Slashdot Mirror

We managed to update running systems with package management for how long?

Yes, but we didn't actually manage to make it work with modern desktops. Firefox and Chromium both fail when the packages of libraries that they depend on are moved underneath the running process. It is fine to say that this is a package manager problem, but it has been a known problem for years and is still unfixed (telling the sysadmin that his users should all manually restart Firefox is not a seamless solution, even assuming that he notices the message and acts on it). You can hardly blame people for wanting a more robust desktop, with applications that don't start randomly crashing when the sysadmin (or an automated script) runs a background update.

Is that the same bug? Looks like the bug that the post is talking about is here:

http://security-tracker.debian.org/tracker/CVE-2006-0744

Fixed a long time ago.

Easy victory for debian, as there was no manager who had to wonder how fixing the bug ASAP vs. schedule the fix for later could potentially affect his career...

According to Debian's security tracker, the stable version is still vulnerable (and unstable was fixed only two days ago).

"I understand that they want the interface to be easy for anyone to approach, but what about those of us who want to do more than just browse the web and share pictures of the grandkids?"

This is kinda cool. You can choose stuff. It's a fork of Ubuntu or something.

http://www.debian.org/

The Debian and Ubuntu PowerPC ports are alive and well. Main lack for modern use is Flash. But I long dual-booted Ubuntu PPC on a G3/G4. A more reliable DVD burner than Mac OS X 10.4, and wider hardware support.

Debian seems to argue this as well - they point out that in Linux, support for ZFS is unlikely to be there in the mainline kernel.

Which is what makes me wonder - why are people trying this? If ZFS is what is needed, why not go w/ a perfectly good and compatible FOSS unix, such as either OpenIndiana or FreeBSD? The former has the same license as ZFS, while the latter's license is not incompatible w/ CDDL.

If the objective was to get some of the nice things from the Linux world, that might be possible w/ Debian kFreeBSD, which has some support for ZFS. Might be a better option than Linux.

Perl is slower than Javascript for many things- 30 to 100 times slower:
http://shootout.alioth.debian.org/u32/benchmark.php?test=all&lang=v8&lang2=perl

slashdot is written in perl

in todays i7's , perl runs as fast as P4.

http://shootout.alioth.debian.org/u64/perl.php

Sure, a mandolbrot may be way faster in java, but real world solutions, are not 100% math, but a combination of data transforms, logic and db acesss or rest access, with disk io or net io.

And if most of your data is text, ie xml/json/csv, then perl can be as fast as java, but with less layers.

Huh?

http://www.ubuntu.com/download/arm
http://www.debian.org/ports/arm/
http://archlinuxarm.org/
http://www.armedslack.org/
http://maemo.org/

Fedora and Suse are working on an arm version.

All these guys even have a standards committee setup linaro.

I was using this package to track absolutely everything I ate. Its advice was always to eat absurd amounts of fish flakes because fish flakes had something that it thought I was low in. Iron, maybe? So whenever I would be winding down my day and I would ask it what to eat, it would always be something like a pound of dried herring flakes.

I wonder if that's this book's conclusion. The last page just goes something like this:

"tl;dr: eat pounds of fish flakes"

I think you are missing some info: what do you want to do? Depending on what you want to develop, a different language may fit your needs:

• - You say you don't like low-level languages such as C, so I'm guessing you don't want to write Linux drivers or create software for tiny embedded systems.
• - Do you want an scriptable language? Then you may consider Python.
• - Do you want a GUI-friendly language? On what platform? Just Windows? Then use .NET (C#, Basic...). On Linux? Then use C++ with GTK or QT libraries. On MAC? Well, MAC is another world. Notice I'm saying "GUI-friendly", not "GUI-capable". I mean, such languages are often accompanied by GUI designers that autogenerate code for creating interfaces.
• - Portability? You may use Java.
• - Would you prefer writing code in a smart, concise, functional way? Then you should have a look to functional programming languages such as Haskell, Scala or F# (the latter by Microsoft). Depending on what you want to develop, functional programming could save you a lot of time. However it can turn to be tedious to learn.
• - What about Web apps? Are you planning to write desktop or web applications? For web apps you should consider Python, PHP, Javascript and a myriad of XML deriving languages.
• - Now somebody will probably shoot on my head, but what about LISP? Are you planning to develop critical software that need to be modified "on-the-run"? LISP is used in very special environments where code and data are indistinguishible, and has a great macro system that allows you creating "your sublanguage". I started learning LISP and, on the face of it, it seemed to me a terrible ugly language. But, with time, that feeling disappeared.

• On the other hand, a language without a good library is a condemned language, unless it is focused for some sort of special, reduced purpose, where such library is simply a small set of functions written to do a specific task.

  If you are evaluating "speed", then you should have a look to this site:

  http://shootout.alioth.debian.org

  Cheers

Yeah, and it's time to discuss something new. I didn't submit the following story, since it's now a month old, but I did find it interesting that Debian, which has always spoken the language of Free Software, decided to join the OSI.

Why? Debian itself has what it calls its 'Free software guidelines', and in it, Debian follows pretty much the same language as the FSF. However, the FSF refuses to recognize Debian as free software. Why? B'cos Debian 'also provides a repository of nonfree software. According to the project, this software is “not part of the Debian system,” but the repository is hosted on many of the project's main servers, and people can readily learn about these nonfree packages by browsing Debian's online package database.'

In other words, b'cos Debian decides to give its users the choice of using 'non-Free' software if they want to, the FSF blacklists them and doesn't endorse their distro. I fully agree w/ Hairy and Barbara above that this is a religion, and nothing illustrates my point better than this. It's not enought that an organization provide 'the community' w/ 'free software' - it's also necessary that they not provide anyone - not even those who want it - w/ non-Free software. This is what makes the FSF a bunch of fanatics who sane people would avoid.

Debian has done a good first step in joining OSI - something that IMO they should have done many moons ago. Next thing I hope they do is substitute the deceptive term 'free software' w/ 'open-source' wherever they can. Like they say in the announcement, they needn't embrace every license that the OSI has accepted as Open-Source, but at least, by being more business friendly than the FSF, and using more industry accepted terminology, they can get more monetary backing for Debian from the industry.

Yeah, and it's time to discuss something new. I didn't submit the following story, since it's now a month old, but I did find it interesting that Debian, which has always spoken the language of Free Software, decided to join the OSI.

Why? Debian itself has what it calls its 'Free software guidelines', and in it, Debian follows pretty much the same language as the FSF. However, the FSF refuses to recognize Debian as free software. Why? B'cos Debian 'also provides a repository of nonfree software. According to the project, this software is “not part of the Debian system,” but the repository is hosted on many of the project's main servers, and people can readily learn about these nonfree packages by browsing Debian's online package database.'

In other words, b'cos Debian decides to give its users the choice of using 'non-Free' software if they want to, the FSF blacklists them and doesn't endorse their distro. I fully agree w/ Hairy and Barbara above that this is a religion, and nothing illustrates my point better than this. It's not enought that an organization provide 'the community' w/ 'free software' - it's also necessary that they not provide anyone - not even those who want it - w/ non-Free software. This is what makes the FSF a bunch of fanatics who sane people would avoid.

Debian has done a good first step in joining OSI - something that IMO they should have done many moons ago. Next thing I hope they do is substitute the deceptive term 'free software' w/ 'open-source' wherever they can. Like they say in the announcement, they needn't embrace every license that the OSI has accepted as Open-Source, but at least, by being more business friendly than the FSF, and using more industry accepted terminology, they can get more monetary backing for Debian from the industry.

http://www.debian.org/

Where is the download now button? It's just masses and masses of documentation and that's why no one wants to use Debian because it's just a complete mess of policy and other bureaucracy that wastes your time.

By Linux I'm referring to Linux. If one doesn't know that Linux is a kernel then it makes further discussion fruitless, now, doesn't it.

OK, then, if one doesn't know that "Linux" rather than "a Linux distribution" corresponds to "Windows NT's kernel-mode code" rather than "Windows", it makes further discussion fruitless.

I shall assume that when you said "It would be very different, because Linux was written from the ground up by competent engineers with portability in mind. Windows was written by some very competent engineers, and many more with -shall we say - much less competence.", you really meant "It would be very different, because Linux was written from the ground up by competent engineers with portability in mind. Windows NT's kernel-mode code was written by some very competent engineers, and many more with -shall we say - much less competence.", in which case I'd like to see some data to support your assertion that some of the engineers working on said kernel-mode code weren't competent enough to avoid making "[assumptions] about internal representation of data structures, word size, endian-ness, and a host of other issues".

Of course, the post to which you replied was speaking not just of the kernel-mode code, but all the code:

Porting a Linux distro to ARM does not mean rewriting the code from the ground up, it means recompiling with different flags... why would it be any different for Windows?

(emphasis mine), so merely speaking about the Linux kernel would not validly reply to that post. Perhaps some of the code in Windows-as-a-whole was written by engineers who couldn't avoid making said assumptions (although earlier versions of NT ran on MIPS, Alpha, and PowerPC, and apparently it was originally developed on i860 machines, and current versions run on x86 in 32-bit mode and x86-64 in 64-bit mode and earlier versions ran on Itanium in 64-bit mode, so at least for those versions they managed to avoid making some incorrect assumptions of that sort). However, not all of the code in pick-your-Linux-distribution-as-a-whole necessarily avoids making those assumptions, and porting the distribution as a whole might not be a snap (although Debian, at least, supports several architectures, so either the Debian packages that work on those architectures avoided making those assumptions or got them fixed when they failed to work on those architectures).

Why not just use Debian with KDE? http://pkg-kde.alioth.debian.org/

Are you looking for something in particular that is only available in some distro?

I'm rather disappointed to report that every major distro has let me down on updates and patches over the years. I really have to question the value of a "maintenance service" if the maintenance introduces bugs instead of fixing them. (RedHat, Mandrake, SuSe, and now Ubuntu have all done this at one time or other.)

Debian did it too with openssl.
http://www.debian.org/security/2008/dsa-1571

Luciano Bello discovered that the random number generator in Debian's openssl package is predictable. This is caused by an incorrect Debian-specific change to the openssl package (CVE-2008-0166). As a result, cryptographic key material may be guessable.

I hate this rapid-release fad. The downsides far outweigh the upsides for me.

Solution: http://www.debian.org/releases/stable/

"In my experience Go is probably 5-10x faster than C or C++. I estimate that programs that take me half an hour to write in Go would take me about 2½ hours in C and 5 hours in C++."

Extraordinary claims require extraordinary evidence. And the evidence that is out there states that the claim Go is 5-10x faster than C or C++ is bullshit. In fact, it's slower.

Notice the benchmark where Go took 30x the time the C program did to finish the benchmark?

Well said. I'd say open source projects are successful when they give those contributing what they want. In my case, I volunteer for projects that help blind/VI people, projects like Vinux - Linux for the Vision Impaired, SpeechHub - free voices everywhere, NVDA - the free screen reader for the blind, and Orca - the Linux screen reader. I also contribute algorithms, such as libsonic - speeding up speech for speed listeners, and an enhanced FFT algorithm for speech recognition.

So, my win is helping the blind and otherwise disabled with computing technology. In Open Source Land, it's whatever floats your boat.

Wow! Rarely seen concentration of FUD in this post, you should be up for a promotion, or a bonus of some kind.

Reality is :
-it's way (I mean WAYYYY) faster and easier to install Linux on just about anything than to get windows working
-on the various machines I've used to make a living as a developper over the last six years (currently EeePC), everything worked seamlessly right out of box

To check it out :
Burn a Knoppix CD (or USB key) and boot your machine with it, you'll get a risk free trial (you should have networking and office suite all up and running)
http://knoppix.net/

Once you love it :
install Debian, following instructions here at
http://wiki.debian.org/QuickInstall
(other distributions exist)

This will preserve your Windows partitions, and the content will be accessible easily from Linux

You'll soon discover that computers are actually fun.

Mod parent up, the firmware still remains proprietary. See e.g. http://packages.debian.org/squeeze/firmware-linux-nonfree

Be sure to read the copyright file which includes gems like "Binary redistribution", "All Rights Reserved", "No reverse engineering, decompilation, or disassembly of this Software
  is permitted"...

Doesn't sound too free to me.

It was a typo, I actually meant 2-3 orders of magnitude. And yes, I mean 100-1000 times slower. Well, actual measurements show Python stuck overall about 2 orders of magnitude slower than C++ with some excursions out in the general direction of three orders of magnitude. That is for compute intensive work, which of course is the topic here. I personally have measured Python running up to a thousand times slower than C++, running things like the good old sieve bench.

So no, expecting a Python program that actually does any significant amount of computation to come as close as 2-3.5 times slower than C++ is wild fantasy.

The real reason why Python will still be slower than Java is because it's a much more dynamic language.

That's actually a rather bad excuse. Dynamic languages don't need to be that much slower. For instance, have a look at JavaScript V8 or LuaJIT 2 for some incredibly fast dynamic language implementations. If I recall correctly, last time I checked, LuaJIT was about a factor 3 behind optimized C code in the median case on the Computer Language Benchmarks Game (although it seems to be no longer included in that benchmark for some reason).

Yep. Debian gets them from the VLC website. I'm running Wheezy.

http://packages.debian.org/sid/libx264-118

There's a debian squeeze port available now right here. The problem is that it was never compiled to do floating point instructions in hardware, so you're going to lose some seroius performance by using it over Fedora.

There are two "stock" Debian ARM distros. The one in stable (the "Arm EABI" port) doesn't support floating point. There's also one in the unstable branch called "armhf" which has support for ARM hardware floaitng point, but only for ARMv7 and up. Raspberry Pi is ARMv6 (notes for armhf platform are here..

Ah, young idealism, trying to be the Debian. I was there, once. It is true that it's better to have open-source drivers, but you need a stable, open, documented hardware platform. PCs are, Android is neither.

Debian includes access to a "non-free" official repository that his non-open-source drivers. Please don't refer to the Debian Foundation as "young ideal[ists]." They've done a great job balancing idealism and pragmatism.

PCs are not an "open, documented hardware platform." Here's an interesting thread from 2004 about this same issue debated at Debian http://lists.debian.org/debian-vote/2004/02/msg00136.html

Aterm seems to be ok.
i don't see it listed as using libVTE

http://packages.debian.org/wheezy/aterm

Debian hasn't packaged 1.2 yet, these are third-party packages.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=585409

Apparently one of the issues why the newer versions can't be packaged is that the maintainer wants to package and upload all versions between the last one and 1.4 in order. Since nobody has the time to do so, there isn't any progress towards packaging the newer ones.

Did you look at the list of files? libgnustep-base-dev clearly includes files like NSObject.h. Unless Debian went out of their way to write their own shims to reproduce the functionality of the GNUstep runtime based on the GNU objc runtime (which uses Object instead of NSObject), it seems far more likely to me that the GNUstep runtime is included in the libgnustep-base package. Both runtimes can co-exist just fine, to the best of my knowledge.

Yes, I'll admit that the linkage to libobjc2 is curious, but that may be nothing more than some minimal GCC-based dependencies on startup code or the like. I'll no longer assert with perfect certainty that your hypothesis is incorrect, but I still find it highly implausible.

But in any case, the bottom line is that even if the GNUstep runtime isn't included, it has absolutely nothing to do with the license. Unlike the FSF, the Debian project has absolutely no preference for any DFSG-license over another. If they aren't both there, it's most likely because nobody has taken the effort not only to package both but to make sure they can co-exist and cooperate smoothly and effortlessly. If that's the case, the first thing I would do is talk to the packagers, and find out what the issues are, then see if I can find someone who is willing to solve those issues. It doesn't even have to be a Debian developer--you only need a developer to audit the packages once they're ready, and sponsor the upload.

But first you should make sure you're correct, because, as I said, it seems extremely implausible.

Your point was funny and well illustrated, but I'm not sure it's correct. Is Debian actually biased toward the GPL over other F/OSS licenses? Their Debian Free Software Guidelines and Software License FAQ explicitly suggests the BSD and MIT licenses for authors who want their code to be useable by everyone. They also call out the Artistic License by name in the "What Does Free Mean?" section of the "Introduction to Debian".

I've never thought of Debian as particularly pro-GPL in particular so much as pro-Free Software in general.

Your point was funny and well illustrated, but I'm not sure it's correct. Is Debian actually biased toward the GPL over other F/OSS licenses? Their Debian Free Software Guidelines and Software License FAQ explicitly suggests the BSD and MIT licenses for authors who want their code to be useable by everyone. They also call out the Artistic License by name in the "What Does Free Mean?" section of the "Introduction to Debian".

I've never thought of Debian as particularly pro-GPL in particular so much as pro-Free Software in general.

And doesn't Debian actually actively work for make sure the packages it distributes are GPL?

Not at all. They just tend to make selections of the projects which actually work rather than the hundreds of projects that never go anywhere. The Debian Free Software Guidelines mean that main distribution software has to be free, but basically anyone who has motivation and acceptable software can get their package in.

Simply put, if a package isn't in Debian then it mostly very specialised, quite new or isn't worth touching. If there are several Debian packages and you don't know which to go for, then go for the one which is in Red Hat since that will be the most professionally maintained package.

The first survey may have been representative of packages which people start developing, but this is more representative of packages which are actually useful.

And doesn't Debian actually actively work for make sure the packages it distributes are GPL?

So, not only is he cherrypicking but he picked a project that strives to use Copyleft.

http://www.debian.org/News/2012/20120219

The actual study mentioned in the talk came out last month and was written up here.

http://www.itwire.com/business-it-news/open-source/52838-gpl-use-in-debian-on-the-rise-study

John Sullivan even called picking only one distribution as "scientific". I'm not sure he knows what the word means.

rpm has many more features that dpkg.

Debian gets around this with Debian Policy which is why Debian takes so long to do a release, and why even Debian unstable tends to be pretty stable.

http://www.debian.org/

I think it's a fork of Ubuntu or something.

This is why the episode of BSG where they didn't want to network the systems together because of the Cylons hacking in remotely is so laughable

It's a little less laughable when you consider that the Cylons owned most of the electronics manufacturing business on the twelve colonies.

(it would take a single firewall rule in that case... deny all incoming traffic)

That's a nice thought, but it doesn't help when your firewall switches to "allow all" once it sees the right magic packet. Which is exactly what happened in the pilot episode.

there's consumer grade encryption available that far exceeds the capability of the most advanced military computers to crack within a practical amount of time

There sure is. But all it takes is one little "mistake" to turn it from unbreakable into child's play.

Imagine a world where one company in, say, China makes more than half of the world's consumer electronics, including parts used for high security applications. In such a world it would be easy to see why people lie awake at night dreaming of Ken Thompson style hacks.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=461309 shows why xmms was kicked out of debian. Presumablly ubuntu just followed them. Not sure about the others.

System76 and Zareason are two US vendors I've heard of which sell all Ubuntu machines. Debian also keeps a solid list of vendors here http://www.debian.org/distrib/pre-installed

As far as I know the differences in directory layout you've mentioned flowed from RedHat's RPM packaging down to Mandrake. They did some useful innovation in addressing the desktop section of the market that RedHat abandoned; part of their problem is that they never offered a compelling server alternative to RedHat that was different enough to distinguish itself.

In the case of PostgreSQL, there definitely wasn't any conscious decision on the part of RedHat/Mandrake here. By default when compiled from source, PostgreSQL expects the database configuration file to be in the same directory as the database itself. RedHat and Mandrake don't change that; Debian goes to a lot of trouble to move those to /etc instead. And even that's not complete, one of the configuration files needed for replication setup (recovery.conf) still goes into the database directory, even on Debian.

Agreed, I actually like Gnome3.

Disclaimer: I spend most of my time in a Windows environment (not by choice, it's the work I need to do) so when I use one of the modern Linux GUIs it's mostly just to try it out or on friends or relatives' PCs. I've used them, but not for months at a time. I do enjoy them though.

FWIW, if you don't like any of the new stuff and prefer ye olde Gnome2, then Debian is still happy to provide you with an excellent OS.

Since this bug was introduced in Linux 2.6.39 Debian Stable (squeeze, Linux 2.6.32) is not affected. Unstable(sid, Linux 3.1) has already been patched, though Testing (wheezy) is still vulnerable.

More information here

no, I'm from the present. you know that is based on FreeBSD 8.1

No - you really are from the past.

that Debian frankenstein is not production stable, just a trial balloon.

Wrong again. And you don't get to determine what's "production stable" as you clearly don't have a clue what you're talking about.

We're talking about Debian GNU/LInux compared to FreeBSD here

You're talking pure bullshit. Do you even use BSD? (not that anything you say has a shred of credibility).

I looked it up. The NASA Open Source Agreement 1.3 is OSI certified, but the FSF deems it non-free.
Since NASA World Wind is in Debian's nonfree repository, I assume that would be where this will go too.

"The NASA Open Source Agreement, version 1.3, is not a free software
license because it includes a provision requiring changes to be your
“original creation”. Free software development depends on combining
code from third parties, and the NASA license doesn't permit this."

http://lists.debian.org/debian-legal/2011/04/msg00075.html

Seems you really didn't get it at all! The standard is called UEFI secure booting, and that's the main topic of this slashdot news. If you want to know more about it, and understand how it works, read this recent post on debian-devel: http://lists.debian.org/debian-devel/2012/01/msg00168.html

I run Debian 6 on both physical hardware and VMs that have less resources than you've described, and it works perfectly fine. For desktop environments, I use either plain Gnome or XFCE. For the VMs, I access their desktops by running FreeNX, and on both my LAN and over the Internet it's almost as fast as a native desktop. These machines are routinely used for web browsing, email, and office documents.

The answer is obviously yes, http://packages.qa.debian.org/k/kfreebsd-9.html

Debian will get there someday

Are you from the past?

Replace FreeBSD with Debian,

*cough*.

Replace Ubuntu.

TFTFY