I can't imagine it's any easier to get a new plugin approved than a complete new browser. It shouldn't be, since most programs' plugins can execute arbitrary code.
You thought I was going to say Linux. Seriously, interface isn't what made FF popular. If it was that simple, Opera would have crushed FF a long time ago.
No, Opera failed (in that it hasn't conquered the internet) because of the ads/paying. That was all. FF is popular because it's an IE alternative and there's an insane level of fanboyism for it, particularly here. Partly it's that firefox is in many ways the open source poster child. Everyone and their dog is switching their grandma to firefox to get away from IE, wheras people didn't do that with Opera.
It's customizability. One of Dreamweaver's most competitive features is the ability to write plugins for it in known languages -- JavaScript and XML. Guess what FF plugins are written in.
Customisability isn't what makes the difference, if it was winamp would still be king of the media players. "You can get that through plugins" all too often becomes an excuse for not including features. You complain that you can't even drag tabs around in firefox and the response is "get tabbrowser extensions". And extensions have been responsible for most of the publicised flaws I've seen in firefox. A program which is designed to deal with remote data should not be running complete programming languages, and even sandboxed execution should be avoided whenever possible. Anything less is a security nightmare.
A product with a locked featureset is going to suffer against a product you can upgrade piece by piece for free.
New features are what new versions are for. Opera 8 added integrated bittorrent (I think). And you don't need to hunt down new versions of your 600 plugins when you upgrade.
There are people out there who hate this, but they're slowly figuring out they're the minority.
I don't care if I'm the last person thinking this, a product with more features is still better than one with less even if the one with less supports plugins. If I really need a feature I'll write the code to do it myself.
But even if FF is losing momentum, it's still going to be with us for some time -- or another cross-platform open-source alternative.
Hopefully it will be another cross-platform open-source alternative - one that's better.
Faster startup and rendering. It's not worth it to me - better to render it slowly and correctly than fast and wrong, and I won't even start on security - but firefox can be a real dog on older machines.
So, it's actually just another alternative IE shell, with the slight difference that it's loaded by the IE executable rather than being a separate executable and using the IE rendering dll. Tabbed browsing etc. can be had in plenty of such shells, e.g. slimbrowser, and most will include a "firewall" or similar, not that they're usually effective at stopping all the malware IE is all too happy to run but it's better than nothing if the user is an idiot. It doesn't fix the most fundamental problem with IE, its broken nonstandard rendering.
OT but how does he get modded up for this? It's not original, there's even someone saying it in the next post above him, it's not insightful, it sounds like it's being repeated by rote.
2) The degree of fabrication could and IMHO should only be a factor in determining restitution. It has nothing to do with determing the validity of the litigation. Even if the blogger could show he/she truly believed the comments were accurate and took every possible action to verify the statement, that does not change the fundamental fact that the statement is WRONG.
So a blogger can never publish a negative statement about anyone? Because that seems to me to be the result of your viewpoint. Someone acting in good faith for the public interest should not be breaking the law.
3) Please define "reasonable person".
Uh, it has the same meaning as in all the other parts of the law it's used. It's not a new concept he's introducing.
Correct what? The fact that other countries are full of hackers that constantly attack you and you have little recourse to stop it? I suggest blocking them. Duh.
I'd suggest just keeping your services secure. Automated attacks are aimed at the lowest common denominator, even basic security steps will stop them. My smb server gets connect attempts at a rate of around 2 per second, and has done for the last six months or so. So far none have got in. I only take action if I'm getting hammered by a single IP, and then I'm more likely to complain to his ISP than block him.
I have an extensive ban list on my firewall including tons of/8 and/16's but mostly/24's. If someone cannot e-mail me it's because they are likely using a residential cable/DSL account
As well they should. The internet should be a community, not controlled by big corporations like other media.
and I suggest to them to either use AIM or a viable webmail service like GMail (hotmail and yahoo are banned).
Ooh, because an attacker is obviously so much less likely to use GMail than hotmail. After all, it's made by the holy Google who say "Do no evil", and everyone knows MS are always evil.
I am an individual. I don't run a corporate network and I am not required to put up w/a bunch of shit from other people.
If you want to be a part of the internet rather than a passive consumer of it, you should let everyone access what you're serving. Anything less is worse than nothing at all.
If it were up to me, I would want entire countries in their own easy to block IP address space. Want to block.br? Here's the single block that does it. Want to block.kr,.cn, and.nz? Go for it. Right now it's entirely too difficult and it requires some real work to do what you need to do.
Why do you want to block entire countries? Assuming Brazilians are evil because one tried to hack you is pure prejudice and as bad as any other kind.
Firewalls of any sort are a menace. They're not part of the open internet. Every port of every publicly routable IP should either be open, because it's providing a service accessible from the open internet, or closed, in which case it should respond appropriately when it gets packets there and not just drop them. I don't actively block them, but I try to avoid enabling any options on my services that would help firewalled users.
I want everything to be free software to the driver level too. Its not all that long ago that nVidia released a Linux driver that broke old and low-end cards and didn't bother to release a fix for months.
They've fixed it? I'm finding that very issue hugely annoying, and I didn't see any fix in the forum.
Just like discussions of Linux sound server issues underscore that the real problem is that it's insane that the user of a desktop operating system ever encounters something called a "sound servers"...
The user doesn't have to, whatever their distro shipped with will do fine. The fact that you have a choice and can change it if you want to is the great thing about linux.
I think OSS has a low view of "usability" from experience. Gnome recently had a lot of usability work done on it. The noticable consequences of this were:
Reversing the button order of the confirmation dialog, confusing everyone. Eventually they were forced to put it back.
Changing the file manager viewing mode to a different, arguably superior or inferior but certainly requiring adjusting to, one, with no option to change it back. Eventually they were forced to make it an option.
Shrinking the save dialog, and making it impossible to resize (so it looks horribly tiny on my 1600x1200), and making you do an extra click to get the familiar browse window, again with no option to change it back. I can only hope they are forced to change this.
If this is the effect a usability focus has on a desktop environment, I hope KDE never pays any attention to it.
The point of the article is that peer review fails because reviewers approve papers that agree with their personal biases, rather than those with good science.
The problem some people have with the GPL is that lots of things that really don't need to be GPL'd are, and are less useful as a result. For example, why is readline GPL? This means people making software with GPL incompatable licenses (even if they are open source) cannot use it. Is it really a justifiable concern that someone might use readline to make a commercial, non-free version of readline? There's no incentive to do that, and even if someone did, the open source version would still be there and just as good. So if readline were licensed under a more free license like a BSD or MIT or ISC license, or even the LGPL it would be a more useful piece of software for more people.
There is at least one program which is open source solely because readline is GPL (Stallman or someone else at the FSF sent a C&D to someone who was using it in propriety software, who responded by making their software GPL). So by being GPL it makes the software world as a whole better for everyone.
Furthermore, if RMS wants people to be remembered, he should fight for everyone to be remembered, whether he agrees with their development model or not.
The right to have the OS that you fricking wrote called anything that you like is so entirely obvious it doesn't need fighting for. Except, apparently, in this case.
some GNU were indeed there before linux but i don't think they had a runnable totally free system before linux existed due to thier screwing about with microkernels.
True, but linux was just the last piece added to the jigsaw, it's not any more important than the rest of the system.
true but only because GCC happened to be one of the best C compilers arround when they were developed.
Were there any others that Linus could have afforded?
however i've never heared of a system that uses the gnu userland as its main userland, doesn't use the linux kernel, and that actually has reached the point of stable releases that are usable. would you like to name one?
A friend of mine had a file server running on Debian GNU/kFreeBSD for a while. The release isn't called stable but given that this is Debian that's understandable. It may not support as much hardware as linux and it doesn't support anything like the complete debian package tree, but it's perfectly usable.
You can enable it, yes, but by default it isn't possible. In the default Ubuntu setup everyone can do admin tasks with their ordinary account, just like windows. It is of course a matter of personal preference and situation-dependent as to whether or not you think that's a more secure situation.
I didn't notice any security with bittorrent, with FTP there is some, ineffective as it may be. For a public filesserver it doesn't matter.
it's difficult to get working if either the client or server are behind a firewall or NAT,
That's their own fault, but anyway it's not difficult if the client's behind either. They need to use passive mode, yes, but that's the default in most clients anyway, and once they have passive mode selected they get full speed transfers. As for the server, of course you will have problems if the server's behind a firewall, that's true for any protocol. A BT tracker would be the same.
and the fact that it differentiates between ASCII and binary files is stupid.
The C programming language does the same. The fact that the line terminator is different on different systems is what's stupid. If it's a problem, just set all files as binary, but then your editor will have to handle foreign files.
SFTP is an ideal replacement for it for normal system administration tasks,
Agreed
and for mass distribution of files I'd rather use HTTP, BitTorrent, eDonkey, or any other number of file sharing protocols.
Why? Security is irrelevant for publicly distributed files, you'll have problems with any protocol if your server's firewalled/NATed, and the ascii/binary distinction is handled automatically by any client worth its salt and ensures people can read text files they download.
While BT isn't perfect in that regard, it is better. FTP uses two ports, 21 and 20; 21 is for control, 20 is for data transfers. At the very least, port 21 has to be forwarded from the firewall to the server. Furthermore, there are two different ways of activating the data connection, active and passive; "active" involves the FTP server initiating the connection to the client's data port, while "passive" involves the client initiating the connection to the server. If you want to use "active" connections, you have to have port 20 forwarded to the client; if you want to use "passive" connections, you have to have port 20 forwarded to the server
Servers always need ports forwarded - an apache server needs port 80 and possibly 443, a BT tracker needs ports forwarded to it.
Depending on what FTP client you use, things can be further complicated; you never know whether active or passive will be the default.
That's a flaw in those clients, not the protocol.
Furthermore, you never know whether the client or server will report their internet IP or their internal network IP;
No, but again that's the case for any server, apache needs to be told its (external) hostname.
so, for example, if you're a client behind a firewall and you want to use an active connection,
That's a very contrived situation, the whole point is that you use passive from behind firewalls.
On the other hand, with BT, the only port forwarding that's actually required is if you have to forward a port to the tracker.
And the only port forwarding that's actually required with FTP is two ports to the server. OK it's 2 rather than 1, but it's not that much harder.
Clients can connect and transfer files without forwarding any ports at all,
True, but the protocol can only sustain a minority of clients doing this.
although you will generally get faster speeds if you have a port forwarded to your computer.
Wheras with FTP you can get full speed connections without having to forward a port. Sounds like a win for FTP to me.
I don't personally think eDonkey is a good solution either though. Poisoned files are too rampant these days to make these types of services viable anymore. What we need is something analogous to IRC where all but the higher intelligence/true geeks can do their thing without it being flooded by idiots and junk. Also some sort of file verification system is needed beyond a user rating system that is pretty fool-proof and solid to keep things legit and flowing smoothly.
I think gnunet is the network of the future. All encrypted, it can be purely F2F, a good reward system, and integrated digital signing of content. Now if only more people would use it. Hopefully the new version will have a nicer GUI and be more usable.
You are stack with the old process, you need to understand that 2.6 vanilla aren't considered stable, in the same way 2.4 and 2.5 were.
2.5 wasn't stable, it's a development branch. Linus has officially declared that 2.6 is stable - if it wasn't, it should have remained 2.5 - and what we should use, and 2.4 is the legacy branch.
Now, why do you want so much to upgrade to 2.6 if you are so happy with 2.4 kernel?
I don't, but my distro no longer supports using 2.4.
I really can't see what is your point.
2.6 is officially the stable branch of the kernel, and distributions are flamed if they do not switch to it. As such, it should actually be stable.
Slashdot Mirror
I can't imagine it's any easier to get a new plugin approved than a complete new browser. It shouldn't be, since most programs' plugins can execute arbitrary code.
Everyone is "weak willed" at that age, they don't know any better. You can't blame elementary school kids for being impressionable, they're kids.
No, Opera failed (in that it hasn't conquered the internet) because of the ads/paying. That was all. FF is popular because it's an IE alternative and there's an insane level of fanboyism for it, particularly here. Partly it's that firefox is in many ways the open source poster child. Everyone and their dog is switching their grandma to firefox to get away from IE, wheras people didn't do that with Opera.
It's customizability. One of Dreamweaver's most competitive features is the ability to write plugins for it in known languages -- JavaScript and XML. Guess what FF plugins are written in.
Customisability isn't what makes the difference, if it was winamp would still be king of the media players. "You can get that through plugins" all too often becomes an excuse for not including features. You complain that you can't even drag tabs around in firefox and the response is "get tabbrowser extensions". And extensions have been responsible for most of the publicised flaws I've seen in firefox. A program which is designed to deal with remote data should not be running complete programming languages, and even sandboxed execution should be avoided whenever possible. Anything less is a security nightmare.
A product with a locked featureset is going to suffer against a product you can upgrade piece by piece for free.
New features are what new versions are for. Opera 8 added integrated bittorrent (I think). And you don't need to hunt down new versions of your 600 plugins when you upgrade.
There are people out there who hate this, but they're slowly figuring out they're the minority.
I don't care if I'm the last person thinking this, a product with more features is still better than one with less even if the one with less supports plugins. If I really need a feature I'll write the code to do it myself.
But even if FF is losing momentum, it's still going to be with us for some time -- or another cross-platform open-source alternative.
Hopefully it will be another cross-platform open-source alternative - one that's better.
Faster startup and rendering. It's not worth it to me - better to render it slowly and correctly than fast and wrong, and I won't even start on security - but firefox can be a real dog on older machines.
So, it's actually just another alternative IE shell, with the slight difference that it's loaded by the IE executable rather than being a separate executable and using the IE rendering dll. Tabbed browsing etc. can be had in plenty of such shells, e.g. slimbrowser, and most will include a "firewall" or similar, not that they're usually effective at stopping all the malware IE is all too happy to run but it's better than nothing if the user is an idiot. It doesn't fix the most fundamental problem with IE, its broken nonstandard rendering.
OT but how does he get modded up for this? It's not original, there's even someone saying it in the next post above him, it's not insightful, it sounds like it's being repeated by rote.
So a blogger can never publish a negative statement about anyone? Because that seems to me to be the result of your viewpoint. Someone acting in good faith for the public interest should not be breaking the law.
3) Please define "reasonable person".
Uh, it has the same meaning as in all the other parts of the law it's used. It's not a new concept he's introducing.
I'd suggest just keeping your services secure. Automated attacks are aimed at the lowest common denominator, even basic security steps will stop them. My smb server gets connect attempts at a rate of around 2 per second, and has done for the last six months or so. So far none have got in. I only take action if I'm getting hammered by a single IP, and then I'm more likely to complain to his ISP than block him.
I have an extensive ban list on my firewall including tons of /8 and /16's but mostly /24's. If someone cannot e-mail me it's because they are likely using a residential cable/DSL account
As well they should. The internet should be a community, not controlled by big corporations like other media.
and I suggest to them to either use AIM or a viable webmail service like GMail (hotmail and yahoo are banned).
Ooh, because an attacker is obviously so much less likely to use GMail than hotmail. After all, it's made by the holy Google who say "Do no evil", and everyone knows MS are always evil.
I am an individual. I don't run a corporate network and I am not required to put up w/a bunch of shit from other people.
If you want to be a part of the internet rather than a passive consumer of it, you should let everyone access what you're serving. Anything less is worse than nothing at all.
If it were up to me, I would want entire countries in their own easy to block IP address space. Want to block .br? Here's the single block that does it. Want to block .kr, .cn, and .nz? Go for it. Right now it's entirely too difficult and it requires some real work to do what you need to do.
Why do you want to block entire countries? Assuming Brazilians are evil because one tried to hack you is pure prejudice and as bad as any other kind.
Firewalls of any sort are a menace. They're not part of the open internet. Every port of every publicly routable IP should either be open, because it's providing a service accessible from the open internet, or closed, in which case it should respond appropriately when it gets packets there and not just drop them. I don't actively block them, but I try to avoid enabling any options on my services that would help firewalled users.
They've fixed it? I'm finding that very issue hugely annoying, and I didn't see any fix in the forum.
The user doesn't have to, whatever their distro shipped with will do fine. The fact that you have a choice and can change it if you want to is the great thing about linux.
If this is the effect a usability focus has on a desktop environment, I hope KDE never pays any attention to it.
That's not SIP sucking, it's NAT and firewalls sucking.
The point of the article is that peer review fails because reviewers approve papers that agree with their personal biases, rather than those with good science.
That's not about what you're allowed to call it, it's about what you should call it. Different things.
There is at least one program which is open source solely because readline is GPL (Stallman or someone else at the FSF sent a C&D to someone who was using it in propriety software, who responded by making their software GPL). So by being GPL it makes the software world as a whole better for everyone.
The right to have the OS that you fricking wrote called anything that you like is so entirely obvious it doesn't need fighting for. Except, apparently, in this case.
True, but linux was just the last piece added to the jigsaw, it's not any more important than the rest of the system.
true but only because GCC happened to be one of the best C compilers arround when they were developed.
Were there any others that Linus could have afforded?
however i've never heared of a system that uses the gnu userland as its main userland, doesn't use the linux kernel, and that actually has reached the point of stable releases that are usable. would you like to name one?
A friend of mine had a file server running on Debian GNU/kFreeBSD for a while. The release isn't called stable but given that this is Debian that's understandable. It may not support as much hardware as linux and it doesn't support anything like the complete debian package tree, but it's perfectly usable.
Patents take a while to go through the system (A post below says this one was filed in 2000)
No. The whole point of the patent system is to encourage people to publish.
You can enable it, yes, but by default it isn't possible. In the default Ubuntu setup everyone can do admin tasks with their ordinary account, just like windows. It is of course a matter of personal preference and situation-dependent as to whether or not you think that's a more secure situation.
I didn't notice any security with bittorrent, with FTP there is some, ineffective as it may be. For a public filesserver it doesn't matter.
it's difficult to get working if either the client or server are behind a firewall or NAT,
That's their own fault, but anyway it's not difficult if the client's behind either. They need to use passive mode, yes, but that's the default in most clients anyway, and once they have passive mode selected they get full speed transfers. As for the server, of course you will have problems if the server's behind a firewall, that's true for any protocol. A BT tracker would be the same.
and the fact that it differentiates between ASCII and binary files is stupid.
The C programming language does the same. The fact that the line terminator is different on different systems is what's stupid. If it's a problem, just set all files as binary, but then your editor will have to handle foreign files.
SFTP is an ideal replacement for it for normal system administration tasks,
Agreed
and for mass distribution of files I'd rather use HTTP, BitTorrent, eDonkey, or any other number of file sharing protocols.
Why? Security is irrelevant for publicly distributed files, you'll have problems with any protocol if your server's firewalled/NATed, and the ascii/binary distinction is handled automatically by any client worth its salt and ensures people can read text files they download.
Servers always need ports forwarded - an apache server needs port 80 and possibly 443, a BT tracker needs ports forwarded to it.
Depending on what FTP client you use, things can be further complicated; you never know whether active or passive will be the default.
That's a flaw in those clients, not the protocol.
Furthermore, you never know whether the client or server will report their internet IP or their internal network IP;
No, but again that's the case for any server, apache needs to be told its (external) hostname.
so, for example, if you're a client behind a firewall and you want to use an active connection,
That's a very contrived situation, the whole point is that you use passive from behind firewalls.
On the other hand, with BT, the only port forwarding that's actually required is if you have to forward a port to the tracker.
And the only port forwarding that's actually required with FTP is two ports to the server. OK it's 2 rather than 1, but it's not that much harder.
Clients can connect and transfer files without forwarding any ports at all,
True, but the protocol can only sustain a minority of clients doing this.
although you will generally get faster speeds if you have a port forwarded to your computer.
Wheras with FTP you can get full speed connections without having to forward a port. Sounds like a win for FTP to me.
I think gnunet is the network of the future. All encrypted, it can be purely F2F, a good reward system, and integrated digital signing of content. Now if only more people would use it. Hopefully the new version will have a nicer GUI and be more usable.
2.5 wasn't stable, it's a development branch. Linus has officially declared that 2.6 is stable - if it wasn't, it should have remained 2.5 - and what we should use, and 2.4 is the legacy branch.
Now, why do you want so much to upgrade to 2.6 if you are so happy with 2.4 kernel?
I don't, but my distro no longer supports using 2.4.
I really can't see what is your point.
2.6 is officially the stable branch of the kernel, and distributions are flamed if they do not switch to it. As such, it should actually be stable.