And yes really how often are you without internet.
Never.
No, seriously, absolutely fucking never. Disregarding the possibility of mobile connections, which would also give me Internet as a passenger on a road trip, I currently have Internet at every building on campus, including large swaths of central campus. Also pretty much every coffee shop or restaurant, even the cheap pizza joints have wifi now. Two of the last three times I've flown, there's been in-flight wifi.
And that's disregarding mobile connections.
A better question might be, when are you without power? I think you'll find a similar answer. Internet is a utility, as plentiful as power, heat, and running water.
Ooh can't connect to a wifi/my internet is out , I guess I wont do my homework...
Yes, because all homework necessarily requires a computer.
I find that most of my homework falls neatly into two categories -- either it practically (or actually) requires an Internet connection, or it doesn't require a computer at all.
Programming assignment -- if I can use the standard libs, and I haven't memorized them, having multiple online references is helpful. Even if I were to download them, pasting a particular error into Google is invaluable, not to mention blog entries and such covering things that might not be obvious from the reference.
Example: How do I make a string lowercase in C++? Does std::string have anything like java.lang.String.toLowerCase()? I'd have to read the entire reference for std::string, and then start searching for a way to make a character lowercase... Or I can just Google "lowercase string C++" and find some examples to get me started.
Then, since this is C++, for consistency, we have to verify that our code works on a particular machine we have ssh access to, in order to ensure that architectural differences don't make the program work for us but not for the TA who grades it.
It's also common that the homework isn't sufficiently specified, so I go to the discussion boards and ask. Even when I think I understand what's going on, there are often useful insights there.
My English projects can be pretty neatly broken into three parts: Preliminary research, reading, and actual writing. The preliminary research pretty much _needs_ Internet access. The actual reading either needs Internet access (because it's a Web article), or doesn't need a computer at all (a physical book from the library). The actual writing is the only step which would benefit from a computer without Internet, and that's a minority of it.
Math? Physical textbook, homework is best done on pencil and paper.
Physics? Physical textbook. Relatively small amounts require a computer. I do most of it on a computer by choice, but most of my classmates just print out the assignments and do everything except the plots on pencil and paper. The plots take maybe five minutes once the preliminary work is done.
That's with the exception of the online homework, and the online English quizzes, both of which must be taken in a browser, and, of course, online.
So for the roughly ten hours per year I don't have Internet, I would be able to do exactly as much homework with something like Chrome OS that I can do with my current OS -- which is still a considerable amount.
I'm not saying I would use Chrome OS, or that I never use native applications, but even the native apps I use at least benefit from Internet access, so I'm really getting sick of this argument. Mount Fucking Everest has 3G -- Google it if you don't believe me. This is roughly like arguing you shouldn't use a computer at all because the power might go out.
Wait, how would it force a specifically Google account?
Presumably, you can associate different local files with different web applications, which is about the only thing Chrome OS does other than show you webpages. Although Google has a spreadsheet offering, one of their Chrome OS demos was sticking a USB thumb drive into the netbook, clicking on an Excel document, and having it open in Microsoft's own Office Live.
So if you're that fucking paranoid, there's absolutely nothing about Chrome OS which forces you to use a Google account. Or, technically, any account -- nothing's stopping you from using your own server, or, y'know, just using it for the Internet -- I doubt it was ever meant to replace a desktop OS. Even if their was, Chromium OS exists.
Programming is not a production line, and trying to turn it into that leads to inefficient programmers, bad code, and maintenance nightmares. Programming is an art, a creative process, and a science, and there are definitely people who do it better than others, and platforms which make it easier than others.
That's important. Think about a typical ad agency, special effects company, or pretty much any design field where you can hire a contractor for a project. You hire them based on their work, because their work is recognizable and valuable. You also hire them based on prior experience working with them, how well you can communicate your ideas to them, and so on. You can pretend they're replaceable if you want, which is partly true -- there are always other design companies you can go to -- but you certainly don't think of them as cogs in an assembly line.
You sure as hell don't try to design your process so you can replace a single artist at any time.
However, ideas are valuable. I can't speak for other programmers, but I'm absolutely lost on the business side of things. From my perspective, sales, marketing, ideas, and so on are just some of the things I'm very glad other people do, all as part of the Development Abstraction Layer. I'm hopeless without them, to the point where on one-man projects, I usually end up asking every prospective customer, investor, or just friends and family, for advice on things like naming a price.
I'm not sure how I feel about idea people learning to program. They try anyway, with spreadsheets. Sometimes it ends well, but often it ends in disaster. It's usually not a good idea to hire a dedicated full-time programmer to work on spreadsheets, and the whole point of spreadsheets is to enable end-users to do these things. Still, a few basic programming concepts would go a long way, even if they are in spreadsheets.
(No, I don't mean VBA. Either program or don't, but to half-ass it by crawling up out of excel into VBA is only going to end in tears.)
And I do like to think I'm working on something really cool. I certainly want my "code poetry" to have a point. It's not that I can't appreciate idea people or their ideas, it's that I'm not much of an idea person myself -- or at least, my ideas don't tend to be the sort that are likely to make me money.
I dismissed your link because you did seem to be assuming quite a bit, and at least one of your assumptions was incorrect -- "restrict that account to pure scp transfer".
The fact that you used a TLA makes me suspect astroturf. RIA? What?
Nothing wrong with Rich Internet Applications per se, so long as they're actually based on standard technologies. To base them on Silverlight, which has exactly one decent implementation with exactly two supported platforms (Windows and Mac), is worse than basing them on Flash, which has exactly one decent implementation and supports Windows, Mac, and Linux. If I include the mobile space, it really doesn't change much -- pretty much everyone except Apple supports Flash and is proud of it, and pretty much only Windows Mobile supports anything like Silverlight.
So it's not that they're coming out of Redmond, it's that they fundamentally make certain websites Windows-only, or Windows/Mac only. Which means if you're truly an "avid Linux user," you're likely to run into a number of sites which don't work because they use Silverlight. I'm aware of Moonlight, and it just isn't there yet -- it's roughly like suggesting we all adopt Windows EXEs as a common execution platform because Wine exists.
Oh, and it also solves what's essentially a non-problem. HTML5 is good enough for the vast majority of what anyone wants to do with the Web. Extending it in small ways might be nice, and that's what people are doing (WebGL, etc.), but dropping a giant inner platform into it just makes things worse for everyone.
Apple has done more than anyone to standardize the Web, to make it multiple vendor safe and consumer friendly.
Yeah, I'll buy that when they natively support WebM or Theora/Vorbis.
They've done exactly what's in their best interest, and stopped just short of actually having to take the slightest amount of risk (perceived or real) to support the platform. Had they been willing to support any modern open video format, the whole codec war would've been over instantly.
Ok, let's compare, to the extent such comparisons are possible...
A modern system is 1-3 ghz.
Divide by 12 mhz. It's roughly one to three hundred times faster.
Assume it needed the full speed of the 386. Then 1-3% is exactly what you'd expect it to take on a modern CPU.
In other words, 1-3% is actually ridiculously low for something written in JavaScript on a modern system, compared to the hand-optimized C or ASM that would've been used to power the 386-based emulator.
When a 12mhz 386 can do fullspeed and it takes up to 5% on a recent machine, something is still deeply wrong.
Actually, regardless of the platform, if I can do something with 5% CPU that people actually want done, where's my incentive to make it take 0.5% CPU? There comes a point where ease/speed of development, readability, maintainability, stability, and security trump raw CPU usage.
I mean, giving Wikileaks an ssh account (as they're asking for) is pretty stupid, security-wise, but it's nowhere near as bad as giving _everyone_ your password by using FTP. You'd think Firesheep would've taught people something...
Lower the barrier of entry even further, and just throw up a torrent or ten of static files which can be hosted anywhere, without fear of compromising your own server.
In what way is Java more sandboxed than JavaScript? Both run in a VM with a limited, controlled API to access the rest of the world. What makes you think exposing the DOM to Java would make it any more secure than exposing it to JavaScript?
That doesn't seem to work in this case -- I've tried Kate.
What I don't get is when one of the buffers is clearly empty -- it's not pasting garbage, it's pasting nothing -- why wouldn't it check another? And why is this even an issue? On my system, I use Firefox or Chrome, depending on the situation. Both use GTK+. Why would this bug only affect Chrome, and only on certain websites, under certain conditions? Why do I never see this issue anywhere else? Why will it paste into other locations in Chrome, like the address bar or even other textboxes, just not a Slashdot comment box?
Even if it's X-related, the fact that other browsers get it right suggests that it's Chrome's bug, so again, where's the bug we can all vote up till someone actually fixes it?
Because the alternative, at least in Chrome, is obnoxious -- it doesn't have the option to download something to a temp folder and open it. Instead, if it sees anything it can't handle itself, either internally or with plugins, it's going to download it, automatically and instantly, cluttering my download folder to hell. It's especially bad when people put PDFs in iframes...
Acrobat actually isn't that bad, it's mostly the startup time that's a killer, but this is an improvement all around.
Actually, that is a security risk, currently mitigated on modern systems by not giving the root account a password, and only allowing root access through tools like sudo -- so in order to root a system, first you need to figure out which account is mine, only then can you break it.
First, you've got a purpose-built application. What's stopping you from implementing whatever protocol you want? Why on earth would you use FTP?
Second, IP-based security isn't security.
Third, why would you re-implement encryption, authentication, and authorization in your application when you can get them for free from FTPS, SFTP, HTTPS, etc?
Here, try this: "Oscar Meyer wiener". That song just came on in your head, didn't it?
Nope, actually took a bit of thought to remember it.
Know why? Because I avoid advertising. I buy or rent DVDs, or I torrent, but I don't watch network TV. I do watch YouTube, but I avoid channels that allow video ads. I don't use a preconfigured adblock, but I do block annoying ads, and I tend to avoid sites that have interstitials.
So, force me to wait five seconds, and you won't make me memorize your little jingle or slogan. You'll make me go out of my way to avoid seeing your ad, which is likely going to end worse for you than if you just let me skip it.
Slashdot Mirror
And yes really how often are you without internet.
Never.
No, seriously, absolutely fucking never. Disregarding the possibility of mobile connections, which would also give me Internet as a passenger on a road trip, I currently have Internet at every building on campus, including large swaths of central campus. Also pretty much every coffee shop or restaurant, even the cheap pizza joints have wifi now. Two of the last three times I've flown, there's been in-flight wifi.
And that's disregarding mobile connections.
A better question might be, when are you without power? I think you'll find a similar answer. Internet is a utility, as plentiful as power, heat, and running water.
Ooh can't connect to a wifi/my internet is out , I guess I wont do my homework...
Yes, because all homework necessarily requires a computer.
I find that most of my homework falls neatly into two categories -- either it practically (or actually) requires an Internet connection, or it doesn't require a computer at all.
Programming assignment -- if I can use the standard libs, and I haven't memorized them, having multiple online references is helpful. Even if I were to download them, pasting a particular error into Google is invaluable, not to mention blog entries and such covering things that might not be obvious from the reference.
Example: How do I make a string lowercase in C++? Does std::string have anything like java.lang.String.toLowerCase()? I'd have to read the entire reference for std::string, and then start searching for a way to make a character lowercase... Or I can just Google "lowercase string C++" and find some examples to get me started.
Then, since this is C++, for consistency, we have to verify that our code works on a particular machine we have ssh access to, in order to ensure that architectural differences don't make the program work for us but not for the TA who grades it.
It's also common that the homework isn't sufficiently specified, so I go to the discussion boards and ask. Even when I think I understand what's going on, there are often useful insights there.
My English projects can be pretty neatly broken into three parts: Preliminary research, reading, and actual writing. The preliminary research pretty much _needs_ Internet access. The actual reading either needs Internet access (because it's a Web article), or doesn't need a computer at all (a physical book from the library). The actual writing is the only step which would benefit from a computer without Internet, and that's a minority of it.
Math? Physical textbook, homework is best done on pencil and paper.
Physics? Physical textbook. Relatively small amounts require a computer. I do most of it on a computer by choice, but most of my classmates just print out the assignments and do everything except the plots on pencil and paper. The plots take maybe five minutes once the preliminary work is done.
That's with the exception of the online homework, and the online English quizzes, both of which must be taken in a browser, and, of course, online.
So for the roughly ten hours per year I don't have Internet, I would be able to do exactly as much homework with something like Chrome OS that I can do with my current OS -- which is still a considerable amount.
I'm not saying I would use Chrome OS, or that I never use native applications, but even the native apps I use at least benefit from Internet access, so I'm really getting sick of this argument. Mount Fucking Everest has 3G -- Google it if you don't believe me. This is roughly like arguing you shouldn't use a computer at all because the power might go out.
Wait, how would it force a specifically Google account?
Presumably, you can associate different local files with different web applications, which is about the only thing Chrome OS does other than show you webpages. Although Google has a spreadsheet offering, one of their Chrome OS demos was sticking a USB thumb drive into the netbook, clicking on an Excel document, and having it open in Microsoft's own Office Live.
So if you're that fucking paranoid, there's absolutely nothing about Chrome OS which forces you to use a Google account. Or, technically, any account -- nothing's stopping you from using your own server, or, y'know, just using it for the Internet -- I doubt it was ever meant to replace a desktop OS. Even if their was, Chromium OS exists.
Can't telecommute?
I know the feeling, though. Even noise-cancelling headphones aren't a substitute for actual silence.
I'll settle for relative quiet, though -- take a walk.
I honestly don't think either is true.
Programming is not a production line, and trying to turn it into that leads to inefficient programmers, bad code, and maintenance nightmares. Programming is an art, a creative process, and a science, and there are definitely people who do it better than others, and platforms which make it easier than others.
That's important. Think about a typical ad agency, special effects company, or pretty much any design field where you can hire a contractor for a project. You hire them based on their work, because their work is recognizable and valuable. You also hire them based on prior experience working with them, how well you can communicate your ideas to them, and so on. You can pretend they're replaceable if you want, which is partly true -- there are always other design companies you can go to -- but you certainly don't think of them as cogs in an assembly line.
You sure as hell don't try to design your process so you can replace a single artist at any time.
However, ideas are valuable. I can't speak for other programmers, but I'm absolutely lost on the business side of things. From my perspective, sales, marketing, ideas, and so on are just some of the things I'm very glad other people do, all as part of the Development Abstraction Layer. I'm hopeless without them, to the point where on one-man projects, I usually end up asking every prospective customer, investor, or just friends and family, for advice on things like naming a price.
I'm not sure how I feel about idea people learning to program. They try anyway, with spreadsheets. Sometimes it ends well, but often it ends in disaster. It's usually not a good idea to hire a dedicated full-time programmer to work on spreadsheets, and the whole point of spreadsheets is to enable end-users to do these things. Still, a few basic programming concepts would go a long way, even if they are in spreadsheets.
(No, I don't mean VBA. Either program or don't, but to half-ass it by crawling up out of excel into VBA is only going to end in tears.)
And I do like to think I'm working on something really cool. I certainly want my "code poetry" to have a point. It's not that I can't appreciate idea people or their ideas, it's that I'm not much of an idea person myself -- or at least, my ideas don't tend to be the sort that are likely to make me money.
Why would you assume I'm "worked up"?
I dismissed your link because you did seem to be assuming quite a bit, and at least one of your assumptions was incorrect -- "restrict that account to pure scp transfer".
To be fair, Chrome has always compiled JavaScript to native code, in a process similar to what modern JVMs do.
Nope, they're using it for rsync at least.
And why would you assume I'm on CentOS, or anything RedHat-derived?
Who are you replying to?
-1 Offtopic.
The fact that you used a TLA makes me suspect astroturf. RIA? What?
Nothing wrong with Rich Internet Applications per se, so long as they're actually based on standard technologies. To base them on Silverlight, which has exactly one decent implementation with exactly two supported platforms (Windows and Mac), is worse than basing them on Flash, which has exactly one decent implementation and supports Windows, Mac, and Linux. If I include the mobile space, it really doesn't change much -- pretty much everyone except Apple supports Flash and is proud of it, and pretty much only Windows Mobile supports anything like Silverlight.
So it's not that they're coming out of Redmond, it's that they fundamentally make certain websites Windows-only, or Windows/Mac only. Which means if you're truly an "avid Linux user," you're likely to run into a number of sites which don't work because they use Silverlight. I'm aware of Moonlight, and it just isn't there yet -- it's roughly like suggesting we all adopt Windows EXEs as a common execution platform because Wine exists.
Oh, and it also solves what's essentially a non-problem. HTML5 is good enough for the vast majority of what anyone wants to do with the Web. Extending it in small ways might be nice, and that's what people are doing (WebGL, etc.), but dropping a giant inner platform into it just makes things worse for everyone.
Apple has done more than anyone to standardize the Web, to make it multiple vendor safe and consumer friendly.
Yeah, I'll buy that when they natively support WebM or Theora/Vorbis.
They've done exactly what's in their best interest, and stopped just short of actually having to take the slightest amount of risk (perceived or real) to support the platform. Had they been willing to support any modern open video format, the whole codec war would've been over instantly.
Ok, let's compare, to the extent such comparisons are possible...
A modern system is 1-3 ghz.
Divide by 12 mhz. It's roughly one to three hundred times faster.
Assume it needed the full speed of the 386. Then 1-3% is exactly what you'd expect it to take on a modern CPU.
In other words, 1-3% is actually ridiculously low for something written in JavaScript on a modern system, compared to the hand-optimized C or ASM that would've been used to power the 386-based emulator.
When a 12mhz 386 can do fullspeed and it takes up to 5% on a recent machine, something is still deeply wrong.
Actually, regardless of the platform, if I can do something with 5% CPU that people actually want done, where's my incentive to make it take 0.5% CPU? There comes a point where ease/speed of development, readability, maintainability, stability, and security trump raw CPU usage.
It was tk last I checked. Has that changed?
I mean, giving Wikileaks an ssh account (as they're asking for) is pretty stupid, security-wise, but it's nowhere near as bad as giving _everyone_ your password by using FTP. You'd think Firesheep would've taught people something...
It also requires me to give them ssh access. No thanks.
Lower the barrier of entry even further, and just throw up a torrent or ten of static files which can be hosted anywhere, without fear of compromising your own server.
You can use document.write to output more JS, that will then be interpreted after the current script block.
And how's document.write any harder to intercept than eval?
As many layers as you like can be used.
I'm not sure how adding an extra layer makes it harder, either.
In what way is Java more sandboxed than JavaScript? Both run in a VM with a limited, controlled API to access the rest of the world. What makes you think exposing the DOM to Java would make it any more secure than exposing it to JavaScript?
That doesn't seem to work in this case -- I've tried Kate.
What I don't get is when one of the buffers is clearly empty -- it's not pasting garbage, it's pasting nothing -- why wouldn't it check another? And why is this even an issue? On my system, I use Firefox or Chrome, depending on the situation. Both use GTK+. Why would this bug only affect Chrome, and only on certain websites, under certain conditions? Why do I never see this issue anywhere else? Why will it paste into other locations in Chrome, like the address bar or even other textboxes, just not a Slashdot comment box?
Even if it's X-related, the fact that other browsers get it right suggests that it's Chrome's bug, so again, where's the bug we can all vote up till someone actually fixes it?
This has been annoying me for awhile now. Where's a bug we can all vote for and Slashdot?
Because the alternative, at least in Chrome, is obnoxious -- it doesn't have the option to download something to a temp folder and open it. Instead, if it sees anything it can't handle itself, either internally or with plugins, it's going to download it, automatically and instantly, cluttering my download folder to hell. It's especially bad when people put PDFs in iframes...
Acrobat actually isn't that bad, it's mostly the startup time that's a killer, but this is an improvement all around.
True, but pagination is alive and well in the normal Web search.
Actually, that is a security risk, currently mitigated on modern systems by not giving the root account a password, and only allowing root access through tools like sudo -- so in order to root a system, first you need to figure out which account is mine, only then can you break it.
That's moronic.
First, you've got a purpose-built application. What's stopping you from implementing whatever protocol you want? Why on earth would you use FTP?
Second, IP-based security isn't security.
Third, why would you re-implement encryption, authentication, and authorization in your application when you can get them for free from FTPS, SFTP, HTTPS, etc?
Here, try this: "Oscar Meyer wiener". That song just came on in your head, didn't it?
Nope, actually took a bit of thought to remember it.
Know why? Because I avoid advertising. I buy or rent DVDs, or I torrent, but I don't watch network TV. I do watch YouTube, but I avoid channels that allow video ads. I don't use a preconfigured adblock, but I do block annoying ads, and I tend to avoid sites that have interstitials.
So, force me to wait five seconds, and you won't make me memorize your little jingle or slogan. You'll make me go out of my way to avoid seeing your ad, which is likely going to end worse for you than if you just let me skip it.