DD-WRT is open source in the same sense that the original Linksys firmware was open-source. Clearly, the GPL parts are open source, including all kernel modules and command line tools based on BSD/Linux. And yes, it must be possible to compile a bootable image with minimal shell support (otherwise they wouldn't be complying with the GPL). However, (this was true two years ago--haven't checked sense) DD-WRT has several binary blobs and closed-source components that handle higher-level tasks (for example, at the time I was looking into this, it was not possible to extend the webserver.)

Additionally, DD-WRT was still on the age-old nvram model of configuration, rather than using a read-write overlay filesystem to allow editing any configuration files. This means that some things were a pain in the ass to change once you have flashed the router, and building a custom image requires compiling a 10GB svn checkout. I'm sure you got it to compile, but I'm just saying that compiling isn't as easy as it should be. I (as did many other angry slashdotters) wasted several hours trying to compile DD-WRT. This is why the words "open source" in the description gave such a backlash.

Anyway, I didn't bother to figure out the compilation process, and I just went over to OpenWRT for my Linksys WRTSL54GS (kernel 2.6 broadcom with b43 -- works really well), Airlink AR-430W, and D-Link DIR-615. They all work really well.

That said, DD-WRT is a fine firmware if you want something that works and does more than the default images--I have friends who love it. It does Client Bridging which is the one feature I sorely miss from openwrt. So in my opinion it's a good choice if you are the sort of person who wants things to work and doesn't plan to write scripts or tweak things from source. And because fewer things are configurable and Brainslayer tests it on a ton of routers, you can be sure that an image will work on your hardware without tweaking anything (if it's on the Supported Devices list).

This is why an encryption key is never "temporary" -- it shows no discretion on the part of the journalists to leak a key. This is not a password that can be revoked--it's a key. If you have a key for your previous house, you don't ever give the key away while telling people the address -- the lock has probably not been changed.

Honestly I don't know why he didn't use SCP or SFTP, giving the journalist the fingerprint+password over a second channel... It's easy to revoke a password, and hard to MITM the leap-of-faith while maintaining the correct fingerprint. But hindsight is 20-20... I wouldn't have thought of this issue either.

I know most people are complaining about the irony of a leak at wikileaks, but has nobody considered the fact that the gpg-encrypted file was publicly available on a "temporary server", probably for at least a few hours (it must have taken Leigh some time to drive home and start the download).

At the time, wikileaks may not have been as popular, but it's not a stretch to imagine somebody was randomly browsing the IP address of that "temporary server" at the time, and noticed the encrypted file. Wikileaks is not your ordinary file host with uninteresting data on it--every file on there can be considered politically sensitive, and it may have been downloaded by several governments the instant Assange started the http daemon.

So it's not a stretch to imagine somebody downloads the file and leaves it on his hard drive waiting for the password to come out. Heck, I may have done this once or twice to the "insurance" file--and the only thing more obvious than "insurance" is a file named "cables.gpg".

It would be kind of hard to send that message in five minutes in this universe. Perhaps the speed of light is different where you come from.

I think you meant five million years behind--but I'll excuse you, the keys are like right next to each other.

NO, that is not a solution! You're doing exactly what the RI/MPAA want. The RI/MPAA have a secret, and that is that they want people like you to pirate their movies, because some of your friends will then pay money to watch them (since they have convinced the public that piracy=murder).

If you really want to stop the movie industry from bribing our public officials and criminalizing us, you should boycott the movie industry. This means, *do not consume* music/movies covered by RI/MPAA--whether you paid for something legally or not is a moot point (unless you get caught).

If you squint your eyes at MPAA, you will notice that the last letter in the acronym stands for *America*. America does not have the only movies/music in the world, and there is nothing these corporations would fear more than American media becoming irrelevant. A boycott of american media (or election reform to prevent bribery of public officials through "donations") is the only way we can stop corporations from controlling our laws and controlling us.

There are other countries that produce music and movies, and some of it is as good or better than our mass-produced hollywood media. If everyone watched half as much American media and watched some movies from other countries, that means 50% less money going to the RI/MPAA, and 50% less bribes to our representatives.

1. Yep, just run tune2fs and enable the ext4-specific features (google for 'upgrade ext 3 ext4') Then, make sure to edit your /etc/fstab. -O extents is the magic that makes them incompatible, but you don't need to use extents to get the benefits of ext4.

2. The rename issue was about bad assumeptions made by some gnome/kde programs about when to call fsync(), and those have long been fixed. I think it was that ext3 used low time to sync, so it was almost impossible to run into this, and ext4 set the *default* sync delay much higher--it is easy to change this in /etc/fstab, so google about this if you want.

I've been using ext4 since before it was supported in most distributions (soon after the announcement that it was marked stable in linux) and I've had no such issues (or I've never noticed). My /home partition has survived crashes due to the faulty seagate drives of 2.5 years ago (on RAID 1) and survived an abnormally high number of daily motherboard/psu-related crashes with no data loss (of stuff synced to disk).

Saying Slackware doesn't support GRUB is like saying Dell doesn't support Linux. It's a bootloader, and aside from installing it, it's completely unrelated to the OS. They probably kept LILO as the default since it's works easily out of the box.

Just grab a copy of grub2, make, make install, install it to the bootloader, and set up a linux64 menu.lst to load into your OS. Unlike LILO, you can actually type in commands at the boot prompt and tab complete to get a list of OS's, so it's kind of hard to mess up if you have the documentation handy.

That said, there's not much point in changing something that works--unless you're intent on booting on new hardware that uses EFI or something.

Make sure you're using EXT4 for your filesystem... it's really simple to upgrade, and you can basically change /etc/fstab, and optionally run some tune2fs parameters to enable extents if you are happy with making it permanent.

Just changing fstab to say "ext4" instead of "ext3" alone cuts fsck time by about a factor of 10 (but make sure your version of grub supports ext4 before turning on extents). My 900GB ext4 raid partition will fsck in roughly the same amount of time as my 20GB ext3 root partition

That's mostly true -- however, my MSI motherboard seems to spend 5 seconds in each of the AHCI controllers, with the annoying "Press Alt-F2 for RAID setup" message (and switching to IDE emulation makes ubuntu's Disk I/O roughly 25% slower in terms of boot time). Aside from that, I can hardly see my BIOS boot logo--it flashes and goes to GRUB.

Unfortunately Ubuntu seems to have poor support for UEFI at the moment... not too long ago, apt-get dist-upgrade was forcing me to remove grub-efi-amd64 in exchange for grub-pc, and I had to do some nasty dpkg hold to avoid that.

I find this number pretty surprising--my 11.04 boot isn't anywhere near 12 seconds, even with UEFI. Even if you leave out the time to start the ubuntuone and nepomuk I/O heavy processes, my system spends roughly 10 seconds before anything prints out (ureadahead and kernel modules).

LightDM should help a fair amount, so I'm happy about that... but I'm worried that's just offloading the gnome libraries to later in the boot process, so I don't think this will help much for time to a usable desktop...that is, unless you use strictly Unity WM and no kde/gnome apps whatsoever. I guess this might not be for me.

Why aren't search paths disabled by default? They seem like a huge security hole. I don't want to be getting "google.com.mitm.comcast.net" when I type in "google.com".

Search paths should be enabled explicitly, since I've only ever seen them legitimately used on corporate networks where they control all the computers anyway.

First to answer your question: I agree about using Opera in business--I actually think that makes a lot of sense for businesses concerned about stability. I'm sure Opera would sell support agreements, and they don't have an insane release schedule, though they manage to keep up with standards. I suspect Opera doesn't have brand name recognition, so no IT manager would bother suggesting it out of fear of a backlash. For example, what if an obscure version of Oracle's timecard crapware fails on Opera? Then you would need to tell people to use IE or Firefox to run internal apps -- and you're back to the same problem.

As for Chrome-- "Google Chrome" is not open source because it includes a version of flash player, a custom trademark, and PDF support, among other things. (Firefox has this distinction too--the difference between Firefox and Iceweasel.). But in general, the developers actually want to release as much as open source as they can. They need to have a really good reason before deciding to make something closed source (usually a result of legal requirements, as in Adobe Flash's redistribution license).

However, Chromium comprises most of the code (enough that most developers modify and test chromium, and wouldn't notice the difference). Chromium is what you should be using: it is shipped entirely under the revised BSD license (it's about the most liberal open source license out there), and you probably can't tell the difference with the exception of missing a few google-specific features like Sync (most of these are open source but disabled by default).

If you want to use the open source version, feel free to download nightlies here (updater is disabled here since it doesn't make sense):
http://build.chromium.org/f/chromium/snapshots/
Or, if you want--just compile Chromium yourself. It's a cakewalk on Mac/Linux -- and it's easy on windows if you have Visual Studio.

Captain Keming to the rescue! (Or, in this case, he may be called Captain Keiviing)

It's okay: we won't have another leap day until 2016, so Microsoft didn't need to code that logic into their software. 365 days is enough for anyone.

Don't worry: in 2015, Microsoft will release a new version, Office 366, which will offer you the full yearly experience for only one of the cheap monthly prices listed below (assuming you pick the right plan)!

  . /- $2/mo for Plan E
  / $4/mo for Plan K1
  \ $6/mo for Plan P
< $10/mo for Plan E1 or K2
  / $16/mo for Plan E2
  \ $24/mo for Plan E3
    \- $27/mo for Plan E4

Pick your plan today, before it's too late. The 366th day cometh!

My title was misleading, sorry. I don't mean to say there is no value in stereo video. Just as stereo and 5.1 audio improve the experience, stereo video has the ability to improve the experience. But I'm tired of the media treating this as though it's a whole new paradigm. (Ever been to an IMAX movie and watched the beginning where they show off the 10 speaker surround sound system? My reaction usually is "That's really cool. But start the movie already")

Personally, I'd rather they spent their efforts on adding more resolution and better touch accuracy to cell phone screens. Hardware space/development time is finite, and 3D screens are going to draw effort away from things that affect my everyday usage of these devices.

The one good thing about this technology catching on is that people may start to record more data in stereo video, which could lead to interesting use cases in things like photosynth and 3D model scanning (this time I mean real 3D as in three axes, not stereo).

Lame... I was hoping this article was about phones getting better EGL or WebGL support for Video Games.

I remember one time when I was testing a "3D" (stereoscopic) device, and it seemed like the coolest thing ever--I felt I really noticed the 3d effect, and it seemed "awesome".

Turns out, the left eye of the glasses wasn't working, and I guess the 3D effect was my own brain inferring the depth the whole time. I'd suggest trying this out if you're ever testing a "3D" technology: cover/close your left (or right) eye and see if it looks more "realiztic" or "cool" than with it open.

Also, as an analogy for audio, how many times have you played mono audio files and not noticed? (hint: anything recorded on a phone or microphone is not stereo) Video is hardly different.

Mod parent up! (Though please don't put WTF and idiot in your subject, or you sound like a troll)

Why are we wasting our time with screenshots when this barcode technology has existed for ages. I'm actually surprised that QR codes haven't really taken off in the US--I guess if the iPhone doesn't support it, nobody cares.

Thanks for that bookmarklet--it'll come in handy.