naturally, *this* still wouldn't do a much to avoid phishing, because the email could come from foobar@citibank-customerservice.com or some equally bogus name, but at least it wouldn't be as completely ineffective as SPF.
I am very, very frusterated with SPF. Angry even, which is why I'm so combative. It reminds me of the Bulk Priority flag created by the Internet2 people -- the folks involved were just not *thinking* about how they're impacting existing people. I am fed up with mail servers rejecting email from my personal mail server because it's on dial-up (bad spam filtering idea #1), ISPs blocking my mail server from connecting out because "no reasonable home user would want a mail server" (bad spam filtering idea #2), ISPs silently dropping my email because I had to relay through their server just to get mail out through their system, then moved to a different ISP and forgot to update the relay host (bad spam filtering idea #3)...
Honestly, I have far more dislike for most of the people in the antispam arena (not all, there are some folks that put out carefully reasoned systems) than the spammers. Spammers make me waste a bit of time setting up some filters. They have wasted some of my bandwdith. I don't like them much. Yet they have *still* not managed to inhibit functionality or make my email get dropped, like a small but frusterating subset of antispam people that insist on breaking existing systems for short-term "improvements".
Here's my argument -- if you can't make a permanent solution that will be significantly useful in stopping the problem at hand (reducing spam) for at *least* ten years (remember, Internet email's been around for much longer) and your solution is a huge pain in the ass for regular users, please *don't* push such a system. Spammers are not going to just continue trying to spam into SPF, which is the assumption one has to make to assume that SPF is going to solve anything. They'll just buy the latest idiot-proof tool to avoid SPF, and continue spamming.
I'm getting the impression that a lot of the people who post objections haven't bothered to read the FAQ.
I *have* read the FAQ. Some of my complaints about SPF are based on its content. I've steadily posted objections to Slashdot in most SPF stories -- at first, I spent ages on each post, writing out lengthy lists of issues. It was only after any SPF people failed to respond that I stopped listing item-by-item issues.
Of course, he told me it would never work and that it was unreasonable to use DNS to provide such a service.
He's also dead-on right. There are many ways to simply bypass SPF; SPF-like schemes impose a number of limitations on regular email use; the use of DNS is a major part of the problem in SPF, as it's entirely inappropriate for a trusted transport.
The only reason anyone's deploying SPF, broken as it is, is because people will do *anything* at this point if someone promises them *any* reduction in spam -- look at the RIAA, and how they keep trying copy-protection scheme after copy-protection scheme. They've had PhDs doing studies come forward and say "there isn't any feasible way to copy-protect simple stereo audio as you're currently selling it", but they so badly want to believe that they keep buying into the latest scheme. ISPs get a huge amount of flak from customers over spam, and are more than willing to take even idiotic and nonfunctional measures if it (a) lets them believe that they're doing something that will be long-term effective and (b) lets their marketing people have something to trumpet about ("Earthlink's advanced spam-blocking technology")
Really, it sucks that there's no visual association between child and parent windows (like a string attaching them, or something). If a dialog comes up from a Javascript, how are you to know what frame it belongs to?
The idea up throwing up dialogs really predates the need to provide a trusted interface to the user.
Plant a couple of desktops in Iraq, and we'll finally have some weapons of mass destruction to justify the whole invasion, occupation, and installation of puppet government figures thing.
Good job, Osgyth. We've had so many piss-poor Ask Slashdots for so long that I was startled to see over a thousand responses in a story that is broadly useful (not a question with an answer that only has value to you), unique (this data can't be effectively obtained elsewhere), and valuable (finding good publications is very nice).
True, if by "most of us" you mean "those of us who happen to be morons." Guess why nobody sends credit card numbers over e-mail?
*Legally* secure. A criminal isn't going to care how he gets ahold of credit card numbers. If you're sending a bunch of potentially valuable business ideas, however, they're secure.
Most already do.
They don't without *informing* you that they're doing so. Employers cannot read your email without telling you in advance.
It's a free service. They should be able to do whatever the hell they feel like. Read the usage agreement.
Not without including such a notification in their usage agreement, is the point. Now, all the people that *didn't* explicitly say that they might be reading your email *can*.
If you think an ISP wouldn't cooperate with the FBI without a warrant, then you are a moron.
If the FBI is abusing their privileges, it's easy to raise a stink (and has been done in the past).
I can't speak for anyone else, or even for my own ISP, but if the FBI came in to request data on a system that I was administering, didn't have a warrant, and I had no instructions from my boss telling me to hand it to them, they certainly wouldn't be getting anything. The warrant is the way the judicial system acts as a check on law enforcement -- it's terribly important to require its use.
My employees sign a user agreement that acknowledges the company's right to read emails and any information stored on a company computer.
And that agreement was put in place either by a lawyer or because someone based it on agreements at other institutions that were installed by a lawyer, because to do otherwise would be illegal. The user knows that the company may monitor their emails. This eliminates that.
I tell the employees that the email system is like the company's phone system: a few personal emails is OK, but abuse will not be tolerated.
I'm just curious -- are employees at your work ever contacted out of work for any reason, expected to work weekends or after hours, or expected to use personal assets (like fuel in their car) for business purposes? Because it's not as if there isn't any justification for personal use of the network -- obviously, if someone is causing a problem or not fulfilling their duties, there is an issue, but I think that if an employee is not willing to strictly stay off employee time and resources, they should not expect an employee not to utilize business time and resources.
That being said, I'm pretty religious about not misusing business resources -- I won't take pens or disks with me if I need one for personal data, nor do I browse the network for anything non-business related except after I've finished up for the day and am waiting to leave with someone (and that doesn't mean porn browsing -- it means reading research papers and software stuff that I'm interested in.) However, I'd consider it quite unreasonable for my employer to have a problem with this.
Oh, c'mon. It'll be fun. You can pretend that you're the 51st state. You just have to give up your liberal drug law and immigration law and TV copyright law and you get to be part of the last world superpower! Won't that be great?
* Email, the dominant form of online communication, which most of us have regarded as fairly secure, is now grabable by federal authorities or police *without a warrant*.
* Your employer may now read all your email -- previously, he had to at least inform you that he was going to monitor your network traffic ahead of time (admittedly, including such a clause in the usage policy was depressingly common, but still).
* Free email providers like Yahoo, Microsoft, and Google now are free to do anything they want with all the mail that you've ever sent or has been sent to you.
I'm sure that the EFF is scrambling to try and do something at the moment -- it'll be their most important case yet.
*IF* this is not overturned, it means that it is *impossible* to have legal privacy protection for any form of communication that is asynchronous across hosts. This affects a vast number of potential protocols.
This means that voicemail systems are *not* protected by federal wiretapping law. If you *ever* leave a message for anyone, your privacy protections are out the window.
It's debatable over whether or not this applies to web caching -- if police and federal agents can now swipe the content of your ISP's web cache (yeah, the transparent proxy that your cable ISP uses, even though you don't think you're using a proxy), they can obtain web browsing data without warrant.
This is the biggest argument I've seen yet for use of PGP. If you are not using PGP, you *have* no privacy.
Of course, it's more convenient to only point fingers at those whose politics you disagree with.
Almost as convenient as making unfounded and incorrect claims to defend your favorite politician.
Read some discussion on the differences beween the Ashcroft and Reno DoJ and come back and tell me that the Bush administration smells as sweet as the Clinton administration.
Hitler didn't do any mass killing of Jews until well into World War II when things weren't looking so good anymore and there were some concerns about large numbers of pissed-off Jews running around after the War.
Remember that Hitler started out as nothing more than authoritarian leader with a penchant for manipulating terrorist attacks into police power, a tendency to rely on stirring up peole with fear and nationalistic fervor, xenophobia, and a desire to exercise military strength.
Really, the only major differences between Hitler's early reigme and Bush's that I see are:
* Hitler *always* made ethnic issues a major portion of his campaign. Bush does not have an "ethic enemy within".
* Hitler used physical intimination against domestic politicians before he had the legal right to do so -- his party had military strength of its own. Bush has not physically threatened supporters of opposing politicians.
That being said, the extent to which the Bush/Hitler administrations parallel each other vary, but there *are* a startling number. Almost all the major political events have parallels.
Wrong. Even if you think both Bush and Kerry are gutter trash and the Dems and Republicans are useless, if you endorse a more liberal party (the Green Party, say), it is to your interest to see the Demms win. The Demms leech votes away from the other, more liberal parties because their margins of winning are slim, and folks would rather see a Democrat in office than a Republican. The only way that more liberal parties will *get* votes is for the Demms to get a significant majority, and then for those voters to slip away to the more liberal parties. These votes are *not* going to go directly from Republican to Green.
If Demms get twice the number of votes that the Republicans do, then it's likely that Green or others might get a significant number of votes. Anyone who chooses to keep the Demms from winning by a significant margin is simply postponing the day that other parties can become serious contenders.
Splitting voters is one of the most effective ways to gain a majority. What you are doing is one of the most useful tactics that a Republican would want to take to ensure that Bush wins the election.
And the necessary hardware at that stage was rudimentary, slow and bulky.
If you can find cheap, lightweight, hard-to-break stereoscopic VR goggles at the kind of resolution and refresh rate that a 17-inch monitor displays (i.e. 1152x864, at least 60Hz) and at a cost that is comparable to such a monitor (not thousands of dollars), then I suspect that you'll have a market. I'd certainly enjoy the privacy, unobtrusiveness, and utility of being able to wear such goggles.
I've seen several posts here that lament open source's propensity to copy rather than innovate.
The people that post this are wrong. A huge chunk of serious academic research is open source.
What they mean to say is that "most of the popular open source programs have closed source predecessors", which is entirely to be expected, as the closed source programs being succeeded *got popular* because they worked well. There aren't that many groundbreaking popular closed source programs either -- there can be only one person to start the idea of the "office suite" or the "web browser".
Another warning label in life, brought on by a broken US judicial system that awards vast amounts of punitive damages to the occasional winner of a bullshit lawsuit and ignores common sense.
It's a pain, because it's so much harder to build Windows-from-scratch barebones systems than their Linux equivalents. I've seen a lot of Windows kiosks, and they're almost always loaded with scads of things they don't need because it's so hard to really pare down a Windows box.
I'm going to be blunt and say that the best way to do this is with Linux, because it's much easier to pare down.
Set up a bunch of thin clients with netbooting enabled. That means no CD drive, floppy drive, hard drive. Lock the BIOS. Buy cases that are physically securable.
Have one or several Windows Terminal Server boxes set up.
Set up your netboot server to serve a Linux distro something like Red Hat (or an even more bare-bones system), installing a minimal set of packages necessary. You'll want to install rdesktop so that your clients can act as Terminal Server clients, but no terminals or anything. In/etc/inittab, remove all VTs. In/etc/X11/XF86Config, kill the "special" xorg key combinations (like control-alt-backspace). Don't have xterm or any such terminals installed. Use an xsession set up to start rdesktop, and a window manager of your choice that can slap something up fullscreen and disable all other functionality -- almost all can do this, but you'll probably want something more barebones than the sawfish that I use. Have rdesktop running fullscreen. Set up X to respawn logged in to whatever user you have using the program.
The user should have no write access to anything on the Linux distro (if you want to include a small swap drive, you might want to have a local hard drive, but only root should be able to write to the thing).
The user should have no write access to anything on the Windows TS system (unless as required by your application). Hence, the users can't install anything. It's easy to administer. You don't have to pay for each client, since they're running Linux, which makes a decent thin client OS.
Now, you can do whatever you want in a trusted manner on the TS system(s), since the users don't have the ability to reboot or muck with it, since they have no local access (and rebooting or mucking with their thin client does nothing that gives them any influence over what applications are running on the server). Kill all processes that you don't recognize automatically or whatnot.
1. They are creating a much more robust community than they ever have. Check out http://blogs.msdn.com sometime. They have a lot of their developers - and not just low level guys - blogging on a regular basis. It's an interesting thing to watch these people work. And it really gets out of that "faceless corproate entity" mold they were heading down.
It sounds nice, and I went there with some hopeful anticipation that these would be honest, straightforward thoughts from the developers that we could trust for advance data. However, quotes like this: "I see Andy beat me to the punch of being the first VC blogger to mention the new beta VC++ Express SKU. If I wasn't a Microsoft employee, I'd be downloading this right now instead of typing this entry..." make this sound like nothing more than a new marketing ploy than an actual attempt to provide users with honest, BS-free information from the developers. I like mailing lists, bugzilla/open bug trackers/forums because it means that with Open Source, I don't have to swim through oceans of meaningless crap from marketers to find out what I want. It's a *far* better environment for a user.
It really discourages salespeople to have to do this. I've found that a lot of vendors seem to not deal well with younger purchasers that were brought up on the Web being a source of actual solid data. Last time I was trying to evaluate products for a major purchase, I just wanted a damn feature and price list on the software product involved from each vendor. I'd look all over the vendor's website -- nothing, just "contact this number to obtain information". I call them up to get a feature list, and the salesguy fights tooth and nail to avoid giving me a any hard data. He wants me "to specify my needs" so that he can "assist me in making a decision". I'm quite capable of making my own evaluation, and he's quite reachable if I have technical questions, but he is absolutely scared stiff that his product won't be able to stand on its own merit, and needs a healthy dose of BS about how wonderful it is from him before I'll buy anything. It's astounding, completely contrary to the way I've always worked, and a huge waste of my time.
If there's a problem in a piece of software, I want to know about it, not have some completely ridiculous sales assurances that it "isn't an issue" or "will be fixed soon." If you're straightforward, I don't *have* to expect that the product will be perfect. *That's* why I want to be able to see what developers are saying and talk to them. Salesmen are conditioned to feed out bullshit. It's incredibly frusterating for anyone spoiled on the luxuries of the open source world, which has a very simple contract. If you have simple problems, you resolve them yourself (or have a consultant do it) to avoid consuming developer time. You check online documentation to avoid asking duplicate questions. In exchange for this simple outlay of effort, if you have serious issues that still aren't resolved, you can reach the developer, and get no-BS answers. If mutt can't do something, the people on mutt-users are going to say "you can't do that" or suggest a method of making it work, rather than trying to spin everything. It's awfully nice.
2. The software is getting better. Windows is pretty reliable now. It's not perfect by any means, but Windows 2000 was the first shot. Windows XP and Windows 2003 are really quite a bit better. It's easy to joke about "the most reliable Windows ever". In the real world there isn't that dread like there was in the NT4 days about BSOD's and reliablitly problems.
Yes, but "better than Windows used to be" is a lot different than "good". The biggest thing I heard about Win2k for home use was that "It's better than Windows 9x", which was a pretty weak claim. The issue here is how Windows stacks up to Linux, not how it compares to old version of Windows. Linux 2.6 is a lot nicer than Linux 1.0, for instanc
Slashdot Mirror
And so how exactly does this avoid spam, again?
If you want to avoid phishing (especially the economic fraud that you're talking about here), the weak security of the SPF protocol is about enough to inspire a false sense of security without actually doing anything. At *least* use a pubkey-based signing system if you are trying to pull this off.
Hell, if you wanted to use SPF's poor domain-level-granularity and security-depending-upon-nobody-spoofing-DNS, you could just have a big flag for a domain that says "all legitimate mail originating from this domain is S/MIME signed".
Naturally, *this* still wouldn't do a much to avoid phishing, because the email could come from foobar@citibank-customerservice.com or some equally bogus name, but at least it wouldn't be as completely ineffective as SPF.
I'll even provide links to some of my criticizing comments.
I am very, very frusterated with SPF. Angry even, which is why I'm so combative. It reminds me of the Bulk Priority flag created by the Internet2 people -- the folks involved were just not *thinking* about how they're impacting existing people. I am fed up with mail servers rejecting email from my personal mail server because it's on dial-up (bad spam filtering idea #1), ISPs blocking my mail server from connecting out because "no reasonable home user would want a mail server" (bad spam filtering idea #2), ISPs silently dropping my email because I had to relay through their server just to get mail out through their system, then moved to a different ISP and forgot to update the relay host (bad spam filtering idea #3)...
Honestly, I have far more dislike for most of the people in the antispam arena (not all, there are some folks that put out carefully reasoned systems) than the spammers. Spammers make me waste a bit of time setting up some filters. They have wasted some of my bandwdith. I don't like them much. Yet they have *still* not managed to inhibit functionality or make my email get dropped, like a small but frusterating subset of antispam people that insist on breaking existing systems for short-term "improvements".
Here's my argument -- if you can't make a permanent solution that will be significantly useful in stopping the problem at hand (reducing spam) for at *least* ten years (remember, Internet email's been around for much longer) and your solution is a huge pain in the ass for regular users, please *don't* push such a system. Spammers are not going to just continue trying to spam into SPF, which is the assumption one has to make to assume that SPF is going to solve anything. They'll just buy the latest idiot-proof tool to avoid SPF, and continue spamming.
I'm getting the impression that a lot of the people who post objections haven't bothered to read the FAQ.
I *have* read the FAQ. Some of my complaints about SPF are based on its content. I've steadily posted objections to Slashdot in most SPF stories -- at first, I spent ages on each post, writing out lengthy lists of issues. It was only after any SPF people failed to respond that I stopped listing item-by-item issues.
Of course, he told me it would never work and that it was unreasonable to use DNS to provide such a service.
He's also dead-on right. There are many ways to simply bypass SPF; SPF-like schemes impose a number of limitations on regular email use; the use of DNS is a major part of the problem in SPF, as it's entirely inappropriate for a trusted transport.
The only reason anyone's deploying SPF, broken as it is, is because people will do *anything* at this point if someone promises them *any* reduction in spam -- look at the RIAA, and how they keep trying copy-protection scheme after copy-protection scheme. They've had PhDs doing studies come forward and say "there isn't any feasible way to copy-protect simple stereo audio as you're currently selling it", but they so badly want to believe that they keep buying into the latest scheme. ISPs get a huge amount of flak from customers over spam, and are more than willing to take even idiotic and nonfunctional measures if it (a) lets them believe that they're doing something that will be long-term effective and (b) lets their marketing people have something to trumpet about ("Earthlink's advanced spam-blocking technology")
Really, it sucks that there's no visual association between child and parent windows (like a string attaching them, or something). If a dialog comes up from a Javascript, how are you to know what frame it belongs to?
The idea up throwing up dialogs really predates the need to provide a trusted interface to the user.
Plant a couple of desktops in Iraq, and we'll finally have some weapons of mass destruction to justify the whole invasion, occupation, and installation of puppet government figures thing.
Good job, Osgyth. We've had so many piss-poor Ask Slashdots for so long that I was startled to see over a thousand responses in a story that is broadly useful (not a question with an answer that only has value to you), unique (this data can't be effectively obtained elsewhere), and valuable (finding good publications is very nice).
My hat's off to you, sir.
True, if by "most of us" you mean "those of us who happen to be morons." Guess why nobody sends credit card numbers over e-mail?
*Legally* secure. A criminal isn't going to care how he gets ahold of credit card numbers. If you're sending a bunch of potentially valuable business ideas, however, they're secure.
Most already do.
They don't without *informing* you that they're doing so. Employers cannot read your email without telling you in advance.
It's a free service. They should be able to do whatever the hell they feel like. Read the usage agreement.
Not without including such a notification in their usage agreement, is the point. Now, all the people that *didn't* explicitly say that they might be reading your email *can*.
If you think an ISP wouldn't cooperate with the FBI without a warrant, then you are a moron.
If the FBI is abusing their privileges, it's easy to raise a stink (and has been done in the past).
I can't speak for anyone else, or even for my own ISP, but if the FBI came in to request data on a system that I was administering, didn't have a warrant, and I had no instructions from my boss telling me to hand it to them, they certainly wouldn't be getting anything. The warrant is the way the judicial system acts as a check on law enforcement -- it's terribly important to require its use.
My employees sign a user agreement that acknowledges the company's right to read emails and any information stored on a company computer.
And that agreement was put in place either by a lawyer or because someone based it on agreements at other institutions that were installed by a lawyer, because to do otherwise would be illegal. The user knows that the company may monitor their emails. This eliminates that.
I tell the employees that the email system is like the company's phone system: a few personal emails is OK, but abuse will not be tolerated.
I'm just curious -- are employees at your work ever contacted out of work for any reason, expected to work weekends or after hours, or expected to use personal assets (like fuel in their car) for business purposes? Because it's not as if there isn't any justification for personal use of the network -- obviously, if someone is causing a problem or not fulfilling their duties, there is an issue, but I think that if an employee is not willing to strictly stay off employee time and resources, they should not expect an employee not to utilize business time and resources.
That being said, I'm pretty religious about not misusing business resources -- I won't take pens or disks with me if I need one for personal data, nor do I browse the network for anything non-business related except after I've finished up for the day and am waiting to leave with someone (and that doesn't mean porn browsing -- it means reading research papers and software stuff that I'm interested in.) However, I'd consider it quite unreasonable for my employer to have a problem with this.
Oh, c'mon. It'll be fun. You can pretend that you're the 51st state. You just have to give up your liberal drug law and immigration law and TV copyright law and you get to be part of the last world superpower! Won't that be great?
Wow. This is a huge, huge, huge deal.
Among other things, this means:
* Email, the dominant form of online communication, which most of us have regarded as fairly secure, is now grabable by federal authorities or police *without a warrant*.
* Your employer may now read all your email -- previously, he had to at least inform you that he was going to monitor your network traffic ahead of time (admittedly, including such a clause in the usage policy was depressingly common, but still).
* Free email providers like Yahoo, Microsoft, and Google now are free to do anything they want with all the mail that you've ever sent or has been sent to you.
I'm sure that the EFF is scrambling to try and do something at the moment -- it'll be their most important case yet.
*IF* this is not overturned, it means that it is *impossible* to have legal privacy protection for any form of communication that is asynchronous across hosts. This affects a vast number of potential protocols.
This means that voicemail systems are *not* protected by federal wiretapping law. If you *ever* leave a message for anyone, your privacy protections are out the window.
It's debatable over whether or not this applies to web caching -- if police and federal agents can now swipe the content of your ISP's web cache (yeah, the transparent proxy that your cable ISP uses, even though you don't think you're using a proxy), they can obtain web browsing data without warrant.
This is the biggest argument I've seen yet for use of PGP. If you are not using PGP, you *have* no privacy.
Of course, it's more convenient to only point fingers at those whose politics you disagree with.
Almost as convenient as making unfounded and incorrect claims to defend your favorite politician.
Read some discussion on the differences beween the Ashcroft and Reno DoJ and come back and tell me that the Bush administration smells as sweet as the Clinton administration.
Hitler didn't do any mass killing of Jews until well into World War II when things weren't looking so good anymore and there were some concerns about large numbers of pissed-off Jews running around after the War.
Remember that Hitler started out as nothing more than authoritarian leader with a penchant for manipulating terrorist attacks into police power, a tendency to rely on stirring up peole with fear and nationalistic fervor, xenophobia, and a desire to exercise military strength.
Really, the only major differences between Hitler's early reigme and Bush's that I see are:
* Hitler *always* made ethnic issues a major portion of his campaign. Bush does not have an "ethic enemy within".
* Hitler used physical intimination against domestic politicians before he had the legal right to do so -- his party had military strength of its own. Bush has not physically threatened supporters of opposing politicians.
That being said, the extent to which the Bush/Hitler administrations parallel each other vary, but there *are* a startling number. Almost all the major political events have parallels.
I find it funny that the only remotely reasonable pro-Bush arguments I've seen are "well, Kerry ain't perfect either", which is pretty awful.
Honestly, that's a significantly more plausible excuse than "it'll delete everything".
Wrong. Even if you think both Bush and Kerry are gutter trash and the Dems and Republicans are useless, if you endorse a more liberal party (the Green Party, say), it is to your interest to see the Demms win. The Demms leech votes away from the other, more liberal parties because their margins of winning are slim, and folks would rather see a Democrat in office than a Republican. The only way that more liberal parties will *get* votes is for the Demms to get a significant majority, and then for those voters to slip away to the more liberal parties. These votes are *not* going to go directly from Republican to Green.
If Demms get twice the number of votes that the Republicans do, then it's likely that Green or others might get a significant number of votes. Anyone who chooses to keep the Demms from winning by a significant margin is simply postponing the day that other parties can become serious contenders.
Splitting voters is one of the most effective ways to gain a majority. What you are doing is one of the most useful tactics that a Republican would want to take to ensure that Bush wins the election.
I was talking about the kid having to sneak the magazine.
I'm curious -- what functionality are you thinking of?
And the necessary hardware at that stage was rudimentary, slow and bulky.
If you can find cheap, lightweight, hard-to-break stereoscopic VR goggles at the kind of resolution and refresh rate that a 17-inch monitor displays (i.e. 1152x864, at least 60Hz) and at a cost that is comparable to such a monitor (not thousands of dollars), then I suspect that you'll have a market. I'd certainly enjoy the privacy, unobtrusiveness, and utility of being able to wear such goggles.
I've seen several posts here that lament open source's propensity to copy rather than innovate.
The people that post this are wrong. A huge chunk of serious academic research is open source.
What they mean to say is that "most of the popular open source programs have closed source predecessors", which is entirely to be expected, as the closed source programs being succeeded *got popular* because they worked well. There aren't that many groundbreaking popular closed source programs either -- there can be only one person to start the idea of the "office suite" or the "web browser".
I'm guessing that it's a pain to set up.
None of the mainstream distros have idiot-proof "check to use encryption" interfaces. It probably took a lot of work to set this beast up.
Another warning label in life, brought on by a broken US judicial system that awards vast amounts of punitive damages to the occasional winner of a bullshit lawsuit and ignores common sense.
Guess who would prosecute treason?
Yup, the Bush Administration.
Guess which department?
Yup, the Department of Justice.
Guess who isn't going to get nailed for treason?
It's a pain, because it's so much harder to build Windows-from-scratch barebones systems than their Linux equivalents. I've seen a lot of Windows kiosks, and they're almost always loaded with scads of things they don't need because it's so hard to really pare down a Windows box.
/etc/inittab, remove all VTs. In /etc/X11/XF86Config, kill the "special" xorg key combinations (like control-alt-backspace). Don't have xterm or any such terminals installed. Use an xsession set up to start rdesktop, and a window manager of your choice that can slap something up fullscreen and disable all other functionality -- almost all can do this, but you'll probably want something more barebones than the sawfish that I use. Have rdesktop running fullscreen. Set up X to respawn logged in to whatever user you have using the program.
I'm going to be blunt and say that the best way to do this is with Linux, because it's much easier to pare down.
Set up a bunch of thin clients with netbooting enabled. That means no CD drive, floppy drive, hard drive. Lock the BIOS. Buy cases that are physically securable.
Have one or several Windows Terminal Server boxes set up.
Set up your netboot server to serve a Linux distro something like Red Hat (or an even more bare-bones system), installing a minimal set of packages necessary. You'll want to install rdesktop so that your clients can act as Terminal Server clients, but no terminals or anything. In
The user should have no write access to anything on the Linux distro (if you want to include a small swap drive, you might want to have a local hard drive, but only root should be able to write to the thing).
The user should have no write access to anything on the Windows TS system (unless as required by your application). Hence, the users can't install anything. It's easy to administer. You don't have to pay for each client, since they're running Linux, which makes a decent thin client OS.
Now, you can do whatever you want in a trusted manner on the TS system(s), since the users don't have the ability to reboot or muck with it, since they have no local access (and rebooting or mucking with their thin client does nothing that gives them any influence over what applications are running on the server). Kill all processes that you don't recognize automatically or whatnot.
1. They are creating a much more robust community than they ever have. Check out http://blogs.msdn.com sometime. They have a lot of their developers - and not just low level guys - blogging on a regular basis. It's an interesting thing to watch these people work. And it really gets out of that "faceless corproate entity" mold they were heading down.
It sounds nice, and I went there with some hopeful anticipation that these would be honest, straightforward thoughts from the developers that we could trust for advance data. However, quotes like this: "I see Andy beat me to the punch of being the first VC blogger to mention the new beta VC++ Express SKU. If I wasn't a Microsoft employee, I'd be downloading this right now instead of typing this entry..." make this sound like nothing more than a new marketing ploy than an actual attempt to provide users with honest, BS-free information from the developers. I like mailing lists, bugzilla/open bug trackers/forums because it means that with Open Source, I don't have to swim through oceans of meaningless crap from marketers to find out what I want. It's a *far* better environment for a user.
It really discourages salespeople to have to do this. I've found that a lot of vendors seem to not deal well with younger purchasers that were brought up on the Web being a source of actual solid data. Last time I was trying to evaluate products for a major purchase, I just wanted a damn feature and price list on the software product involved from each vendor. I'd look all over the vendor's website -- nothing, just "contact this number to obtain information". I call them up to get a feature list, and the salesguy fights tooth and nail to avoid giving me a any hard data. He wants me "to specify my needs" so that he can "assist me in making a decision". I'm quite capable of making my own evaluation, and he's quite reachable if I have technical questions, but he is absolutely scared stiff that his product won't be able to stand on its own merit, and needs a healthy dose of BS about how wonderful it is from him before I'll buy anything. It's astounding, completely contrary to the way I've always worked, and a huge waste of my time.
If there's a problem in a piece of software, I want to know about it, not have some completely ridiculous sales assurances that it "isn't an issue" or "will be fixed soon." If you're straightforward, I don't *have* to expect that the product will be perfect. *That's* why I want to be able to see what developers are saying and talk to them. Salesmen are conditioned to feed out bullshit. It's incredibly frusterating for anyone spoiled on the luxuries of the open source world, which has a very simple contract. If you have simple problems, you resolve them yourself (or have a consultant do it) to avoid consuming developer time. You check online documentation to avoid asking duplicate questions. In exchange for this simple outlay of effort, if you have serious issues that still aren't resolved, you can reach the developer, and get no-BS answers. If mutt can't do something, the people on mutt-users are going to say "you can't do that" or suggest a method of making it work, rather than trying to spin everything. It's awfully nice.
2. The software is getting better. Windows is pretty reliable now. It's not perfect by any means, but Windows 2000 was the first shot. Windows XP and Windows 2003 are really quite a bit better. It's easy to joke about "the most reliable Windows ever". In the real world there isn't that dread like there was in the NT4 days about BSOD's and reliablitly problems.
Yes, but "better than Windows used to be" is a lot different than "good". The biggest thing I heard about Win2k for home use was that "It's better than Windows 9x", which was a pretty weak claim. The issue here is how Windows stacks up to Linux, not how it compares to old version of Windows. Linux 2.6 is a lot nicer than Linux 1.0, for instanc