I think Ubuntu simply may be stretching their resources too thin. It's quite clear that they put a *lot* of customization into the packages they ship (including KDE), but IME Ubuntu feels a lot less stable than Mandrake (many versions ago, my first Linux install) or Gentoo (my current Linux installs). The customizations clearly add value, as most people seem to prefer Ubuntu, but I shy away from it for my own use as it strikes me as unstable (and I much prefer Portage to Apt).
My KDE sessions normally last until the next version of KDE is released (and then I upgrade), and that's the way I like it.
Now, I'll give you that I'm not using the largest, baddest laptop around. I'm using a Thinkpad T-60 which, as far as 17" display laptops go, is pretty reasonably sized. I had a Dell D810 for awhile, though, which was a monster of a laptop. It still fit my purpose above, though.
It is actually. ALSA's dmix allows multiple programs to use the sound card concurrently and transparently, regardless of hardware support. Dmix has been on by default for at least a year or two now (since ALSA 1.0.8, IIRC), and you could enable if you wanted it before that.
So, is a RAM drive not a drive then, simply because it doesn't use SATA/IDE/etc?
I don't see how it matters at all that the FusionIO uses PCIe instead of SATA, you're still going to use it like a regular hard drive (or solid state drive, for that matter).
If the FusionIO was exactly the same except it connected via SATA instead of PCIe, would you consider it a drive then?
Direct 3D rendering is slower than on Windows [with fglrx].
If it's still slower, it's only by a tiny margin now. fglrx 8.42 is *hugely* faster than 8.40 and below. My Radeon X1400 now runs Doom 3 at 37.1 fps (1024x768, medium quality), compared to the 21 fps it got before with the same settings.
fglrx still has issues, but its 3D performance is no longer one of them.
Alky is vapourware, don't hold your breath waiting for it.
As the posters above noted, you can already use (most) Wine dlls on Windows. Currently the Wine d3d10 implementation isn't particularly complete, but that will change with time.
No, I think the GP is just a Mac person (or hasn't used OOo since the 1.x days). AFAIK the interface for MS Office on the Mac is non-trivially different in comparison to the Windows version, and the GP is pining for the Mac version's interface in OOo. Also, the GP complains about a lack of integration (using the system icons, fonts, etc), which is a non-issue if you're using KDE or Gnome (at least, maybe Windows too, I don't know) as OOo will already use a set of appropriate icons, use the correct file open/save dialogues, etc.
I was attacking this assertion by noting that the trick does not work (and expose_php does NOT affect this one. You're just wrong about that.)
Yes, and you're wrong. I just verified it with both PHP 5.2.2 and 4.4.6, and expose_php most definitely does affect the previously mentioned urls, in that with expose_php = Off, the easter eggs cannot be produced.
So keep on thinking you're clever if you want, but you're not. You're missing the point, arguing the wrong issues, and despite all that you still manage to be factually incorrect.
I'd tell you the same thing, but I'm guessing you'll continue not listening.
This also would lead to security problems. Unless you can get the resources to do it all yourself, which would be pretty problematic give the no-cost nature, you are going to rely on distributing this to volunteers. What happens if those people are evil or incompetent? Say someone offers to run a distro for you. They are willing to dedicate good servers, give you good dedicated bandwidth, and have plenty of burstable capacity, and are a legit university. So you go for it, this is the kind of thing you need. What you don't know is that the guy who runs it is a moron basically. He's been working there for 50 years and most of his knowledge hasn't been updated in nearly that long. So the system is insecure, the malware people get in it, and you start officially distributing malware.
Most of this attack vector can be removed by cryptographic signing of packages, and AFAIK, every package management system already does this. You download the signatures from the main "trusted" server(s), then you can grab the packages themselves from arbitrary mirrors and then verify their signatures. This means you don't need to trust all (or any) of the mirrors, while you can still offload the vast majority of the bandwidth usage to them. This still isn't perfect (you still need to trust a small subset of the servers, as well as your DNS servers, etc), but it's much less bleak than the situation you present.
That that url doesn't work only shows that either expose_php is off, or that the page isn't powered by PHP. Allegedly it did work at the time of that blog posting, and given that everysingleother37 Signalswebsite is powered by PHP, it seems pretty plausible that rubyonrails.com also uses PHP. Regardless, I don't particularly care either way about that point.
You attacked the original poster's method, saying that it would give false positives, which it clearly does not. I pointed out the flaw in your argument, and you assumed that that somehow meant that I agreed with the original poster, which I do not (I don't disagree either, however).
I'm not defending anything, I'm merely pointing out the blatantly wrong.
Oh, for comparison's sake, you would have to fully load a T1 connection over a quarter of a month to hit the 120GB limit. You would have to be using more than half a T1 connection to hit the 250GB mark. Cable is a shared resource. If you need a dedicated resource, maybe a T1 is right for you.
That might sound like a lot of bandwidth, but a T1 is really not very high bandwidth at all, they're about 1.5 Mbit/s (both up and down). In comparison, that's about 1/10th the speed of a typical broadband connection around here. Also, if you were to fully load said broadband connection (15 Mbit) for one day, you'd be well over 100G already.
I'm using a C2D T7400, with a Radeon X1400 and gmplayer.
Setting TexturedVideo "Off" leaves me with no Xv support at all, so I can't compare with and without. CPU usage with AVIVO was about 55% in X and about 35% in gmplayer.
A similarly powerful (at least in terms of CPU power) system (Athlon X2 4200+, Geforce 7950 GT) drops quite a few frames on high bitrate areas of some videos but plays the rest of them fine.
It should be enabled by default on sufficiently new drivers (within the last year and a bit). If you run xvinfo, you'll see something like this if it's working:
Slashdot Mirror
That assumes that you have Windows machines, of course...
I think Ubuntu simply may be stretching their resources too thin. It's quite clear that they put a *lot* of customization into the packages they ship (including KDE), but IME Ubuntu feels a lot less stable than Mandrake (many versions ago, my first Linux install) or Gentoo (my current Linux installs). The customizations clearly add value, as most people seem to prefer Ubuntu, but I shy away from it for my own use as it strikes me as unstable (and I much prefer Portage to Apt).
My KDE sessions normally last until the next version of KDE is released (and then I upgrade), and that's the way I like it.
Well, no wonder, the T60 isn't a 17" laptop... T60s come with 14.1", 15.0" or 15.4" displays.
It is actually. ALSA's dmix allows multiple programs to use the sound card concurrently and transparently, regardless of hardware support. Dmix has been on by default for at least a year or two now (since ALSA 1.0.8, IIRC), and you could enable if you wanted it before that.
So, is a RAM drive not a drive then, simply because it doesn't use SATA/IDE/etc?
I don't see how it matters at all that the FusionIO uses PCIe instead of SATA, you're still going to use it like a regular hard drive (or solid state drive, for that matter).
If the FusionIO was exactly the same except it connected via SATA instead of PCIe, would you consider it a drive then?
You could always play Game! - The Witty Online RPG.
Hmm, that's interesting. I've got one of these:
...and its worked flawlessly with snd_hda_intel since I got the machine about a year ago.
If it's still slower, it's only by a tiny margin now. fglrx 8.42 is *hugely* faster than 8.40 and below. My Radeon X1400 now runs Doom 3 at 37.1 fps (1024x768, medium quality), compared to the 21 fps it got before with the same settings.
fglrx still has issues, but its 3D performance is no longer one of them.
You did something very wrong, modern laptops can build a kernel in under 5 minutes:
That doesn't work anymore:
Alky is vapourware, don't hold your breath waiting for it.
As the posters above noted, you can already use (most) Wine dlls on Windows. Currently the Wine d3d10 implementation isn't particularly complete, but that will change with time.
No, I think the GP is just a Mac person (or hasn't used OOo since the 1.x days). AFAIK the interface for MS Office on the Mac is non-trivially different in comparison to the Windows version, and the GP is pining for the Mac version's interface in OOo. Also, the GP complains about a lack of integration (using the system icons, fonts, etc), which is a non-issue if you're using KDE or Gnome (at least, maybe Windows too, I don't know) as OOo will already use a set of appropriate icons, use the correct file open/save dialogues, etc.
Yes, and you're wrong. I just verified it with both PHP 5.2.2 and 4.4.6, and expose_php most definitely does affect the previously mentioned urls, in that with expose_php = Off, the easter eggs cannot be produced.
I'd tell you the same thing, but I'm guessing you'll continue not listening.
Most of this attack vector can be removed by cryptographic signing of packages, and AFAIK, every package management system already does this. You download the signatures from the main "trusted" server(s), then you can grab the packages themselves from arbitrary mirrors and then verify their signatures. This means you don't need to trust all (or any) of the mirrors, while you can still offload the vast majority of the bandwidth usage to them. This still isn't perfect (you still need to trust a small subset of the servers, as well as your DNS servers, etc), but it's much less bleak than the situation you present.
You make far too many assumptions.
That that url doesn't work only shows that either expose_php is off, or that the page isn't powered by PHP. Allegedly it did work at the time of that blog posting, and given that every single other 37 Signals website is powered by PHP, it seems pretty plausible that rubyonrails.com also uses PHP. Regardless, I don't particularly care either way about that point.
You attacked the original poster's method, saying that it would give false positives, which it clearly does not. I pointed out the flaw in your argument, and you assumed that that somehow meant that I agreed with the original poster, which I do not (I don't disagree either, however).
I'm not defending anything, I'm merely pointing out the blatantly wrong.
If you actually read the link in the GP's post, you'd find out that he's not trolling. The PHP easter eggs only work on PHP pages, not any page served by a server that supports PHP. For example, compare http://otc.dyndns.org/foo.html?=PHPE9568F36-D428-11d2-A769-00AA001ACF42 and http://otc.dyndns.org/?=PHPE9568F36-D428-11d2-A769-00AA001ACF42
I don't think you realized the alternative meaning there, but I found it amusing.
That might sound like a lot of bandwidth, but a T1 is really not very high bandwidth at all, they're about 1.5 Mbit/s (both up and down). In comparison, that's about 1/10th the speed of a typical broadband connection around here. Also, if you were to fully load said broadband connection (15 Mbit) for one day, you'd be well over 100G already.
Probably 16-bit colour instead of 24 bit colour.
Try a
instead. (You might need to first.)Ah, but it's easy to placate your customers when you don't have any.
I can easily get ~6 hours of battery life out of my Thinkpad T60 with the 9 cell battery pack.
We have Best Buys as well. Despite being basically the same store, they often vary (significantly) on prices for the same item.
I'm using a C2D T7400, with a Radeon X1400 and gmplayer.
Setting TexturedVideo "Off" leaves me with no Xv support at all, so I can't compare with and without. CPU usage with AVIVO was about 55% in X and about 35% in gmplayer.
A similarly powerful (at least in terms of CPU power) system (Athlon X2 4200+, Geforce 7950 GT) drops quite a few frames on high bitrate areas of some videos but plays the rest of them fine.
The X2 would probably be able to play all of the above without issue if it was overclocked 10-20% though.
It should be enabled by default on sufficiently new drivers (within the last year and a bit). If you run xvinfo, you'll see something like this if it's working:
If it's not working, try adding
to your device section, like so. Also make sure DRI is working, as it's required for AVIVO.Google doesn't seem to yield a whole lot of information about about AVIVO, but here's the release notes from the first release with AVIVO support.