Unfortunately, I've yet to find anther with the feature set of these -- especially of KDE's window manager. I'm probably just not looking hard enough yet, though.
That's one advantage Linux has held long. Nothing in and around the Kernel expects KDE to be there, so the temptation to make a call into KDE is less.
It's still there, of course. There seem to be advantages to C++ development (especially with Qt) versus straight C development, even if you're not developing a GUI app.
But you're right -- there's far more desktop-environment-agnostic stuff there. The fact that Windows still needs a video card to run a headless server seems pretty ridiculous.
Learning that KDE meddle with the Bluetooth stack was therefore surprising, but I assume they're not exactly meddling with the stack itself but instead handles stuff like authorizing devices.
Correct. And moreover, they are not actually doing anything directly to the Bluetooth stack.
Essentially, what we have is, there's a bluetooth stack in there that actually handles things. It's roughly equivalent to ifconfig + NetworkManager + wpasupplicant, and so on -- all the low-level, fiddly little pieces you assemble such that you can configure and use bluetooth.
On Linux, that stack is called BlueZ.
On Windows and OS X, I'm pretty positive it is not done using BlueZ. And I've got no clue what other BSDs use.
So the idea of KDE's bluetooth layer is to manage the Bluetooth stack in an abstract and cross-platform way. It means that either KDE will have built-in bluetooth support, or applications will be written which run in the system tray and configure bluetooth, or both, and the whole thing can be ported, the entire BlueZ stack can be ripped out from under it and replaced with something else.
Unfortunately, all this great engineering becomes a great WTF when even the native BlueZ support isn't working. Due to some version mismatch snafus, Kubuntu Intrepid shipped without working Bluetooth.
but does away with the flat config files. You may think that is a bad thing, but the truth is that you can backend gconf with files or databases or anything you like (theoretically) and that is, well, great.
If they would only backend it with something more easily human-editable. XML does not count. YAML would be much better, IMO.
And having the use of gconf be the standard for GNOME means less stupid dotfiles and dotdirs in your homedir. I don't know about you, but mine has gotten kind of out of control lately.
If it means they just push.foo into.gconf/foo, I don't know that counts. That means I'll just have a cluttered.gconf instead.
At least, I assume it's something like that. KDE will generally store the Foo config in.kde/share/config/foorc, and other files (state, downloads, etc) in.kde/share/apps/foo. Which means I now have two cluttered directories instead of one.
It's a desktop environment, sort of, but it doesn't provide an entire library of applications.... Same thing with Fluxbox and other pure window managers - absolutely no apps
That's fine. Believe me, I know what I'm getting into.
My progression is pretty much: GNOME (as a newbie), then I toyed with Enlightenment and Windowmaker (probably in that order), then Fluxbox (my machine was really weak), then Beryl (Compiz fork, which I used the way I used Fluxbox), and finally KDE.
There was also a period of entirely too much time spent trying to get X to work -- as in, a week or three spent entirely in VTs. Towards the end of WindowMaker, and throughout much of Fluxbox (and then Beryl), what I mainly used was a keystroke to bring up a launcher of some sort, then type a command. Basically, alt+f2 in KDE. I then used Katapult in kde3, then alt+f2 in kde4.
And I use the commandline (now in konsole) as a file browser.
What I like most about KDE now is how flexible it is in terms of keystrokes, both global keystrokes (map a key to pause/play in Amarok, say) and window manager keystrokes (win+arrow keys packs windows in that direction, so I can very quickly arrange things with the keyboard)
However, I'm thinking at this point that I'd really like one that's either easily scriptable, or written from scratch in a scripting language. For example, I would love to have a keystroke to switch between similar windows, and by some order other than most-recently-used -- one thing sorely missed from OS X was command+right/left to cycle through open terminal windows. It seems Linux UIs tend to use tabs where I'd use Windows -- come on, people, I have a 3840*1200 area to fill up, I'd much rather fill it with little 80x24 terminals, not one gigantic tabbed terminal!
Why would you lose faith in a product with a long track record because of one bad / too early release?
Because of how profoundly bad it was, and because it shows they are more committed to their release schedule than to releasing a quality product.
This was a known issue, and they decided to just drop support.
It is going to get fixed.
Yes. And I wonder what will be broken next release?
Time was, things didn't break with new releases. I could just emerge --update world and everything would work. (I think that's the command; Gentoo was a long time ago.)
there is no evidence of a deep structural permanent problem.
With KDE, the largest evidence I see of a problem is that they don't realize just how badly they fucked up.
It's like Steve Ballmer calling Vista a success, with a straight face. That's when you know he's lost it.
Instead, I installed the gnome bluetooth manager. That does show how pathetic the state of bluetooth support on kde is, though. How the hell is this not a priority?
It's really not the trolls' fault. KDE4.0 was never intended to be actually used by the mainstream. This wasn't made sufficiently explicit
They released a dot-oh release. If that was the only thing they did, it shows that they don't know how to name their product. Just about anything else, I can clearly look at the version number -- worst case, even number stable, odd number unstable.
Perhaps more importantly, they've stopped adding new features to v3, and some projects have stopped supporting it at all. I realize Amarok is a separate project, but it's the best example -- I've seen similar things in core -- the latest kde3 version of AmaroK has transcoding broken, which won't be fixed because development on that branch has stopped. The latest kde4 version hasn't even begun to implement transcoding.
Most cases, you see the v4 version is much improved, and has actually addressed all of these issues -- but then you have to use and deal with v4. For example: I have no complaints about Okular vs KPDF, other than that there's no KDE3 version of Okular.
perhaps they felt that those who forgot the lessons of history (remember KDE 3.0? me neither
Did they make the same mistake with 3.0, I wonder?
But the point is, Kubuntu should have stuck with KDE 3.5,
It was already 4.1 by the release, not just 4.0. And 4.1 was the release that was supposed to be "so much better" and "actually ready to replace 3.5".
Now it's 4.2 that's supposed to be -- but I'm wondering if we'll get feature parity with 3.5 by the time we get 4.5.
It's great for guys like Monty Python, who are already famous and already have merchandise in production and have millions of fans.
I think the ponit is that if something like that was released today, and gained millions of fans on YouTube, that they would also be able to sell a few DVDs. Which is what Monty Python just did. It's not as if they were selling Dead Parrot Plushies -- it was just DVDs.
Not until YouTube starts putting in ads or something and paying the content producers.
There's no way it would improve their average sales to have those actual films previewable on YouTube.
I think that says more about their average quality than it does about YouTube.
It also would very likely force the MPAA's feet into their mouths, if they had to essentially make the argument that "Most of our movies suck so much that if people actually saw them first, no one would want to buy them. We make most of our money by selling people crap they don't really want."
thankfully, for me ATI has never initiated any MITM attacks.
Seems I missed this before...
No, MITM is not about ATI -- Man In The Middle attack. As in, someone between you and ATI. Given you're on wireless, that's not entirely implausible -- but it is completely preventable, by at least signing the software, if not providing it through a common, trusted channel.
What impresses me about Linux, that no one else has done so far, is that it not only provides that additional security, but it makes the process of installing it easier than other methods -- including the Windows methods of downloading or inserting a CD.
The closest, oddly enough, is Apple's App Store, but that's entirely too limited.
If you've already got cable, it seems to me that getting TV with your fiber could still be an attractive offer. The $65 is the cheapest plan, but that's phone+internet, no TV.
Seriously, if I could get $65 fiber, I might, but where do I need to live ?
If a town of 10k people, in the middle of nowhere, can get that -- or crappy DSL, or sort of decent cable, and there's another, business-oriented ISP around selling fiber, too -- why is Comcast winning everywhere else?
Eum, isn't KDE a shell? Why does it need to support Bluetooth, isn't that the job for the OS?
Well, the OS supports the physical hardware. Most of the logic of managing which devices are allowed to connect and which aren't, among other things, is managed in user space by a stack called BlueZ, which mostly runs as a daemon and is controlled through arcane config files.
Now, I'm not afraid of arcane config files, but I was a bit spoiled. It was a few clicks to get my mouse working in KDE3. It would probably take me a few hours to learn enough to do it manually with BlueZ.
In KDE3, the bluetooth manager was a separate application. In KDE4, that's still true... sort of. It's also part of the "solid" system, I believe -- which is KDE4's hardware abstraction magic. It wires GUIs to potentially OS-specific backends -- looking at the config pane, it looks to support power management, network management, and bluetooth.
But the idea is that a KDE bluetooth manager should also work on Windows and OS X, neither of which will be running BlueZ. Similarly, the KDE network manager should work on Windows and OS X, neither of which will be running the Ubuntu-like NetworkManager.
Like so many parts of KDE4, it is a really good idea, and you can see how it has the potential for greatness.
Unfortunately, Ubuntu shipped incompatible versions of parts of this stack -- I believe it was that a new BlueZ was required by the new kernel, but the new BlueZ was incompatible with the old Solid. Which means that, out of the box, my mouse didn't work.
That was my introduction to KDE4: Why doesn't my fucking mouse work anymore? It's 2008, and my mouse doesn't work?!
I wish I could say it got better after that. It didn't -- it got worse.
I actually got sick of having to learn and spend time on every new thing, and having people laugh at me while I was reading the iwconfig manpage, while they just clicked a menu on OS X and joined a wireless network.
With Kubuntu Hardy and KDE3, the joke was on them. Everything Just Worked, out of the box, with far more configurability than just about anything else. All those "extra applications" includes things like wifi, bluetooth, sound, and USB mass storage hotplug, as simple, intuitive GUIs that require no more learning than "Let's try right-clicking the Bluetooth icon... Oh, I get how this works."
Then came Kubuntu Intrepid and KDE4. Bluetooth didn't work. Wireless became more complex, and no longer uses kdewallet to store things. I'm taking a long, hard look at things like xfce, fluxbox, or just rolling my own.
Or I could be lazy and use KDE, which, instead of forcing me to use arcane commandline utilities and XML, provides me with a nice GUI and a much simpler, much more UNIX-y set of config files. KDE4 screwed it up a lot, but it's still nowhere near as bad as GNOME.
I'll remind people one of the older reasons Linus chose KDE: There's a nice GUI for configuring what each mouse button on the title bar of a window does. In GNOME, this functionality simply wasn't available. I assume it wasn't in a config file either, because Linus ended up having to write a patch. Once he wrote it, he couldn't figure out where to send it.
Now, if Linus fucking Torvalds can't figure out where to send a patch, you have a problem.
And it is for that reason that I have such frustration with it...
It used to be, I could in good conscience make jokes about Windows, about how when Microsoft makes a "beta" release, it's what the rest of the world would call an Alpha, the release is really Beta quality, and SP1 is release candidate 1. By SP2, the product might be ready.
I could laugh about how Microsoft, and occasionally other proprietary shops, would follow that model, as opposed to the open source model, where the versioning seems to go, alpha is unstable (so beware), beta is good enough to use, release candidates are pretty solid, and release versions you can bet your business on.
But KDE4 was an alpha release. 4.1 was a beta release. Surrounding projects have done no better -- Amarok currently will not transcode automatically from flac to aac for ipods; it insists on mp3. This is a bug; it used to work. The stable Amarok won't fix the bug, because it's being depricated in favor of the kde4 version of Amarok, which doesn't yet support transcoding. WTF?
Kubuntu has done spectacularly bad as well. My mouse didn't work. Why? Because they included an update to the Bluez stack, to support a change to the kernel, but the KDE4 Bluetooth support hadn't been updated to support that new Bluez stack. Their solution? Drop bluetooth support in Kubuntu Intrepid. WTF?
It has been pretty much my own private Daily WTF as I continue to use KDE4. It's not yet so bad I'm going back to GNOME, but by this time next year, I suspect I'll be using something like Fluxbox again.
It is all situational, what is good for one is not automatically good for everyone
Agreed.
I will, however, say that this should be left to the individual, whether it's Internet, coffee, or crack cocaine. The damage from trying to prevent other people from making their own mistakes is far worse than the mistakes themselves.
Now, it would be nice if you could simply say "you follow the spam link posted on our board and you'll be banned from the board", but you can't enforce it. How do you want to know whether someone followed a spam link?
That's trivial, actually.
You could provide a redirect. Thus, the link is not to http://spam.me/, it's to http://my.board/spam.me. (Roughly -- you could store the whole URL in there, too.)
Or, you could leave the link alone, but use Javascript to intercept it, and submit a statistic to you before following it.
Or, you could make the link not a link at all, but a span styled as a link, with javascript that does whatever you want.
Of these, I'd prefer a combination of the first and the second -- transparent to middle-click-open-in-tab and javascript-free users, but enough javascript to hide from the casual observer (even replacing the status bar message) what you're doing, if they aren't middle-clicking on it -- also enough to ensure that the redirect page does not appear in the browser's history.
But however you do it, tracking clicks is really, really not hard.
I don't think it's enough to solve the problem, though. Consider that the automated spambots will continue to hit every single board. Consider that a botnet effectively costs zero to keep running.
Probably the simplest long-term solution, though of course not complete, is to do statistical analysis on the body of the comment itself. Categorize comments by ISP, and if a sufficient amount of spam traffic comes from one ISP to actually start costing bandwidth/cpu/money, block it wholesale, and fire an email off to abuse@example.net letting them know why you've blocked them.
Please, don't suggest something stupid AND already obsolete, we might get saddled with it.
Fortunately, it has two advantages:
First, for those who aren't using botnets, or sufficiently large botnets, it's a significant impediment.
Second, more cycles increases the chance that people will notice their computers slowing down and figure out its a botnet.
Finally, it really doesn't matter whether we get saddled with it or not -- since it's just using Javascript, it's no more cumbersome than Slashdot's current comment system. And if it's completely ineffective, it could be turned off with no ill effects.
Why aren't there more ISPs who just sell pure bandwidth?
My biggest complaint about my current ISP is the bundling -- there is actually no way to just get Internet, it's Internet + Phone, or Interent + TV, or all three. However, it's also $65 for 100 mbit fiber, no installation fee, and in my experience, no throttling, no bullshit.
I don't really care for the ToS, so I suppose if they do start sucking, there won't be a lot I can do about it. But so far, they've been good.
So my question is, why is this so rare? Why are craptastic ISPs like Comcast so common?
Your advice is not going to make things simpler for the topic starter.
If successful, it will make things simpler for other students in the future.
Besides which, I don't know about you, but especially in college, I'd have the idealism to want an interesting, maybe important life, not necessarily an easy one. It seems every US president ages ten years for every four spent in office -- does that mean children shouldn't want to grow up to be president?
This is bigger than that, too. Rosa Parks would probably have had an easier time giving up her seat than going to jail.
start selling packages like X gigabytes of 1st class traffice, Y gigabytes of 2nd class traffic, and Z gigabytes of 3rd class traffic.
That seems like it would be a bit complex, and also problematic if they did not ship their own routers and instructions for configuring it.
Why not just sell pure bandwidth, and if people want to prioritize things, let them do it within their own networks? If I'm saturating my connection with BitTorrent, it's really up to me to QoS it down until Skype works. But, if I'm saturating my connection with BitTorrent, and someone else is having problems with Skype, that suggests they should buy more bandwidth.
I can see where TOS might be easier for the ISPs, let them squeeze a bit more out of their networks, maybe oversell a bit more and acknowledge that your torrent will slow to a crawl (but your voice will still work) during "peak" hours.
On the other hand, Amazon seems to be able to put a relatively cheap, relatively constant price on all network traffic to Amazon Web Services. I don't know if my bill is typical, but I pay $65 for fiber -- split evenly, that would be 300 gigs upload and 176 gigs download, or 150 gigs up and 265 gigs down... per month. I mean, I might do more than that torrenting, but not much, and I imagine that's a good deal more than Comcast currently provides.
Is that OK with you, or do you want to shove these things down their throat as well?
No one's forcing you to use them. Nor will anyone force you to use the Internet, except for essential communication.
They are, however, already forcing you to pay for them. It's called a public library. And they are already forcing you to use the phone and the Internet, on occasion -- filing for unemployment in my area takes place over the Internet, even if you walk into the office, and you have to report in over the phone.
Slashdot Mirror
Unfortunately, I've yet to find anther with the feature set of these -- especially of KDE's window manager. I'm probably just not looking hard enough yet, though.
That's one advantage Linux has held long. Nothing in and around the Kernel expects KDE to be there, so the temptation to make a call into KDE is less.
It's still there, of course. There seem to be advantages to C++ development (especially with Qt) versus straight C development, even if you're not developing a GUI app.
But you're right -- there's far more desktop-environment-agnostic stuff there. The fact that Windows still needs a video card to run a headless server seems pretty ridiculous.
Learning that KDE meddle with the Bluetooth stack was therefore surprising, but I assume they're not exactly meddling with the stack itself but instead handles stuff like authorizing devices.
Correct. And moreover, they are not actually doing anything directly to the Bluetooth stack.
Essentially, what we have is, there's a bluetooth stack in there that actually handles things. It's roughly equivalent to ifconfig + NetworkManager + wpasupplicant, and so on -- all the low-level, fiddly little pieces you assemble such that you can configure and use bluetooth.
On Linux, that stack is called BlueZ.
On Windows and OS X, I'm pretty positive it is not done using BlueZ. And I've got no clue what other BSDs use.
So the idea of KDE's bluetooth layer is to manage the Bluetooth stack in an abstract and cross-platform way. It means that either KDE will have built-in bluetooth support, or applications will be written which run in the system tray and configure bluetooth, or both, and the whole thing can be ported, the entire BlueZ stack can be ripped out from under it and replaced with something else.
Unfortunately, all this great engineering becomes a great WTF when even the native BlueZ support isn't working. Due to some version mismatch snafus, Kubuntu Intrepid shipped without working Bluetooth.
He can send it to the people who made linux usable for the masses.
That's just it -- he couldn't find them.
He was all set to send it in, improve the project, make Linux that much more usable.
But their own development process wasn't "usable" enough to even find a mailing list or a bug tracker.
but does away with the flat config files. You may think that is a bad thing, but the truth is that you can backend gconf with files or databases or anything you like (theoretically) and that is, well, great.
If they would only backend it with something more easily human-editable. XML does not count. YAML would be much better, IMO.
And having the use of gconf be the standard for GNOME means less stupid dotfiles and dotdirs in your homedir. I don't know about you, but mine has gotten kind of out of control lately.
If it means they just push .foo into .gconf/foo, I don't know that counts. That means I'll just have a cluttered .gconf instead.
At least, I assume it's something like that. KDE will generally store the Foo config in .kde/share/config/foorc, and other files (state, downloads, etc) in .kde/share/apps/foo. Which means I now have two cluttered directories instead of one.
It's a desktop environment, sort of, but it doesn't provide an entire library of applications.... Same thing with Fluxbox and other pure window managers - absolutely no apps
That's fine. Believe me, I know what I'm getting into.
My progression is pretty much: GNOME (as a newbie), then I toyed with Enlightenment and Windowmaker (probably in that order), then Fluxbox (my machine was really weak), then Beryl (Compiz fork, which I used the way I used Fluxbox), and finally KDE.
There was also a period of entirely too much time spent trying to get X to work -- as in, a week or three spent entirely in VTs. Towards the end of WindowMaker, and throughout much of Fluxbox (and then Beryl), what I mainly used was a keystroke to bring up a launcher of some sort, then type a command. Basically, alt+f2 in KDE. I then used Katapult in kde3, then alt+f2 in kde4.
And I use the commandline (now in konsole) as a file browser.
What I like most about KDE now is how flexible it is in terms of keystrokes, both global keystrokes (map a key to pause/play in Amarok, say) and window manager keystrokes (win+arrow keys packs windows in that direction, so I can very quickly arrange things with the keyboard)
However, I'm thinking at this point that I'd really like one that's either easily scriptable, or written from scratch in a scripting language. For example, I would love to have a keystroke to switch between similar windows, and by some order other than most-recently-used -- one thing sorely missed from OS X was command+right/left to cycle through open terminal windows. It seems Linux UIs tend to use tabs where I'd use Windows -- come on, people, I have a 3840*1200 area to fill up, I'd much rather fill it with little 80x24 terminals, not one gigantic tabbed terminal!
Why would you lose faith in a product with a long track record because of one bad / too early release?
Because of how profoundly bad it was, and because it shows they are more committed to their release schedule than to releasing a quality product.
This was a known issue, and they decided to just drop support.
It is going to get fixed.
Yes. And I wonder what will be broken next release?
Time was, things didn't break with new releases. I could just emerge --update world and everything would work. (I think that's the command; Gentoo was a long time ago.)
there is no evidence of a deep structural permanent problem.
With KDE, the largest evidence I see of a problem is that they don't realize just how badly they fucked up.
It's like Steve Ballmer calling Vista a success, with a straight face. That's when you know he's lost it.
A few clicks of what, exactly?
Trackpad, in my case, though if I had learned the keystrokes, there is a way to use the numpad to control the cursor.
3. Learn config file.
4. Add to config file.
5. Restart bluez.
Instead, I installed the gnome bluetooth manager. That does show how pathetic the state of bluetooth support on kde is, though. How the hell is this not a priority?
It's really not the trolls' fault. KDE4.0 was never intended to be actually used by the mainstream. This wasn't made sufficiently explicit
They released a dot-oh release. If that was the only thing they did, it shows that they don't know how to name their product. Just about anything else, I can clearly look at the version number -- worst case, even number stable, odd number unstable.
Perhaps more importantly, they've stopped adding new features to v3, and some projects have stopped supporting it at all. I realize Amarok is a separate project, but it's the best example -- I've seen similar things in core -- the latest kde3 version of AmaroK has transcoding broken, which won't be fixed because development on that branch has stopped. The latest kde4 version hasn't even begun to implement transcoding.
Most cases, you see the v4 version is much improved, and has actually addressed all of these issues -- but then you have to use and deal with v4. For example: I have no complaints about Okular vs KPDF, other than that there's no KDE3 version of Okular.
perhaps they felt that those who forgot the lessons of history (remember KDE 3.0? me neither
Did they make the same mistake with 3.0, I wonder?
But the point is, Kubuntu should have stuck with KDE 3.5,
It was already 4.1 by the release, not just 4.0. And 4.1 was the release that was supposed to be "so much better" and "actually ready to replace 3.5".
Now it's 4.2 that's supposed to be -- but I'm wondering if we'll get feature parity with 3.5 by the time we get 4.5.
It's great for guys like Monty Python, who are already famous and already have merchandise in production and have millions of fans.
I think the ponit is that if something like that was released today, and gained millions of fans on YouTube, that they would also be able to sell a few DVDs. Which is what Monty Python just did. It's not as if they were selling Dead Parrot Plushies -- it was just DVDs.
Not until YouTube starts putting in ads or something and paying the content producers.
Which they are, to some of them.
There's no way it would improve their average sales to have those actual films previewable on YouTube.
I think that says more about their average quality than it does about YouTube.
It also would very likely force the MPAA's feet into their mouths, if they had to essentially make the argument that "Most of our movies suck so much that if people actually saw them first, no one would want to buy them. We make most of our money by selling people crap they don't really want."
thankfully, for me ATI has never initiated any MITM attacks.
Seems I missed this before...
No, MITM is not about ATI -- Man In The Middle attack. As in, someone between you and ATI. Given you're on wireless, that's not entirely implausible -- but it is completely preventable, by at least signing the software, if not providing it through a common, trusted channel.
What impresses me about Linux, that no one else has done so far, is that it not only provides that additional security, but it makes the process of installing it easier than other methods -- including the Windows methods of downloading or inserting a CD.
The closest, oddly enough, is Apple's App Store, but that's entirely too limited.
If you've already got cable, it seems to me that getting TV with your fiber could still be an attractive offer. The $65 is the cheapest plan, but that's phone+internet, no TV.
Seriously, if I could get $65 fiber, I might, but where do I need to live ?
Fairfield, Iowa.
If a town of 10k people, in the middle of nowhere, can get that -- or crappy DSL, or sort of decent cable, and there's another, business-oriented ISP around selling fiber, too -- why is Comcast winning everywhere else?
Eum, isn't KDE a shell? Why does it need to support Bluetooth, isn't that the job for the OS?
Well, the OS supports the physical hardware. Most of the logic of managing which devices are allowed to connect and which aren't, among other things, is managed in user space by a stack called BlueZ, which mostly runs as a daemon and is controlled through arcane config files.
Now, I'm not afraid of arcane config files, but I was a bit spoiled. It was a few clicks to get my mouse working in KDE3. It would probably take me a few hours to learn enough to do it manually with BlueZ.
In KDE3, the bluetooth manager was a separate application. In KDE4, that's still true... sort of. It's also part of the "solid" system, I believe -- which is KDE4's hardware abstraction magic. It wires GUIs to potentially OS-specific backends -- looking at the config pane, it looks to support power management, network management, and bluetooth.
But the idea is that a KDE bluetooth manager should also work on Windows and OS X, neither of which will be running BlueZ. Similarly, the KDE network manager should work on Windows and OS X, neither of which will be running the Ubuntu-like NetworkManager.
Like so many parts of KDE4, it is a really good idea, and you can see how it has the potential for greatness.
Unfortunately, Ubuntu shipped incompatible versions of parts of this stack -- I believe it was that a new BlueZ was required by the new kernel, but the new BlueZ was incompatible with the old Solid. Which means that, out of the box, my mouse didn't work.
That was my introduction to KDE4: Why doesn't my fucking mouse work anymore? It's 2008, and my mouse doesn't work?!
I wish I could say it got better after that. It didn't -- it got worse.
Linus will be back. KDE 4.2 is turning out very nice and I'm sure he will give it a try.
Indeed. I'm told it's even approaching the quality of a 4.0 release.
As it is, I've pretty much lost all faith. Kubuntu broke my Bluetooth with the Intrepid release. WTF?
I actually got sick of having to learn and spend time on every new thing, and having people laugh at me while I was reading the iwconfig manpage, while they just clicked a menu on OS X and joined a wireless network.
With Kubuntu Hardy and KDE3, the joke was on them. Everything Just Worked, out of the box, with far more configurability than just about anything else. All those "extra applications" includes things like wifi, bluetooth, sound, and USB mass storage hotplug, as simple, intuitive GUIs that require no more learning than "Let's try right-clicking the Bluetooth icon... Oh, I get how this works."
Then came Kubuntu Intrepid and KDE4. Bluetooth didn't work. Wireless became more complex, and no longer uses kdewallet to store things. I'm taking a long, hard look at things like xfce, fluxbox, or just rolling my own.
You could always just grow a brain and use gconf,
Or I could be lazy and use KDE, which, instead of forcing me to use arcane commandline utilities and XML, provides me with a nice GUI and a much simpler, much more UNIX-y set of config files. KDE4 screwed it up a lot, but it's still nowhere near as bad as GNOME.
I'll remind people one of the older reasons Linus chose KDE: There's a nice GUI for configuring what each mouse button on the title bar of a window does. In GNOME, this functionality simply wasn't available. I assume it wasn't in a config file either, because Linus ended up having to write a patch. Once he wrote it, he couldn't figure out where to send it.
Now, if Linus fucking Torvalds can't figure out where to send a patch, you have a problem.
In short, KDE4 is basically a year late.
And it is for that reason that I have such frustration with it...
It used to be, I could in good conscience make jokes about Windows, about how when Microsoft makes a "beta" release, it's what the rest of the world would call an Alpha, the release is really Beta quality, and SP1 is release candidate 1. By SP2, the product might be ready.
I could laugh about how Microsoft, and occasionally other proprietary shops, would follow that model, as opposed to the open source model, where the versioning seems to go, alpha is unstable (so beware), beta is good enough to use, release candidates are pretty solid, and release versions you can bet your business on.
But KDE4 was an alpha release. 4.1 was a beta release. Surrounding projects have done no better -- Amarok currently will not transcode automatically from flac to aac for ipods; it insists on mp3. This is a bug; it used to work. The stable Amarok won't fix the bug, because it's being depricated in favor of the kde4 version of Amarok, which doesn't yet support transcoding. WTF?
Kubuntu has done spectacularly bad as well. My mouse didn't work. Why? Because they included an update to the Bluez stack, to support a change to the kernel, but the KDE4 Bluetooth support hadn't been updated to support that new Bluez stack. Their solution? Drop bluetooth support in Kubuntu Intrepid. WTF?
It has been pretty much my own private Daily WTF as I continue to use KDE4. It's not yet so bad I'm going back to GNOME, but by this time next year, I suspect I'll be using something like Fluxbox again.
It is all situational, what is good for one is not automatically good for everyone
Agreed.
I will, however, say that this should be left to the individual, whether it's Internet, coffee, or crack cocaine. The damage from trying to prevent other people from making their own mistakes is far worse than the mistakes themselves.
Now, it would be nice if you could simply say "you follow the spam link posted on our board and you'll be banned from the board", but you can't enforce it. How do you want to know whether someone followed a spam link?
That's trivial, actually.
You could provide a redirect. Thus, the link is not to http://spam.me/, it's to http://my.board/spam.me. (Roughly -- you could store the whole URL in there, too.)
Or, you could leave the link alone, but use Javascript to intercept it, and submit a statistic to you before following it.
Or, you could make the link not a link at all, but a span styled as a link, with javascript that does whatever you want.
Of these, I'd prefer a combination of the first and the second -- transparent to middle-click-open-in-tab and javascript-free users, but enough javascript to hide from the casual observer (even replacing the status bar message) what you're doing, if they aren't middle-clicking on it -- also enough to ensure that the redirect page does not appear in the browser's history.
But however you do it, tracking clicks is really, really not hard.
I don't think it's enough to solve the problem, though. Consider that the automated spambots will continue to hit every single board. Consider that a botnet effectively costs zero to keep running.
Probably the simplest long-term solution, though of course not complete, is to do statistical analysis on the body of the comment itself. Categorize comments by ISP, and if a sufficient amount of spam traffic comes from one ISP to actually start costing bandwidth/cpu/money, block it wholesale, and fire an email off to abuse@example.net letting them know why you've blocked them.
Please, don't suggest something stupid AND already obsolete, we might get saddled with it.
Fortunately, it has two advantages:
First, for those who aren't using botnets, or sufficiently large botnets, it's a significant impediment.
Second, more cycles increases the chance that people will notice their computers slowing down and figure out its a botnet.
Finally, it really doesn't matter whether we get saddled with it or not -- since it's just using Javascript, it's no more cumbersome than Slashdot's current comment system. And if it's completely ineffective, it could be turned off with no ill effects.
I should rephrase that, then.
Why aren't there more ISPs who just sell pure bandwidth?
My biggest complaint about my current ISP is the bundling -- there is actually no way to just get Internet, it's Internet + Phone, or Interent + TV, or all three. However, it's also $65 for 100 mbit fiber, no installation fee, and in my experience, no throttling, no bullshit.
I don't really care for the ToS, so I suppose if they do start sucking, there won't be a lot I can do about it. But so far, they've been good.
So my question is, why is this so rare? Why are craptastic ISPs like Comcast so common?
Your advice is not going to make things simpler for the topic starter.
If successful, it will make things simpler for other students in the future.
Besides which, I don't know about you, but especially in college, I'd have the idealism to want an interesting, maybe important life, not necessarily an easy one. It seems every US president ages ten years for every four spent in office -- does that mean children shouldn't want to grow up to be president?
This is bigger than that, too. Rosa Parks would probably have had an easier time giving up her seat than going to jail.
start selling packages like X gigabytes of 1st class traffice, Y gigabytes of 2nd class traffic, and Z gigabytes of 3rd class traffic.
That seems like it would be a bit complex, and also problematic if they did not ship their own routers and instructions for configuring it.
Why not just sell pure bandwidth, and if people want to prioritize things, let them do it within their own networks? If I'm saturating my connection with BitTorrent, it's really up to me to QoS it down until Skype works. But, if I'm saturating my connection with BitTorrent, and someone else is having problems with Skype, that suggests they should buy more bandwidth.
I can see where TOS might be easier for the ISPs, let them squeeze a bit more out of their networks, maybe oversell a bit more and acknowledge that your torrent will slow to a crawl (but your voice will still work) during "peak" hours.
On the other hand, Amazon seems to be able to put a relatively cheap, relatively constant price on all network traffic to Amazon Web Services. I don't know if my bill is typical, but I pay $65 for fiber -- split evenly, that would be 300 gigs upload and 176 gigs download, or 150 gigs up and 265 gigs down... per month. I mean, I might do more than that torrenting, but not much, and I imagine that's a good deal more than Comcast currently provides.
Is that OK with you, or do you want to shove these things down their throat as well?
No one's forcing you to use them. Nor will anyone force you to use the Internet, except for essential communication.
They are, however, already forcing you to pay for them. It's called a public library. And they are already forcing you to use the phone and the Internet, on occasion -- filing for unemployment in my area takes place over the Internet, even if you walk into the office, and you have to report in over the phone.