I believe RAR does what they call "solid archiving," which means that a single compression dictionary is created for the entire archive.
ZIP doesn't do that; each file in a ZIP archive is compressed individually, with a separate compression dictionary. That hurts the compression ratio for ZIP archives that contain many files, particularly many small files, particularly many similar small files, like source code, for example. But it does mean that archive operations (like extracting or updating individual files or and adding files to or removing files from an archive) are fast and simple.
It's optional. You can do solid archiving, but it's not the default, for the reasons you mention. The files were several megabytes each, and the zip performance so much worse that I doubt it could be made competitive.
Even with the quick launch option, mozilla and firefox still start much slower. And whilst I've seen many claims that firefox now renders faster, personal experience and the benchmarks I've seen don't agree with that.
I'm not being a crybaby, I'm complaining in the hope that it will change things. Insightful moderations should go to things which actually are insightful.
You're an idiot and a liar who always pops up in threads like this one. GNOME had a lot of usability work done on it, most of which resulted in a system that it light-years ahead of KDE in usability and consequently commercial acceptance.
I've *tried* to use gnome. It is far less usable. This is from simple experience.
What do you mean reversed? Ok/Cancel dialogs have been removed from GNOME... making this argument rather less important.
It may have been the yes/no dialog, someone decided the "positive" button should go on the left and the negative one on the right, rather than the normal way round. It was horrible.
Wrong... it was always an option -- just not one available directly from within Nautilus. P.S. I think Nautilus is a piece of shit.
In the first releases of gnome 2.6 it required a manual config file edit (or using the gnome registry editor). It wasn't a user-accessible option. An option was added pretty quick, certainly by 2.6.2, but there was no option in the first releases.
Now about that file dialog... it's quite clever actually. Simple, elegant and quite customizable if you want it.
How does one customize it? If I could just make it maximisable it would be 100x easier to use. Starting up in "opened out" mode would be a big help too.
Hmm, must be a different issue. My TNT2 is no longer supported in the new drivers. Worse, it happened in the very release that enabled xinerama and glx together. Ah well.
And 2.4 is not a legacy branch, it's still being developed and most things of 2.6 (like new hardware support) are being backported to 2.4
It's older than the current stable, it's legacy. 2.2 and 2.0 are still under development and occasionally get things backported
Both 2.4 kai 2.6 kernels are fully supported by my distros, non of those distros have a policy that forced me to upgrade to 2.6.
Those are relatively major distos. It's difficult for smaller distros to support two separate versions. Given that they had to choose one I can't blame mine for picking 2.6.
Actually most of the linux servers out there are still running on 2.4, and 2.6 is recommended mostly for desktop workstations. Switching to 2.6 is not the only option, since both can be supported. I can think of a reason for a distro to be flamed if it _does stop_ supporting 2.4, but not in the other way around.
Distros which didn't switch to 2.6 as default were flamed quite severely, here for starters. I remember a post along the lines of "why is slackware still using 2.4 as a default 3 months after 2.6 came out? Answer: because it's an obsolete distribution for obsolete people. It just isn't relevant today". I'm pretty sure I saw a comment from Linus along the lines of "2.6 is the stable tree, distributions should be using it, I can't do much more until distributions have it as their default."
When you say that distros are flamed for not switching, flamed by who? Maybe from people like you, eager to upgrade everything to the last version because they think that this is always the right thing to do.
When have I ever said anything like that? I was happy with 2.4, it worked, and I didn't switch to 2.6 until it became necessary.
Maybe you should stop blaming linus for the way he chooses to develop _his_ project, and start complaining to your distro maintainers for quiting on 2.4 and shiping with an unstable 2.6.
Linus is free to develop however he likes, but when he releases a new kernel as the stable kernel, says that it's the stable tree, changes the version number to indicate that it's no longer the development tree it's now the stable tree, tells distributions they should be shipping it, and it isn't fucking stable, there is no one to blame but him. I have no problem with him releasing unstable versions of the kernel, but he shouldn't call it a stable version if it isn't.
So we have a start, but you'll need to look furthur - try to remember what it was doing when it froze and look at the logs next time it starts to find out what it was trying to do - it may even just be an X lockup which would be fixed by a newer or older version of X.
There doesn't seem to be any common factor, running the same programs after a lockup hasn't seemed to produce another one, and there never seem to be any unusual log entries. It's not just X, I've tried to ssh in when it locks up, no good.
That's a start - so look for the differences and leave out modules for hardware you don't need and options you don't want.
I've always built my kernels with the minimum I want/need. The differences in options seem to just be reorganisation of the menus - other than the inclusion of alsa, but I always had the alsa modules loaded when I was using 2.4.
They may be called in different ways, so as I said before, what do the docs for the drivers for each bit of hardware say? Not simple, but if you do want to be on the leading edge not everything is going to be ready at the same time.
I don't want to be on the leading edge. 2.6 is supposed to be a *stable* kernel tree.
I've tried to troubleshoot this before, but it just seems to be a random lockup. It doesn't happen with this particular version and that will have to do.
Both Java and.NET apps need runtime environment to work.
.net doesn't. You can compile it to native code.
Your semantics masturbation won't change the fact that I can run Java programs on more platforms than I can run.NET programs on. Period. And THAT is what matters, not what is more "native" than something else.
If what you're doing with Java counts as running then I can use Bochs to run.net on any platform.
And, unlike every other protocol, FTP requires twoports be forwarded to it.
There are plenty of other protocols which require multiple ports. dhcp uses two, netbios needs three, and using bittorrent effectively from behind a firewall requires forwarding a full eight ports. Forwarding two ports isn't noticeably harder on most devices than forwarding a port for both tcp and udp, which many, I'd say the majority of, services use.
If you want Apache to return specific content depending on what hostname the client is trying to access, then yes, it does need to know, but that's hardly a requirement for just transferring files.
The sample config I got said you can only leave the hostname set as localhost if the server is just accessed locally, to be publicly accessible you need to set it to your external hostname.
So what's the point in having two connection modes at all?
Passive for when you're behind a firewall, active for when you aren't. Though the primary purpose is to allow server to server transfers (FXP).
Except with FTP, if you have very many clients connected, your speed starts dropping rapidly.
That's the case with any server-client protocol, and P2P ones suffer from the opposite problem: if you have few clients connected, your speed drops rapidly.Many FTP servers also impose a hard limit on the maximum number of connected users. I've seen tens of thousands of people connected to a single torrent at once, frequently getting download speeds of over 100 k/s; good luck getting that on an FTP server.
There are enough mirrors around that I've never had trouble maxing out my 2mbps connection over ftp. And on the other side, I've got files from 20 years ago at full speed quite easily by ftp, wheras torrents often dry up in days.
So? Why should a file transfer protocol behave like the C language? They're completely different.
But both are facing the same problem in this case: that text files have different line endings on different platforms. Having separate text and binary modes, horrible as it is, seems to be the best solution anyone has come up with so far.
Besides that, modern file sharing protocols also implement things such as automatic hash checking to ensure that your transfer is never corrupted,
That doesn't belong in the protocol, where there's no way to turn it off. If you want it, you can do it over any protocol via par2 files.
and many also allow you to download from multiple sources at once, which, all other things being equal, will always be faster than downloading from a single source.
Most modern FTP clients will also allow you to download from multiple sources at once.
What software is this? They could have just stopped using readline and used libedit instead, and kept their software closed source. Since they didn't, it would seem that maybe they would have open sourced their code anyways if someone just asked. Hard to say since you didn't mention who it was or what the software was.
I don't know, it's a story Stallman tells in interviews. Perhaps it was before libedit was available.
And consider the counter to your claim that forcing a particular piece of software to be GPLd makes the world better for everyone. It doesn't, it actually only makes the world better for everyone who uses GPL compatable licenses.
No, because the software is available to everyone.
But don't lie and pretend that its helping "everyone". It does nothing to help people who write free software for instance, code without nasty restrictions.
Huh? It is free software, it makes another library available to free software developers. I don't see how you can call the GPL a "nasty restriction" since the only restrictions it has on permission which is being granted. A typical piece of software won't let you redistribute it at all, under any conditions.
Making your code GPL because you are a GNU cult member and want everything to be GPL is fine, but that's not the group of people I was talking about. I was talking about normal developers who simply want to let other people use their code. These people would be better served using an MIT or ISC license, or if they want to make sure anyone who changes and distributes their code releases their changes, then the LGPL would be more appropriate.
If you just want your code to be used the most, then yes, that's what those licenses are for. If your object is to increase the amount of free software around, or get your code used as much as possible in free software, then the GPL is better.
The GPL makes your code useless to people who write free software too, not just closed source software.
Huh? The GPL makes your code give more of an advantage to people who write free software, not only can they use it, but those writing propriety software can't. If it's good enough it may even "convert" them. You seem to be implying you don't think GPL software is free software. Since RMS coined the phrase and started the movement *with* the GPL, it looks like it's you who's trying to redefine the english language to suit your purposes.
remember that IBM has a substantial interest in Linux. If it was the other way around we'd be crying foul about how studies will always find in favour of whoever's funding them. Anyone know if there's ever been a truly independent comparison
If it's a Java app all you need is one flaw in the Sun JVM (none of the others is up to scratch really) and you're toast. Use a real cross-platform language, yes it may require more maintenance but that's the price of diversity. It's also why diversity isn't so popular.
The individual does not trust the CA; at best he/she gives it the benefit of the doubt.
When using OpenPGP the only "root" is you. You need to trust someone, but it only needs to be someone you know.
The CA does not trust the company; at best it checks the status of the company in the chamber of commerce.
You can assign how much weight to give "recommendations", people trusted by people you trust, and decide how many "degrees of separation" are acceptable before you stop trusting people.
The rules of logical reasoning ("implication") do not apply when it comes to trust. Exempli gratia: my brother trusts me, I trust my Peruvian friend, but yet my brother does not trust my friend. I am sure you can come up with similar examples.
This must be because your brother doesn't trust you to verify people as trustworthy. OpenPGP would allow him to do this.
Annoying? It's the most effective compression format around, supports all the features I've seen in any archiver, and includes recovery volumes without needing an external tool (though people normally use par instead anyway)
Why does the Slashdot community, one of the largest Free / Open Source communities on the Net, care when a new proprietary version of some Windows-only software comes out?
Because, sad as it is, most slashdot users are still accessing it through a proprietary browser on a proprietary OS, and don't actually care about software freedom.
It's easy to use: The free zip programs included in Windows, although they are easy enough for most users, just don't feel right - and you can't really expect grandma to use open source utilities or to find other Windows freeware zip progams, however easy you might find it personally.
Why? It's UI is pretty much the same as any other dedicated program's, and if their website's slightly easier to find that doesn't make much difference. If you're getting your software the way most windows users do, by searching download.com, it's just as easy to find a free archiver.
It works well all of the time: People seem to have less difficulties with WinZip than with other archivers, and while this may just be due to familiarity or other reasons, the amount of people who recommend WinZip for everyday use is very telling.
I haven't found that at all. Winrar works all the time, winzip less so. I abandoned winzip when it couldn't open an ace file, not sure if that's still the same.
I have in a pretty similar situation, I was compressing 1.4gb of various binary data in the hope of getting it onto a cdr. Zip and gzip I didn't even bother with (no, ok, I did try them, they sucked, about 1.2gb). 7-zip and ace were better but pretty bad, about 8-900mb each. Arj got it down to about 706mb, just too much without overburning. Rar blew them all away, compressed it down to just over 550mb. Took two hours to do though.
There is nothing inherently evil about search engine optimization.
A good search engine will show me the results I want to see, the most relevant to my query. Therefore any SEO technique must distort this, making their site appear higher than I would want it to. They are improving their profit at the expense of my searching, damaging in a tiny way the quality of life of many people. It's evil, plain and simple.
Slashdot Mirror
ZIP doesn't do that; each file in a ZIP archive is compressed individually, with a separate compression dictionary. That hurts the compression ratio for ZIP archives that contain many files, particularly many small files, particularly many similar small files, like source code, for example. But it does mean that archive operations (like extracting or updating individual files or and adding files to or removing files from an archive) are fast and simple.
It's optional. You can do solid archiving, but it's not the default, for the reasons you mention. The files were several megabytes each, and the zip performance so much worse that I doubt it could be made competitive.
Even with the quick launch option, mozilla and firefox still start much slower. And whilst I've seen many claims that firefox now renders faster, personal experience and the benchmarks I've seen don't agree with that.
I'm not being a crybaby, I'm complaining in the hope that it will change things. Insightful moderations should go to things which actually are insightful.
Then they're going far beyond the call of duty.
I've *tried* to use gnome. It is far less usable. This is from simple experience.
What do you mean reversed? Ok/Cancel dialogs have been removed from GNOME... making this argument rather less important.
It may have been the yes/no dialog, someone decided the "positive" button should go on the left and the negative one on the right, rather than the normal way round. It was horrible.
Wrong... it was always an option -- just not one available directly from within Nautilus. P.S. I think Nautilus is a piece of shit.
In the first releases of gnome 2.6 it required a manual config file edit (or using the gnome registry editor). It wasn't a user-accessible option. An option was added pretty quick, certainly by 2.6.2, but there was no option in the first releases.
How does one customize it? If I could just make it maximisable it would be 100x easier to use. Starting up in "opened out" mode would be a big help too.
No way, konqueror is faster than anything else I've used (as long as you avoid javascript-heavy pages). Except links of course.
Hmm, must be a different issue. My TNT2 is no longer supported in the new drivers. Worse, it happened in the very release that enabled xinerama and glx together. Ah well.
It's older than the current stable, it's legacy. 2.2 and 2.0 are still under development and occasionally get things backported
Both 2.4 kai 2.6 kernels are fully supported by my distros, non of those distros have a policy that forced me to upgrade to 2.6.
Those are relatively major distos. It's difficult for smaller distros to support two separate versions. Given that they had to choose one I can't blame mine for picking 2.6.
Actually most of the linux servers out there are still running on 2.4, and 2.6 is recommended mostly for desktop workstations. Switching to 2.6 is not the only option, since both can be supported. I can think of a reason for a distro to be flamed if it _does stop_ supporting 2.4, but not in the other way around.
Distros which didn't switch to 2.6 as default were flamed quite severely, here for starters. I remember a post along the lines of "why is slackware still using 2.4 as a default 3 months after 2.6 came out? Answer: because it's an obsolete distribution for obsolete people. It just isn't relevant today". I'm pretty sure I saw a comment from Linus along the lines of "2.6 is the stable tree, distributions should be using it, I can't do much more until distributions have it as their default."
When you say that distros are flamed for not switching, flamed by who? Maybe from people like you, eager to upgrade everything to the last version because they think that this is always the right thing to do.
When have I ever said anything like that? I was happy with 2.4, it worked, and I didn't switch to 2.6 until it became necessary.
Maybe you should stop blaming linus for the way he chooses to develop _his_ project, and start complaining to your distro maintainers for quiting on 2.4 and shiping with an unstable 2.6.
Linus is free to develop however he likes, but when he releases a new kernel as the stable kernel, says that it's the stable tree, changes the version number to indicate that it's no longer the development tree it's now the stable tree, tells distributions they should be shipping it, and it isn't fucking stable, there is no one to blame but him. I have no problem with him releasing unstable versions of the kernel, but he shouldn't call it a stable version if it isn't.
There doesn't seem to be any common factor, running the same programs after a lockup hasn't seemed to produce another one, and there never seem to be any unusual log entries. It's not just X, I've tried to ssh in when it locks up, no good.
That's a start - so look for the differences and leave out modules for hardware you don't need and options you don't want.
I've always built my kernels with the minimum I want/need. The differences in options seem to just be reorganisation of the menus - other than the inclusion of alsa, but I always had the alsa modules loaded when I was using 2.4.
They may be called in different ways, so as I said before, what do the docs for the drivers for each bit of hardware say? Not simple, but if you do want to be on the leading edge not everything is going to be ready at the same time.
I don't want to be on the leading edge. 2.6 is supposed to be a *stable* kernel tree.
I've tried to troubleshoot this before, but it just seems to be a random lockup. It doesn't happen with this particular version and that will have to do.
.net doesn't. You can compile it to native code.
Your semantics masturbation won't change the fact that I can run Java programs on more platforms than I can run .NET programs on. Period. And THAT is what matters, not what is more "native" than something else.
If what you're doing with Java counts as running then I can use Bochs to run .net on any platform.
There are plenty of other protocols which require multiple ports. dhcp uses two, netbios needs three, and using bittorrent effectively from behind a firewall requires forwarding a full eight ports. Forwarding two ports isn't noticeably harder on most devices than forwarding a port for both tcp and udp, which many, I'd say the majority of, services use.
If you want Apache to return specific content depending on what hostname the client is trying to access, then yes, it does need to know, but that's hardly a requirement for just transferring files.
The sample config I got said you can only leave the hostname set as localhost if the server is just accessed locally, to be publicly accessible you need to set it to your external hostname.
So what's the point in having two connection modes at all?
Passive for when you're behind a firewall, active for when you aren't. Though the primary purpose is to allow server to server transfers (FXP).
Except with FTP, if you have very many clients connected, your speed starts dropping rapidly.
That's the case with any server-client protocol, and P2P ones suffer from the opposite problem: if you have few clients connected, your speed drops rapidly.Many FTP servers also impose a hard limit on the maximum number of connected users. I've seen tens of thousands of people connected to a single torrent at once, frequently getting download speeds of over 100 k/s; good luck getting that on an FTP server.
There are enough mirrors around that I've never had trouble maxing out my 2mbps connection over ftp. And on the other side, I've got files from 20 years ago at full speed quite easily by ftp, wheras torrents often dry up in days.
But both are facing the same problem in this case: that text files have different line endings on different platforms. Having separate text and binary modes, horrible as it is, seems to be the best solution anyone has come up with so far.
Besides that, modern file sharing protocols also implement things such as automatic hash checking to ensure that your transfer is never corrupted,
That doesn't belong in the protocol, where there's no way to turn it off. If you want it, you can do it over any protocol via par2 files.
and many also allow you to download from multiple sources at once, which, all other things being equal, will always be faster than downloading from a single source.
Most modern FTP clients will also allow you to download from multiple sources at once.
I don't know, it's a story Stallman tells in interviews. Perhaps it was before libedit was available.
And consider the counter to your claim that forcing a particular piece of software to be GPLd makes the world better for everyone. It doesn't, it actually only makes the world better for everyone who uses GPL compatable licenses.
No, because the software is available to everyone.
But don't lie and pretend that its helping "everyone". It does nothing to help people who write free software for instance, code without nasty restrictions.
Huh? It is free software, it makes another library available to free software developers. I don't see how you can call the GPL a "nasty restriction" since the only restrictions it has on permission which is being granted. A typical piece of software won't let you redistribute it at all, under any conditions.
Making your code GPL because you are a GNU cult member and want everything to be GPL is fine, but that's not the group of people I was talking about. I was talking about normal developers who simply want to let other people use their code. These people would be better served using an MIT or ISC license, or if they want to make sure anyone who changes and distributes their code releases their changes, then the LGPL would be more appropriate.
If you just want your code to be used the most, then yes, that's what those licenses are for. If your object is to increase the amount of free software around, or get your code used as much as possible in free software, then the GPL is better.
The GPL makes your code useless to people who write free software too, not just closed source software.
Huh? The GPL makes your code give more of an advantage to people who write free software, not only can they use it, but those writing propriety software can't. If it's good enough it may even "convert" them. You seem to be implying you don't think GPL software is free software. Since RMS coined the phrase and started the movement *with* the GPL, it looks like it's you who's trying to redefine the english language to suit your purposes.
remember that IBM has a substantial interest in Linux. If it was the other way around we'd be crying foul about how studies will always find in favour of whoever's funding them. Anyone know if there's ever been a truly independent comparison
If it's a Java app all you need is one flaw in the Sun JVM (none of the others is up to scratch really) and you're toast. Use a real cross-platform language, yes it may require more maintenance but that's the price of diversity. It's also why diversity isn't so popular.
Your sig is broken. I presume you wrote "#include ", slashcode sees the as a html tag and strips it out.
When using OpenPGP the only "root" is you. You need to trust someone, but it only needs to be someone you know.
The CA does not trust the company; at best it checks the status of the company in the chamber of commerce.
You can assign how much weight to give "recommendations", people trusted by people you trust, and decide how many "degrees of separation" are acceptable before you stop trusting people.
The rules of logical reasoning ("implication") do not apply when it comes to trust. Exempli gratia: my brother trusts me, I trust my Peruvian friend, but yet my brother does not trust my friend. I am sure you can come up with similar examples.
This must be because your brother doesn't trust you to verify people as trustworthy. OpenPGP would allow him to do this.
Annoying? It's the most effective compression format around, supports all the features I've seen in any archiver, and includes recovery volumes without needing an external tool (though people normally use par instead anyway)
Because, sad as it is, most slashdot users are still accessing it through a proprietary browser on a proprietary OS, and don't actually care about software freedom.
Why? It's UI is pretty much the same as any other dedicated program's, and if their website's slightly easier to find that doesn't make much difference. If you're getting your software the way most windows users do, by searching download.com, it's just as easy to find a free archiver.
It works well all of the time: People seem to have less difficulties with WinZip than with other archivers, and while this may just be due to familiarity or other reasons, the amount of people who recommend WinZip for everyday use is very telling.
I haven't found that at all. Winrar works all the time, winzip less so. I abandoned winzip when it couldn't open an ace file, not sure if that's still the same.
Drag over all the zipfiles, right click, extract. Just as simple if you're used to gui concepts.
I have in a pretty similar situation, I was compressing 1.4gb of various binary data in the hope of getting it onto a cdr. Zip and gzip I didn't even bother with (no, ok, I did try them, they sucked, about 1.2gb). 7-zip and ace were better but pretty bad, about 8-900mb each. Arj got it down to about 706mb, just too much without overburning. Rar blew them all away, compressed it down to just over 550mb. Took two hours to do though.
KDE now has a very nice and efficient implementation (kpdf), which will be available for all major platforms eventually.
A good search engine will show me the results I want to see, the most relevant to my query. Therefore any SEO technique must distort this, making their site appear higher than I would want it to. They are improving their profit at the expense of my searching, damaging in a tiny way the quality of life of many people. It's evil, plain and simple.