People have done things like adding randomness via microphone noise, but I'm not really sure how reliable that is.
The rest of it either isn't necessarily random, or isn't necessarily cheap enough / fast enough. And PRNGs can be made hard enough to guess that no one will. It's kind of like how RSA is possible to crack, if someone guesses the right prime factors, but with a sufficiently large key size, you can get to where all of the matter in the Universe, assembled into chips that vaguely resemble today's processors, will not be able to guess it before the heat death of the Universe.
Of course, if quantum computers ever scale, they change the math on that and pretty much demolish RSA, but they might also make true randomness easier.
And by the way -- IANAC -- I Am Not A Cryptographer.
Indeed, that is one of the things I love about Linux.
But I don't think for a second that it has anything to do with the software being open. After all, the same mechanism delivers updates to my proprietary nvidia drivers.
In my case, because I didn't know about xrandr, originally did this on OS X, and I really only want to rotate that one book. I'd rather not have to rearrange the rest of my windows.
And I didn't use the cursor, I used spacebar, I think. Nice, big target.
Hardly a Microsoft invention. Of course, one of the two I mentioned might actually run on XP -- or you could switch all the client OSes to virtual machines and run Coda on Linux.
Oh, I see... is the issue that people are running older versions of Acrobat?
If they can't be bothered to upgrade to the latest version, what makes you think they'll patch themselves? Are you suggesting that the big advantage of me running Free Software here is that I could be running kpdf 0.2 and patch the security holes? Or are you suggesting that someone who can't be bothered to update their software is going to have a better time of it on Linux, for which I've never seen a built-in, GUI way to force auto-updates?
Of course, if you were going to suggest that Free Software doesn't have security bugs, I'd really have to laugh in your face...
It is possible to write provably safe apps. As in, mathematical proofs.
In fact, there is a company which specializes in writing damn-near absolutely safe, bug-free apps. They do it in about as much time as the competition writes buggy, insecure apps, because the lack of bugs in the first place means less of a debugging cycle. They charge about twice as much, because very few other companies provide that much quality.
The main reason I have acroread is because I can -- it's one less program people can whine about not having on Linux, and you never know when I'll run into something kpdf can't handle.
But I also have it because it has one feature I dearly wish kpdf did: the ability to rotate the rendered PDF. Take a widescreen, clamshell laptop/notebook, turn it on its side, and let a page of a book fill the screen, and you have a pretty nice eBook reader.
The W3C may still be living in the era when all websites were looked after directly by an admin, who did everything to do with the site, including write the software. They might share software, just as they might share HTML, but for many of these sites, best practices, or even common sense practices, were completely gone -- so no formal "libraries".
It's not just web pages. That's not all the W3C does.
In fact, they cite things that IE doesn't support natively, like SVG -- unless something's changed.
But before you go blaming MS, try gathering some evidence, maybe? Shouldn't be hard -- you do have access to a machine with IE, right? And Wireshark is easy to find...
It would help if you were more clear about exactly what you need.
I've been pretty much buried in Amazon's AWS stack for awhile now, so a part of me wants to say... If you just need storage, go with S3. It's pay-as-you-go, absurdly simple file storage that happens to be done over http(s). Can be private, or can be on the public Internet.
It's the 30 mhz figure that makes me wonder. If you need to do a batch job very occasionally, and you don't care how long you wait, you could always prepare an ec2 image, and bring the instance up on demand. Costs 10 cents/hour, which means it's a bit expensive to leave on all the time, but if you only leave it up for an hour a day, that's $3/mo. When you shut it down, all data is lost... unless you persisted it to S3.
Or, yeah, what other people are suggesting, if you're wanting a mailserver or something.
Have you considered Amazon S3? I assume your download isn't dynamically generated...
Some quick calculations: Say you upload one version a month, and you might see 50 downloads on a busy day. That's:
$0.10 per gig uploaded * 0.2 gigs = 2 cents $0.01 per 1k PUT requests means it will cost you a cent after 1k months = 83 years $0.15 per gig stored. To make it easy, assume you have five versions archived at any given time: 15 cents per month. $0.01 per 10k GET and HEAD requests means, assuming they HEAD before a GET (for some reason), it costs a cent after 8.3 years $0.18 per gig for the first 10 TB/mo. 50*200 megs = 10 gigs (roughly), so you aren't going over that, so this costs you $1.80.
Adding it all up, this will cost you a little less than $2/mo. You might need additional hosting (and EC2 might cost a bit much) if there are any dynamic parts of your website, but at $2/mo, you can afford to host the big files.
The downside might be that it scales too well for you -- if you get Slashdotted, and you get hit with, say, a thousand unique users, it's going to cost about $35 for that month. (Most of us consider that a nice problem to have.) A million unique users means $35,000, which is going to hurt -- though of course, if your game gets downloaded a million times, you may be able to sell the concept for a few millions yourself. It does support BitTorrent, though, so if you catch your Slashdotting in time, that might save some money.
- the inability to run native windows apps such as photoshop and dreamweaver (no, they don't really work with wine or cedega)
And I can't run KDE on Windows. News flash: Native apps for $OS_A don't work on $OS_B.
- The inability to use suspend/hibernate when you close the lid on the notebook.
Had this in 2003.
"native" wireless drivers for linux mostly suck
Out of the box, on this Toshiba laptop. Automatically scans, builds a list of wireless networks. I can join an open one in two clicks, if it doesn't do that already. And unlike Windows, it won't bug me about wireless when I'm plugged in to Ethernet.
if you're lucky, you can view divx files and play MP3s.
And h.264, and vc1, and wmv9, and flv, and ac3, and... Hell, even codecs I've never heard of, like cinepack.
Added Medibuntu to my Linux. Took maybe two minutes. And I'm on a 64-bit distro, which means I can't easily fall back on win32codecs, but I haven't needed to.
I can't really run a scanner properly (my brother just got an HP F4180 that supports hplip under Linux, but barely).
In other words, it works. What's this "barely"?
For me, it's been mostly hardware support, but even here, the worst I've got now is my laptop's built-in sound refusing to work. Oh well, I was looking for an excuse to get a nice USB-audio device. Of course, YMMV, I realize buying hardware is not a fair thing to ask of everyone.
Paying more for certain connections to receive higher QoS is also OK.... ISP to ISP interconnections can be negotiated based upon total bandwidth, packet volume (even tiered by QoS). That,s OK. But there should be NO pricing based upon identities of each ISP's customers at such interfaces.
Hmm.
What I want to say is, even limiting the number of connections, and particularly limiting the type of connections, could be used for anticompetitive purposes. Example: Limit the customer to, say, 5-10 open TCP connections, and you absolutely kill BitTorrent, without significantly impacting (most) web browsing. And certainly, you could detect things like Skype and single them out individually, blocking that protocol.
Then I remembered that even limiting bandwidth limits possible uses. Limit bandwidth to some 50 or 100 megs a day, and YouTube becomes difficult, let alone any serious BitTorrenting (even legit things like Vuze).
At this point, I'm thinking that ISPs should block certain ports by default, but make it easy for consumers to unblock them, and be allowed to limit the number of concurrent connections, but to something reasonable. But it is getting a little difficult to draw that line as precisely as we want. It's obvious that a 100mbit pipe which I can saturate any way I want is neutral -- and I'm about to have something like that. It's similarly obvious that deliberately blocking a certain protocol (BitTorrent) or certain endpoints (The Pirate Bay) is right out.
But between those is a gray area that I'm not really sure how to define. I wish all ISPs were like mine, but that's not likely to happen.
Transparent proxies are generally a bad idea, I've found.
But yes, redirection, if that technically counts as NAT (as it's done on the iptables "nat" table), is useful. Still, most places you see it used is for port forwarding, which is an ugly hack to make NAT not as painful.
Were the summary written "EFF files lawsuit against data searches", it would make sense, but the summary freely admits that this has been happening for years. I could swear I've seen it discussed on Slashdot before.
of what use is a Maglev is some asshat blows it the hell up?
There's always the possibility that spending money on a Maglev, rather than on killing people, might make it less likely that someone would want to blow it the hell up.
In any case, most of the money being spent in the military has nothing to do with terrorism. In particular, the Iraq war has nothing to do with terrorism, and building up the armed forces has nothing to do with terrorism. Really, about the only thing that would make sense here is spending more money on intelligence.
Mostly because most browsers really don't like it when you leave an HTTP connection open for too long, even if you set them not to timeout.
Google's actually come up with a neat hack to deal with this -- leave the connection open for 30 seconds, and if nothing new comes down it, close that connection and open a new one. Technically "polling", but practically just as fast as XMPP.
Then, when you consider spyware, even pinning it to a particular house or account goes out the window.
But of course, the real problem is that they obviously don't even go that far, most of the time, or how do you explain the lawsuits against people who are mentally or physically incapable of filesharing?
Slashdot Mirror
Nope, it'd be a nag-you security update. No matter how easy, you're still going to have users who just refuse to click on the thing.
Using Kubuntu, and I've used (and developed for) Gentoo. I know how package managers work.
People have done things like adding randomness via microphone noise, but I'm not really sure how reliable that is.
The rest of it either isn't necessarily random, or isn't necessarily cheap enough / fast enough. And PRNGs can be made hard enough to guess that no one will. It's kind of like how RSA is possible to crack, if someone guesses the right prime factors, but with a sufficiently large key size, you can get to where all of the matter in the Universe, assembled into chips that vaguely resemble today's processors, will not be able to guess it before the heat death of the Universe.
Of course, if quantum computers ever scale, they change the math on that and pretty much demolish RSA, but they might also make true randomness easier.
And by the way -- IANAC -- I Am Not A Cryptographer.
Actually, now I see my mistake...
I was assuming 50 downloads on the same day, and no downloads for the rest of the month, as that would be the day the new version was posted.
Yes, that's the nice thing about it being free as in beer.
Again, nothing to do with it being Free Software.
Indeed, that is one of the things I love about Linux.
But I don't think for a second that it has anything to do with the software being open. After all, the same mechanism delivers updates to my proprietary nvidia drivers.
I never did this with Gutenberg.
In my case, because I didn't know about xrandr, originally did this on OS X, and I really only want to rotate that one book. I'd rather not have to rearrange the rest of my windows.
And I didn't use the cursor, I used spacebar, I think. Nice, big target.
You're right. I did mean to say 50/month.
Hardly a Microsoft invention. Of course, one of the two I mentioned might actually run on XP -- or you could switch all the client OSes to virtual machines and run Coda on Linux.
*cough* *sputter* What?
Slashdotters always making me spill my coffee...
Oh, I see... is the issue that people are running older versions of Acrobat?
If they can't be bothered to upgrade to the latest version, what makes you think they'll patch themselves? Are you suggesting that the big advantage of me running Free Software here is that I could be running kpdf 0.2 and patch the security holes? Or are you suggesting that someone who can't be bothered to update their software is going to have a better time of it on Linux, for which I've never seen a built-in, GUI way to force auto-updates?
Of course, if you were going to suggest that Free Software doesn't have security bugs, I'd really have to laugh in your face...
It is possible to write provably safe apps. As in, mathematical proofs.
In fact, there is a company which specializes in writing damn-near absolutely safe, bug-free apps. They do it in about as much time as the competition writes buggy, insecure apps, because the lack of bugs in the first place means less of a debugging cycle. They charge about twice as much, because very few other companies provide that much quality.
Can't remember their name now, though.
Rather, both kpdf and acroread.
The main reason I have acroread is because I can -- it's one less program people can whine about not having on Linux, and you never know when I'll run into something kpdf can't handle.
But I also have it because it has one feature I dearly wish kpdf did: the ability to rotate the rendered PDF. Take a widescreen, clamshell laptop/notebook, turn it on its side, and let a page of a book fill the screen, and you have a pretty nice eBook reader.
The W3C may still be living in the era when all websites were looked after directly by an admin, who did everything to do with the site, including write the software. They might share software, just as they might share HTML, but for many of these sites, best practices, or even common sense practices, were completely gone -- so no formal "libraries".
Of course, I just completely made that up...
It's not just web pages. That's not all the W3C does.
In fact, they cite things that IE doesn't support natively, like SVG -- unless something's changed.
But before you go blaming MS, try gathering some evidence, maybe? Shouldn't be hard -- you do have access to a machine with IE, right? And Wireshark is easy to find...
Why can't the software cache it, then? Lazy developers...
I'm sure good ol HEAD, If-Modified-Since, and Etag/If-None-Match would be a LOT less bandwidth. Or are they getting that much in cache hits alone?
Could also refuse to serve it except in gzip format.
It would help if you were more clear about exactly what you need.
I've been pretty much buried in Amazon's AWS stack for awhile now, so a part of me wants to say... If you just need storage, go with S3. It's pay-as-you-go, absurdly simple file storage that happens to be done over http(s). Can be private, or can be on the public Internet.
It's the 30 mhz figure that makes me wonder. If you need to do a batch job very occasionally, and you don't care how long you wait, you could always prepare an ec2 image, and bring the instance up on demand. Costs 10 cents/hour, which means it's a bit expensive to leave on all the time, but if you only leave it up for an hour a day, that's $3/mo. When you shut it down, all data is lost... unless you persisted it to S3.
Or, yeah, what other people are suggesting, if you're wanting a mailserver or something.
Trying so hard not to be a fanboy, but...
Have you considered Amazon S3? I assume your download isn't dynamically generated...
Some quick calculations: Say you upload one version a month, and you might see 50 downloads on a busy day. That's:
$0.10 per gig uploaded * 0.2 gigs = 2 cents
$0.01 per 1k PUT requests means it will cost you a cent after 1k months = 83 years
$0.15 per gig stored. To make it easy, assume you have five versions archived at any given time: 15 cents per month.
$0.01 per 10k GET and HEAD requests means, assuming they HEAD before a GET (for some reason), it costs a cent after 8.3 years
$0.18 per gig for the first 10 TB/mo. 50*200 megs = 10 gigs (roughly), so you aren't going over that, so this costs you $1.80.
Adding it all up, this will cost you a little less than $2/mo. You might need additional hosting (and EC2 might cost a bit much) if there are any dynamic parts of your website, but at $2/mo, you can afford to host the big files.
The downside might be that it scales too well for you -- if you get Slashdotted, and you get hit with, say, a thousand unique users, it's going to cost about $35 for that month. (Most of us consider that a nice problem to have.) A million unique users means $35,000, which is going to hurt -- though of course, if your game gets downloaded a million times, you may be able to sell the concept for a few millions yourself. It does support BitTorrent, though, so if you catch your Slashdotting in time, that might save some money.
And I can't run KDE on Windows. News flash: Native apps for $OS_A don't work on $OS_B.
Had this in 2003.
Out of the box, on this Toshiba laptop. Automatically scans, builds a list of wireless networks. I can join an open one in two clicks, if it doesn't do that already. And unlike Windows, it won't bug me about wireless when I'm plugged in to Ethernet.
And h.264, and vc1, and wmv9, and flv, and ac3, and... Hell, even codecs I've never heard of, like cinepack.
Added Medibuntu to my Linux. Took maybe two minutes. And I'm on a 64-bit distro, which means I can't easily fall back on win32codecs, but I haven't needed to.
In other words, it works. What's this "barely"?
For me, it's been mostly hardware support, but even here, the worst I've got now is my laptop's built-in sound refusing to work. Oh well, I was looking for an excuse to get a nice USB-audio device. Of course, YMMV, I realize buying hardware is not a fair thing to ask of everyone.
Finally seems appropriate:
andnothingofvaluewaslost
Hmm.
What I want to say is, even limiting the number of connections, and particularly limiting the type of connections, could be used for anticompetitive purposes. Example: Limit the customer to, say, 5-10 open TCP connections, and you absolutely kill BitTorrent, without significantly impacting (most) web browsing. And certainly, you could detect things like Skype and single them out individually, blocking that protocol.
Then I remembered that even limiting bandwidth limits possible uses. Limit bandwidth to some 50 or 100 megs a day, and YouTube becomes difficult, let alone any serious BitTorrenting (even legit things like Vuze).
At this point, I'm thinking that ISPs should block certain ports by default, but make it easy for consumers to unblock them, and be allowed to limit the number of concurrent connections, but to something reasonable. But it is getting a little difficult to draw that line as precisely as we want. It's obvious that a 100mbit pipe which I can saturate any way I want is neutral -- and I'm about to have something like that. It's similarly obvious that deliberately blocking a certain protocol (BitTorrent) or certain endpoints (The Pirate Bay) is right out.
But between those is a gray area that I'm not really sure how to define. I wish all ISPs were like mine, but that's not likely to happen.
Been awhile, but I'm not sure you can rename them to a different directory in your checkout, only to a different name within the same directory.
Not worth booting Windows just to make sure I'm right, though.
Transparent proxies are generally a bad idea, I've found.
But yes, redirection, if that technically counts as NAT (as it's done on the iptables "nat" table), is useful. Still, most places you see it used is for port forwarding, which is an ugly hack to make NAT not as painful.
Were the summary written "EFF files lawsuit against data searches", it would make sense, but the summary freely admits that this has been happening for years. I could swear I've seen it discussed on Slashdot before.
There's always the possibility that spending money on a Maglev, rather than on killing people, might make it less likely that someone would want to blow it the hell up.
In any case, most of the money being spent in the military has nothing to do with terrorism. In particular, the Iraq war has nothing to do with terrorism, and building up the armed forces has nothing to do with terrorism. Really, about the only thing that would make sense here is spending more money on intelligence.
Mostly because most browsers really don't like it when you leave an HTTP connection open for too long, even if you set them not to timeout.
Google's actually come up with a neat hack to deal with this -- leave the connection open for 30 seconds, and if nothing new comes down it, close that connection and open a new one. Technically "polling", but practically just as fast as XMPP.
Then, when you consider spyware, even pinning it to a particular house or account goes out the window.
But of course, the real problem is that they obviously don't even go that far, most of the time, or how do you explain the lawsuits against people who are mentally or physically incapable of filesharing?