Basically, all of the suggestions offerered here and in TFA are things that a browser was never meant to do. HTTP does an extremely good job at what it is designed to do: deliver HTML pages with text and images. Every feature since that people have manged to add since then has been hack upon hack. CGI and PHP were clever indeed, but can only be taken so far in terms of application complexity. Trying to turning the web into an applications platform is almost as ridiculous as trying to turn email into a feature-rich VoIP conferencing solution. Only nobody sees this because the former almost works for a few limited purposes.
But these days, everyone wants the web to do everything. A noble goal, for sure, but not a very realistic one. Writing a full-feature web app is outright painful. The presentation layer is inflexible, the protocols are stateless, the security by and large sucks, and you have to learn a minimum of 5 different languages for the average web app: a general-purpose language, a markup language, a stylesheet language, a database language, and a client-side scripting language. On top of that, browser vendors still can't agree on which subsets of the standards to support.
We don't need new hacks bolted onto the web, we need something that will make the web obsolete as an application environment. I'm not saying chuck the WWW out the window, but design an application environment to complement it. New protocols, new presentation methods, a good security layer, and modular storage components. If done correctly, you should be able to either visit a website and use an application there or download it to your desktop and use it offline. It should be able to degrade gracefully, like HTML does, when devices or clients are unable to support all features (like a mobile web browser or voice-driven browsers for the blind).
It should be noted that there have been attempts at technologies similar to this. Java promised exactly this, and maybe was a bit ahead of its time, but its drawbacks (performance and complexity, among others) outweighed its benefits for the hobbyist coder. XUL might have also been a good start, but failed to gain a critical mass of developers, even among its own designers.
Whoever invents this web application environment and makes it successful will be a millionaire. Now get to work, the future needs you!
I think I'll try 64 bit linux next.. Never tried a 64 bit rev... Any suggestions?
Give the 64-bit version of Ubuntu a spin. I'm running it on my Core Duo Thinkpad have yet to run into a single problem where the 64- versus 32-bit thing is concerned.
I could just as easily sit on my couch in the dark and use my cell phone.
But the cell networks typically fall down and are completely useless during any type of large-scale emergency. Cell phones were completely useless to those ensnared in the Northeast Blackout of 2003 but I never had any problem getting through to people on my land line (until the batteries at the CO ran out anyway). Cell carriers design and maintain their infrastructure under the assumption that only a small percentage of consumers will be using the network at any given time and they don't bother to plan for contingencies. So when something happens that prompts a bunch of people to dial a number and hit Send at once, the whole thing falls down.
This is okay, I guess, except that WRT54GL has been available for a long time and has roughly the same specs.
A few years ago, I thought that open-source Linux-based routers would have been a boon for router manufacturers and end-users alike, but that dream has yet to be realized. Off-the-shelf wifi routers still have little more than basic NAT functionality. The third-party firmware options offer far more features, of course, but every one of the projects seem to either lack focus or developers and consequently don't even come close to realizing the full potential of these semi-powerful little boxes.
DD-WRT was forked from the Sveasoft code because some guy didn't like the fact that they started selling their firmware. A few years later, he turns around and does something similar.
OpenWRT is a great base upon which to build, but even it can be somewhat buggy and lacks a decent interface. The X-WRT Web UI is good start, but nothing I would dare put into production.
For home and small-office routing, it's probably easier to buy some cheap low-power x86-class machine (maybe even an old laptop) and either put something like m0n0wall on it or just install your distribution of choice and tweak away. This is what I plan to do in the near future.
I have a desktop machine at home, my wife's desktop, a laptop, a desktop at work, and a server at work. All five are completely different hardware. When Ubuntu 8.04 came out, I installed it on each of them using the Ubuntu installer. Out of those five machines, only one gave me any grief and that was the work desktop because the proprietary nvidia drivers still can't gracefully handle two monitors of differing size.
So, I pick a name, and McDeepPockets comes along and thinks, hey, that's a great idea - I'll just take that, thank-you. They "dispute" it, and ICANN's response is... well, if you really can't settle your differences, high bidder gets it. Wow... that's going to make for a pretty mercenary internet.
You say this as if it were a new policy of some kind... ICANN has always operated this way.
You can already do that with the current DNS, the heirarchy is just reversed:
this.would.be.such.an.example.com
Similarly, WHOIS spamming seems to be all the rage these days. Just register some fake nameservers and you're all set. Use the standard unix 'whois' command to check it out. Depending on the WHOIS server you hit, you'll get entries like:
I usually don't read YouTube comments because they tend to be racist, trollish, or just plain inflammatory in general. But the first one underneath the video is priceless:
"I think the most disturbing part of this entire video is that every vehicle shown in motion is driving on the wrong side of the road. BPL seems like a minor issue in comparison."
Ditto. Computers have always been my favorite and strongest skill, but it's hard to find job that's either rewarding or high-paying, let alone both. These days, it's difficult to make serious money in I.T. unless you're a hot-shot consultant or come up with the next big thing on the web.
If I could go back and talk to my younger self, I'd recommend getting some kind of financial investment degree.
The problem with artificial intelligence as it's portrayed as robots or machines that are self-aware and make decisions for themselves. We don't need machines that are just as bad at making decisions as humans are. This isn't a very useful application and if we succeed in creating it, the novelty will probably wear off as quickly as the Tamagotchi fad.
The practical application of AI will be interfaces. For years, the average first-world citizen has possessed far more computing power than he or she knows what to do with. What we're lacking is a good way to interface with that power so that we can use it to its fullest potential. We have software that can translate the words that people are speaking into text, but nothing yet that can understand those words, do whatever is required, and return a meaningful result either visually or audibly.
Think Star Trek, where (except in a few episodes) the computer doesn't take on a life of its own, it simply listens for questions and commands and then either provides answers or executes those commands. It's a perfect universal interface between humans and machines. You don't have to know anything about the computer save the fact that it exists. If the machine has access to the Internet, it has instant access to the single largest information database in human history.
This form of AI is the only piece of the puzzle that's missing before we can take computing technology to the next level and make using it as simple as asking a question aloud.
I gave Opera 9.5 a whirl last week and was highly impressed. It's packed with nice features (Where do you think Firefox and IE get most of their ideas?) but still pretty fast and light. Other versions of Opera never did much for me, but this is the first proprietary application that I've run across in a long time that I would seriously consider using on a daily basis. The only areas where it's really lacking are modularity (extensions, instead of everything being built-in to the browser) and of course the fact that it's not free software.
That's a pretty moronic metaphor. No matter how you look at it, a house is not equivalent in any way to a wifi signal and removing things from it is not equivalent to accessing the Internet. See my sig for further explanation.
Webmail alone isn't a bad thing for encryption. It's just that no major webmail providers offer any way at all to use encryption with their service.
If anything, webmail could make encrypted mail easier to use if the provider builds the PKI bits right into the webmail application. Granted, you might have to paste your public key into a text field every time you send a message (so that javascript or some client-side thing can encrypt the contents before sending). Another option would be to simply trust the email provider to store your key securely and not give it out to any governments which might ask. If you're a company and you run your own webmail server with HTTPS, you probably trust yourself just fine and can have the webmail app automatically encrypt messages for intra-company email and merely sign messages to external addresses.
None of this is all that difficult, the only hard part is just doing the work to get everything automated for the end user.
The reason PGP, and GPG as well, fail is because PKI is just too difficult to setup and maintain. I'm sure some nerd who lives in his mom's basement is going to contest this but the fact remains it's too difficult to do in most corporations let alone end users. Making a key, remembering the password, managing keys, revoking keys, it's all just a total pain in the ass.
The thing is, it doesn't have to be. While PKI is rather complex underneath, well-designed software can cover it all up for the end user. While it might help in certain cases, they don't even necessarily need to know the difference between a public and private key (even though a 4-year-old can understand this just fine). The user just clicks a button somewhere in their program that says, "encrypt this message.". There's no reason why public key distribution can't be made almost or entirely automatic. If encryption ever becomes popular enough (one can dream), the software can automatically encrypt everything from mail to IM to files stored on a flash drive without requiring anything extra of the user (except maybe a password to unlock the private key for use).
What I'd like to see is some kind of WORM (write once, read many) USB keyfob for storing a single keypair on. You plug it into your computer whenever you're creating or working on encrypted data or into someone else's computer if you want to give them your public key.
But I installed the Beta on my son's machine, and was shocked at the 'awesomebar'. What a monumentally bad idea, implemented in the most annoying of fashion! It is seriously the one factor keeping me from switching.
*sigh*
This seems to be the standard reaction from people who want a new shiny upgraded program but are absolutely stunned when the developers actually attempt to, I dunno, change the program's behavior in some way. The fact that it's a web browser seems to amplify this effect even more. My god, I remember when people went up in arms because Mozilla changed the freaking icons in Firefox.
I actually like the awesome bar. While its presentation is a tad on the annoying side, it does let you quickly search within all of the pages you've recently viewed, rather than just the URLs you've recently happened to type in. Think of it as Google, only inside your browser. It's worth noting that Opera has had this functionality for quite a long time, so those who like it should should thank the Opera company rather than the FF devs.
And just think, even if you hate it now, a year down the road you won't want to use a browser that doesn't have an awesome bar because you'll have gotten used to using it by then.
I'll have to second this. I never really cared much for Opera in the past, but this 9.5 stuff is pretty darn good. Works on all the sites I can find, has a clean interface, is blindingly fast, and has better standards support than anything else. Opera also just released a bunch of really cool web developer tools to go with it.
I'm not 100% sure whether Opera will be my main browser anytime soon, but this is the first time I've actually taken more than a cursory glance at it.
One of the strengths of Firefox for some time has been that right out of the box, the binary just ran on lots of Linux versions. With FF3 (starting with betas) they broke this.
I downloaded Firefox3, untarred it it to my desktop, and it ran just fine.
A non-trivial portion of the commercial and research Linux user base has to stick with EL4 or a source rebuild from CentOS, Scientific Linux or whatever because of third party tool support requirements. And not everybody wants to upgrade their OS just because a new browser is out.
I posit that open source application developers should not be expected to support every OS that might be in use at the time of release. This is basically how the open source world works: Project X releases a stable version of their source code and then the distribution developers port, test, and package the software for use with their specific distribution.
Since RHEL5/CentOS 5 has been out for quite some time, RHEL4 and variants are considered legacy OSes in many circles, especially when it comes to the fast-changing world of the Linux desktop. It's not fair to blame the Firefox devs for linking against a library that brings them many benefits and new features but might not happen to come pre-installed on any number of old distributions. If anyone's to blame here, it's your "third party tool" vendor because they're locking you into a distribution that rapidly becoming unsupported by the rest of the world.
No, you're absolutely right. Every few months someone comes out with this "running cars on water" thing, and every time it's the same technology. Notice the following quote in the article:
"The car has an energy generator that extracts hydrogen from water that is poured into the car's tank."
This device isn't an energy generator at all, it's a device which requires electricity in order to separate the hydrogen from the oxygen. (I think this is called hydrolysis?) The end result is that you end up expending more energy trying to get at the hydrogen than you get back from burning it. The stories about "water cars" in the popular media always gloss over this little detail.
So yes, it's perfectly possible to make a car that uses water as fuel, but the chemical reactions required to make it work require a lot of electricity which presently is neither cheap nor clean.
Encryption is only good for protecting data after the computer is turned off. In this case, the computer was attacked remotely while it was on. So even if the data on it was encrypted, the encrypted image was likely unlocked with all of its data waiting to be read just like a normal filesystem.
Seems to be just speculation, because the article never says that the attackers in this particular case was the Chinese government, just that they were "Chinese." The author of the article was probably hoping to dupe Slashdot editors.
It may be just a coincidence anyway that the attack came from China. The first rule of unauthorized electronic penetration is that you never attack a system in your home country unless you really know what you're doing, so the fact that the majority of successful U.S. Government hacks originate overseas isn't shocking.
In fairness, that article talks about fabric skins that are then covered with a lacquer that dries to make a hard surface, rather like many applications today use carbon fiber and plastic resin. This BMW skin is uncoated because it requires flexibility.
Slashdot Mirror
Basically, all of the suggestions offerered here and in TFA are things that a browser was never meant to do. HTTP does an extremely good job at what it is designed to do: deliver HTML pages with text and images. Every feature since that people have manged to add since then has been hack upon hack. CGI and PHP were clever indeed, but can only be taken so far in terms of application complexity. Trying to turning the web into an applications platform is almost as ridiculous as trying to turn email into a feature-rich VoIP conferencing solution. Only nobody sees this because the former almost works for a few limited purposes.
But these days, everyone wants the web to do everything. A noble goal, for sure, but not a very realistic one. Writing a full-feature web app is outright painful. The presentation layer is inflexible, the protocols are stateless, the security by and large sucks, and you have to learn a minimum of 5 different languages for the average web app: a general-purpose language, a markup language, a stylesheet language, a database language, and a client-side scripting language. On top of that, browser vendors still can't agree on which subsets of the standards to support.
We don't need new hacks bolted onto the web, we need something that will make the web obsolete as an application environment. I'm not saying chuck the WWW out the window, but design an application environment to complement it. New protocols, new presentation methods, a good security layer, and modular storage components. If done correctly, you should be able to either visit a website and use an application there or download it to your desktop and use it offline. It should be able to degrade gracefully, like HTML does, when devices or clients are unable to support all features (like a mobile web browser or voice-driven browsers for the blind).
It should be noted that there have been attempts at technologies similar to this. Java promised exactly this, and maybe was a bit ahead of its time, but its drawbacks (performance and complexity, among others) outweighed its benefits for the hobbyist coder. XUL might have also been a good start, but failed to gain a critical mass of developers, even among its own designers.
Whoever invents this web application environment and makes it successful will be a millionaire. Now get to work, the future needs you!
Better let legacy applications (and the whole x86 mess too, BTW) fade away
Legacy applications, sure. But good luck with the x86 part.
I think I'll try 64 bit linux next.. Never tried a 64 bit rev... Any suggestions?
Give the 64-bit version of Ubuntu a spin. I'm running it on my Core Duo Thinkpad have yet to run into a single problem where the 64- versus 32-bit thing is concerned.
I could just as easily sit on my couch in the dark and use my cell phone.
But the cell networks typically fall down and are completely useless during any type of large-scale emergency. Cell phones were completely useless to those ensnared in the Northeast Blackout of 2003 but I never had any problem getting through to people on my land line (until the batteries at the CO ran out anyway). Cell carriers design and maintain their infrastructure under the assumption that only a small percentage of consumers will be using the network at any given time and they don't bother to plan for contingencies. So when something happens that prompts a bunch of people to dial a number and hit Send at once, the whole thing falls down.
This is okay, I guess, except that WRT54GL has been available for a long time and has roughly the same specs.
A few years ago, I thought that open-source Linux-based routers would have been a boon for router manufacturers and end-users alike, but that dream has yet to be realized. Off-the-shelf wifi routers still have little more than basic NAT functionality. The third-party firmware options offer far more features, of course, but every one of the projects seem to either lack focus or developers and consequently don't even come close to realizing the full potential of these semi-powerful little boxes.
DD-WRT was forked from the Sveasoft code because some guy didn't like the fact that they started selling their firmware. A few years later, he turns around and does something similar.
OpenWRT is a great base upon which to build, but even it can be somewhat buggy and lacks a decent interface. The X-WRT Web UI is good start, but nothing I would dare put into production.
For home and small-office routing, it's probably easier to buy some cheap low-power x86-class machine (maybe even an old laptop) and either put something like m0n0wall on it or just install your distribution of choice and tweak away. This is what I plan to do in the near future.
In statistics, we call this kind of a sample "utterly fucking useless."
Anecdotes are fun, let me add my own:
I have a desktop machine at home, my wife's desktop, a laptop, a desktop at work, and a server at work. All five are completely different hardware. When Ubuntu 8.04 came out, I installed it on each of them using the Ubuntu installer. Out of those five machines, only one gave me any grief and that was the work desktop because the proprietary nvidia drivers still can't gracefully handle two monitors of differing size.
So, I pick a name, and McDeepPockets comes along and thinks, hey, that's a great idea - I'll just take that, thank-you. They "dispute" it, and ICANN's response is... well, if you really can't settle your differences, high bidder gets it. Wow... that's going to make for a pretty mercenary internet.
You say this as if it were a new policy of some kind... ICANN has always operated this way.
You can already do that with the current DNS, the heirarchy is just reversed:
this.would.be.such.an.example.com
Similarly, WHOIS spamming seems to be all the rage these days. Just register some fake nameservers and you're all set. Use the standard unix 'whois' command to check it out. Depending on the WHOIS server you hit, you'll get entries like:
MICROSOFT.COM.WILL.LIVE.FOREVER.BECOUSE.UNIXSUCKS.COM
MICROSOFT.COM.SOFTWARE.IS.NOT.USED.AT.REG.RU
MICROSOFT.COM.WILL.BE.SLAPPED.IN.THE.FACE.BY.MY.BLUE.VEINED.SPANNER.NET
Yet another abuse of the Internet that nobody seems to care about.
I usually don't read YouTube comments because they tend to be racist, trollish, or just plain inflammatory in general. But the first one underneath the video is priceless:
"I think the most disturbing part of this entire video is that every vehicle shown in motion is driving on the wrong side of the road. BPL seems like a minor issue in comparison."
Ditto. Computers have always been my favorite and strongest skill, but it's hard to find job that's either rewarding or high-paying, let alone both. These days, it's difficult to make serious money in I.T. unless you're a hot-shot consultant or come up with the next big thing on the web.
If I could go back and talk to my younger self, I'd recommend getting some kind of financial investment degree.
The problem with artificial intelligence as it's portrayed as robots or machines that are self-aware and make decisions for themselves. We don't need machines that are just as bad at making decisions as humans are. This isn't a very useful application and if we succeed in creating it, the novelty will probably wear off as quickly as the Tamagotchi fad.
The practical application of AI will be interfaces. For years, the average first-world citizen has possessed far more computing power than he or she knows what to do with. What we're lacking is a good way to interface with that power so that we can use it to its fullest potential. We have software that can translate the words that people are speaking into text, but nothing yet that can understand those words, do whatever is required, and return a meaningful result either visually or audibly.
Think Star Trek, where (except in a few episodes) the computer doesn't take on a life of its own, it simply listens for questions and commands and then either provides answers or executes those commands. It's a perfect universal interface between humans and machines. You don't have to know anything about the computer save the fact that it exists. If the machine has access to the Internet, it has instant access to the single largest information database in human history.
This form of AI is the only piece of the puzzle that's missing before we can take computing technology to the next level and make using it as simple as asking a question aloud.
is is not free now? It used to be shareware, but it is free now.
Free as in beer, yes. Free as in open source, no.
I gave Opera 9.5 a whirl last week and was highly impressed. It's packed with nice features (Where do you think Firefox and IE get most of their ideas?) but still pretty fast and light. Other versions of Opera never did much for me, but this is the first proprietary application that I've run across in a long time that I would seriously consider using on a daily basis. The only areas where it's really lacking are modularity (extensions, instead of everything being built-in to the browser) and of course the fact that it's not free software.
That's a pretty moronic metaphor. No matter how you look at it, a house is not equivalent in any way to a wifi signal and removing things from it is not equivalent to accessing the Internet. See my sig for further explanation.
You're thinking of FireGPG.
Webmail alone isn't a bad thing for encryption. It's just that no major webmail providers offer any way at all to use encryption with their service.
If anything, webmail could make encrypted mail easier to use if the provider builds the PKI bits right into the webmail application. Granted, you might have to paste your public key into a text field every time you send a message (so that javascript or some client-side thing can encrypt the contents before sending). Another option would be to simply trust the email provider to store your key securely and not give it out to any governments which might ask. If you're a company and you run your own webmail server with HTTPS, you probably trust yourself just fine and can have the webmail app automatically encrypt messages for intra-company email and merely sign messages to external addresses.
None of this is all that difficult, the only hard part is just doing the work to get everything automated for the end user.
The reason PGP, and GPG as well, fail is because PKI is just too difficult to setup and maintain. I'm sure some nerd who lives in his mom's basement is going to contest this but the fact remains it's too difficult to do in most corporations let alone end users. Making a key, remembering the password, managing keys, revoking keys, it's all just a total pain in the ass.
The thing is, it doesn't have to be. While PKI is rather complex underneath, well-designed software can cover it all up for the end user. While it might help in certain cases, they don't even necessarily need to know the difference between a public and private key (even though a 4-year-old can understand this just fine). The user just clicks a button somewhere in their program that says, "encrypt this message.". There's no reason why public key distribution can't be made almost or entirely automatic. If encryption ever becomes popular enough (one can dream), the software can automatically encrypt everything from mail to IM to files stored on a flash drive without requiring anything extra of the user (except maybe a password to unlock the private key for use).
What I'd like to see is some kind of WORM (write once, read many) USB keyfob for storing a single keypair on. You plug it into your computer whenever you're creating or working on encrypted data or into someone else's computer if you want to give them your public key.
*sigh*
This seems to be the standard reaction from people who want a new shiny upgraded program but are absolutely stunned when the developers actually attempt to, I dunno, change the program's behavior in some way. The fact that it's a web browser seems to amplify this effect even more. My god, I remember when people went up in arms because Mozilla changed the freaking icons in Firefox.
I actually like the awesome bar. While its presentation is a tad on the annoying side, it does let you quickly search within all of the pages you've recently viewed, rather than just the URLs you've recently happened to type in. Think of it as Google, only inside your browser. It's worth noting that Opera has had this functionality for quite a long time, so those who like it should should thank the Opera company rather than the FF devs.
And just think, even if you hate it now, a year down the road you won't want to use a browser that doesn't have an awesome bar because you'll have gotten used to using it by then.
I'll have to second this. I never really cared much for Opera in the past, but this 9.5 stuff is pretty darn good. Works on all the sites I can find, has a clean interface, is blindingly fast, and has better standards support than anything else. Opera also just released a bunch of really cool web developer tools to go with it.
I'm not 100% sure whether Opera will be my main browser anytime soon, but this is the first time I've actually taken more than a cursory glance at it.
One of the strengths of Firefox for some time has been that right out of the box, the binary just ran on lots of Linux versions. With FF3 (starting with betas) they broke this.
I downloaded Firefox3, untarred it it to my desktop, and it ran just fine.
A non-trivial portion of the commercial and research Linux user base has to stick with EL4 or a source rebuild from CentOS, Scientific Linux or whatever because of third party tool support requirements. And not everybody wants to upgrade their OS just because a new browser is out.
I posit that open source application developers should not be expected to support every OS that might be in use at the time of release. This is basically how the open source world works: Project X releases a stable version of their source code and then the distribution developers port, test, and package the software for use with their specific distribution.
Since RHEL5/CentOS 5 has been out for quite some time, RHEL4 and variants are considered legacy OSes in many circles, especially when it comes to the fast-changing world of the Linux desktop. It's not fair to blame the Firefox devs for linking against a library that brings them many benefits and new features but might not happen to come pre-installed on any number of old distributions. If anyone's to blame here, it's your "third party tool" vendor because they're locking you into a distribution that rapidly becoming unsupported by the rest of the world.
And just like that, you've gotten that insipid song stuck in the heads of thousands of slashdotters.
I hope you're happy.
No, you're absolutely right. Every few months someone comes out with this "running cars on water" thing, and every time it's the same technology. Notice the following quote in the article:
"The car has an energy generator that extracts hydrogen from water that is poured into the car's tank."
This device isn't an energy generator at all, it's a device which requires electricity in order to separate the hydrogen from the oxygen. (I think this is called hydrolysis?) The end result is that you end up expending more energy trying to get at the hydrogen than you get back from burning it. The stories about "water cars" in the popular media always gloss over this little detail.
So yes, it's perfectly possible to make a car that uses water as fuel, but the chemical reactions required to make it work require a lot of electricity which presently is neither cheap nor clean.
Encryption is only good for protecting data after the computer is turned off. In this case, the computer was attacked remotely while it was on. So even if the data on it was encrypted, the encrypted image was likely unlocked with all of its data waiting to be read just like a normal filesystem.
Seems to be just speculation, because the article never says that the attackers in this particular case was the Chinese government, just that they were "Chinese." The author of the article was probably hoping to dupe Slashdot editors.
It may be just a coincidence anyway that the attack came from China. The first rule of unauthorized electronic penetration is that you never attack a system in your home country unless you really know what you're doing, so the fact that the majority of successful U.S. Government hacks originate overseas isn't shocking.
In fairness, that article talks about fabric skins that are then covered with a lacquer that dries to make a hard surface, rather like many applications today use carbon fiber and plastic resin. This BMW skin is uncoated because it requires flexibility.