Copyfree Software

Yes.

4 (80%)

No.

1 (20%)

Total Members Voted: 5

[Alex Libman]

This is gonna be a continuation of my "Software Freedom Scale" thread, since I am no longer allowed to "necro" the old one.  I'll post another one of my "news roundups" shortly.  But first - a simple poll to get things started.

[Zhwazi]

One vote for yes here. I avoid GPL software whenever I can. Mostly for quality reasons though, proprietary and BSD licensed software tends to be good at what it aims to do, whereas any programming wannabe inspired by The Evil Hippie's arrogance that can hack together something in a script that starts with "#!/bin/bash" will put their crap under GPL.

I try to go BSDL as far as I can. My FreeBSD server has an Apache/PostgreSQL/PHP stack running Serendipity. Apache license is the most restrictive in the entire stack and even it's pretty BSD-like in the end. Can't wait until Clang replaces GCC in FreeBSD base. FreeBSD removing the remaining bits of GNU code from it's source tree are a well-deserved "fuck you" to the culty GPL crowd.

[MacFall]

There wasn't an option for "what is this i dont even", so I didn't vote.

[Alex Libman]

You have the option of not voting until you RTFM on permissive (aka copyFREE) vs restrictive (aka viral / "copyLEFT") open source software licenses...  I've been ranting about this for a very long time now...

[Alex Libman]

The "Copyfree Software News Roundup" is back!

• The big news for the past month is obviously the release of FreeBSD version 8.2.  Changes in the core OS include improved Xen virtualization support, LZMA (7z) compression support in tar, stronger crypto, ZFS file system improvements (though that part is restrictively licensed and still far behind Solaris 11), a few new drivers, and bug-fixes.  New release package versions include: Gnome 2.32.1, KDE 4.5.5, Firefox 3.6.13, Gimp 2.6.11, Python 2.6.6, perl 5.12.3, PHP 5.3.5, Apache 2.2.17, and PostgreSQL 9.0.3.  The KDE-based distro of FreeBSD issued a simultaneous PC-BSD 8.2 release with improvements to the installation procedure (particularly partitioning and ZFS support).  The analogous Gnome-centric FreeBSD distro called GhostBSD v2 is still in beta.
  
  The most exciting FreeBSD features, however, are still being held back for version 9.  What might finally compel me to switch from "Copyfreer" OpenBSD is the addition of the permissively-licensed Clang/LLVM compiler infrastructure as a viable alternative to the restrictively-licensed GNUopoly of GCC.  The core system and many key ports (including Chromium) make it through the transition unharmed.  Another great addition will be the ability to finally run FreeBSD on Amazon's cloud framework, which should be stable by the time v9 is released (although, as with most platforms, NetBSD got there first).  Other v9 improvements will include: significant TCP/IP stack improvements, tickless (dynamic tick) mode, and other performance optimizations, as well as USB 3.0 support.  PC-BSD v9 will be the breakthrough release that finally moves away from just KDE and offers users a choice of any desktop environment, as well as better handling of PBI packages with pbi_add.  Progress is also being made in replacing (and eventually removing) the remaining GNU commands from the core system, most of which are rather trivial: cpio, ar, ranlib, bc, dc, find (the BSD version of that command is reaching feature parity with GNU), etc.  But be warned - the current alpha testing versions of 9 are still very unstable, and it's also slower than the production release will be due to the debugging compiler settings and other debugging-related overhead.

• For people sticking with OpenBSD, like myself, the biggest news of the month comes from the Hybrid Source mastermind Sprewell - Chromium 10.0.648 is now available on OpenBSD!  This is just the first compilation, so it's still very shaky, unstable, and slow, but it's a new beginning.  (The previous porting efforts seem to have stalled some time ago...)
  
  I believe -- and, heck, even Microsoft admits -- that HTML5 is the future, so bringing the world's most Copyfree modern Web client to the world's most Copyfree operating system and achieving optimal performance for this combination is very important work.  Of course there are other Copyfree WebKit-based browsers available for OpenBSD, like the minimalist surf, but those are not the kinds of browsers most people can live with on a day-to-day basis.  Chrome / Chromium offers everything that a modern client stack needs: innovative lightning-fast JavaScript, leading HTML5 feature support, a growing collection of extensions, blacklists and other security features, etc.  One innovative feature from Google that I really like is Native Client, which allows secure in-browser execution of x86 compiled code, and it seems to be picking up some momentum.  So, once again, huge kudos to Sprewell for his porting work, and I hope many stability and performance enhancements are yet to come!

• Chad Perrin, the superhero behind Copyfree.org, has been writing many interesting articles / blog posts on sites like Tech Republic, including: Code Reuse and Technological Advancement" (also: Licensing and the Singularity), "When 'Open Source' Software Isn't Truly Open Source", The Difference Between Secrecy and Privacy as Security Concepts" (mentions WikiLeaks), "Key Open Source Security Benefits", "How a Server Side Language Achieves Popularity", "A Skeptic's History of C++", and on pf [2], rssh, logwatch, sysctl, and burncd, among other topics.  Always a great read!

• The February TIOBE programming language popularity index reports remarkable gains for Python, which is still remains my favorite server-side scripting language, as it has been for a very long time.  Python is now at the #4 spot, behind only C/C++ and Java, leapfrogging PHP and making the PHB's who've made me code Perl instead of "that obscure snake language" a decade ago hang their heads in shame!  (Well, not really, and I doubt they'd remember.)  The current stable versions of Python are 2.7.1 and 3.2 (just released), but most UNIX distributions are still on 2.6.x (OpenBSD -stable is mostly still on 2.5.4, although later versions are available, and the most popular Web server OS CentOS is on 2.4.3).
  
  Python's Copyfree status remains imperfect, as is PHP's, but it's definitely Copyfree-er than Mono, Ruby, or Perl.  Not all of Python's components and packages share the same license, however, so a Copyfree purist (and anyone who just wants to avoid confusion and potential legal liabilities) will want to avoid modules like: Git, Paramiko, PyQt, PyGTK, wxPython, PyMedia, Plone, web2py, CubicWeb, SQLObject, Lupy, SimPy, PyMT, Conio, etc, etc, etc.  Be sure to check around and pay attention to licenses for every package you use - there are plenty of Copyfree alternatives available.

• The TIOBE index also shows Java further solidify its #1 spot in programming language popularity, and Java continues to improve in terms of performance as well, but the potential for a viable Copyfree Java stack is looking increasingly grim.  The one project on which I've placed all of my Java-related hopes for the past few years was Apache Harmony, even though it was being developed at a snail's pace, with FreeBSD support being rather lame and support for other BSD's non-existent.  Oracle obviously abandoned that project after acquiring Sun, and in October it was announced that IBM is disengaging from Harmony to back Oracle's restrictively-licensed Java stack instead, which leaves Google as Harmony's sole major backer.  Given the recent legalistic aggression used against it, Google would be wiser to focus its long-term plans on own technology stack, including Native Client and Go.  Now there's something called "IcedRobot" endeavoring "the GNUlization of Android" and moving things from Harmony to the GPL'ed OpenJDK.  So this is the time for Java programmers to strongly consider a plan to move on to something else...

• When jumping between exotic OS'es on bare hardware (i.e. not in virtualization), hardware compatibility becomes a major issue, and the biggest problem usually tends to be wireless connectivity.  Some operating systems support very few (if any) wireless adapters, especially if you need to use the newer 802.11n standard - even Linux and Solaris are often a pain in the butt, much less OS'es like *BSD, MINIX, Haiku, House, QNX, etc.  And the drivers that are present are often buggy, incomplete, offer limited encryption features, etc.  Fortunately all those problems have a simple hardware solution - use a "universal" wifi adapter like NetGear WNCE2001 (currently $59.44 if you search for it on Newegg or Amazon).
  
  This device connects to a standard Ethernet port and doesn't require your operating system to know anything about wireless - all configuration is done via a simple Web-based interface served by the device.  It will work with anything that has an Ethernet port - old computers without USB, Macs, video game consoles (you may need to hook it up to something with a Web browser first to configure it), DVR's, routers (use your old cheap wired hub to set up a wireless bridge), etc.  Ethernet also offers the possibility of using a much longer cable than USB, so you could more easily place it closer to a window, on a car roof, or wherever else the signal is best.  Plus you'll never have to worry about losing the driver CD and not being able to reconnect after reinstalling the OS, as often happens with Windows.  So if you're thinking about buying a USB wifi adapter, I would strongly recommend getting an Ethernet one instead.

• "Free Software Hero Attacked by Communist Fanatic" - that should have been the headline of this article covering Stalinman's bashing of Google Chrome OS.  And, needless to say, his site is still an endless torrent of calls for government violence - unions, taxes, regulations, luddism, government control of media...  Don't let the parts you agree with fool you - all tyrants initially claim to support "freedom", which they define as them being in control.  Sample quote: "evidence shows Obama's economic stimulus worked - and that right-wing budget cuts will cause disaster".  When you use GNU software and don't speak out against it, this is precisely the kind of philosophy you are endorsing!  Silence implies consent!

• Things not being covered include: BSD podcasts (no new episodes in all these months), DraglonFly's lousy benchmark performance (give it time), BSD Magazine (I can't stand PDF's), Haiku OS (things seem to have slowed down), PostgreSQL 9 (it's awesome, but I've been in a NoSQL mood lately), the OpenBSD backdoor allegations (they're FUD), and of course OpenBSD / AerieBSD failing my "culture test" (nobody's perfect).

And, in conclusion...  More Devilettes! 

[LTKoblinsky]

Didn't RTFM (read the...?), but I've got Ubuntu on this laptop. I'm decently satisfied with it, but try to avoid all of the commie, hippie BS behind it. So, is this BSD an OS?

[Zhwazi]

Rant warning.

I was interested in OpenBSD for a little bit. I eventually got turned off by the indignant attitudes and disregard for security understood to mean anything deeper than lack of vulnerability. No RBAC or ACL's, nothing like jails or zones except the insufficient chroot, it's kinda sad. I also can't stand hero worship. A community where dropping names and following drama is the norm is not a community worth participating in, and OpenBSD's community has a lot of name dropping and drama. You never see that in the FreeBSD community. The FreeBSD foundation puts out a statement against the GPLv3, and it's very formal, with no bitterness or disrespect to be found. The analogous event in the OpenBSD community was when RMS himself decided to troll (successfully!) the OpenBSD people on their "recommendations" of unFree Software and stirs up the hive.

Who is the hero of GNU? Richard Stallman.
Who is the hero of Linux? Linus Torvalds.
Who is the hero of OpenBSD? Theo de Raadt.
Who is the hero of FreeBSD?
...
Who is the hero of FreeBSD? The biggest names you'll hear dropped aren't specific to FreeBSD, but generally to all BSDs and Unix in general. Marshall Kirk McKusick is sort of a visible name, but isn't anywhere near the relative hero stature for FreeBSD of the other people named. And he's been doing BSD for what 25 years, that aside he's not much of a hero anymore than Dennis Ritchie or Ken Thompson or Bill Joy, these are the names of prominent Unix developers, not of prominent FreeBSD developers. FreeBSD isn't a personality cult, it's a community.

The Linux community is full of microcosms of this, with every wannabe hero starting a new distro to try to scratch their itch. There's little standardization or consensus, everything is about you should do it this way and how everything else either sucks or is only "okay".

The FreeBSD community is more community than personality cult, which cannot be said of the other projects I named. The biggest fork of FreeBSD right now is probably PC-BSD, and PC-BSD is almost deferential in respect to FreeBSD, including a plain vanilla FreeBSD installer on their installation media.

They don't start up hissyfights about the terrible things that make one sound system suck (OSS) and try to reimplement an entirely new audio subsystem (ALSA) that has been nothing but problems since its inception but which remains backward compatible with the old API because nobody wants to use the new API despite marking the old one as depracated. (They're doing the same thing to X with Wayland. They're adding a new layer of complexity that nobody will use directly so they're going to be running X in Wayland anyways. Way to go Linux idiots.) They don't change their startup control mechanism every 3 years, they don't change device enumeration and abstraction every year, they don't change the userland sound system every time the wind changes direction, they don't acrete 50 incompatible filesystems nothing else ever uses, and they don't bump version numbers just to garner excitement after noticing it dying down. They don't reject great technology like DTrace and ZFS and GCD, they don't use gross misunderstandings and complaints of rampant layering violations to comfort them away from their cognitive dissonance, and insist that they can reimplement the same thing better because they're fucking Linux and they can be everything to everyone and power your router and the world's most powerful supercomputer hooah.

FreeBSD's community is open, collaborative, cooperative, respectful, and generally disinterested in stupid drama. It's the biggest operating system community with these traits. And the OS is nice too.

End Rant.


It's interesting how you get such different mindsets for software from people with different license philosophies. Copyleft software is generally quick and dirty, the implementation sucks and it's constantly being rewritten or reimplemented. Copyfree stuff is generally higher quality, even if it usually has less bells and whistles. Compare PostgreSQL vs MySQL, tmux vs screen, obviously BSD vs Linux, zsh vs bash, clang vs gcc.

I think copyfree software is winning in the long run. They appear to get 90% as much done with 10% of the effort and nowhere near the corporate backing and manipulation. A copyfree desktop would be great, I've tried to use enlightenment and was impressed by how well it does what it does, but it's still not at the point where I can replace my Qt apps and KDE with EFL apps and Enlightenment.

Didn't RTFM (read the...?), but I've got Ubuntu on this laptop. I'm decently satisfied with it, but try to avoid all of the commie, hippie BS behind it. So, is this BSD an OS?

It's more like a family of OSes, including the three big BSD's (FreeBSD, OpenBSD, and NetBSD), Mac OS X (mostly FreeBSD in the middle layers), and to an extent Solaris (less so since the end of the SunOS days). When people say BSD they're mostly talking about FreeBSD, NetBSD, or one of the various derivatives or forks (OpenBSD, DragonflyBSD, lots of smaller ones).

The BSD's have none of that commie hippie BS. There is a little bit of distaste for Windows, but unlike most of the Linux community you'll find, they aren't under any kind of delusion that they're going to destroy proprietary software. I like to put it as, "BSD is for people who love free and open software. Linux is for people that hate proprietary software."

If you're thinking of playing with it get PC-BSD. BSD's are generally rough around the edges, but simple, and if you're like me in this regard you might prefer a simple and robust tool to a beautiful but fragile one. It does almost all the same stuff, and if you must have something that doesn't run natively on BSD it has a Linux binary compatibility layer and Wine for running Windows stuff.

[Alex Libman]

That's an excellent "rant", Zhwazi, thank you!   

I agree with most things that you've said, but I'm a bit less optimistic about FreeBSD.  It is almost as bloated as Linux, with its own accumulation of restrictive licensing crud.  I was famously [2] banned from the official FreeBSD forums (and had ALL of my posts deleted, even constructive / technical ones!) by that fascist asshole DutchDaemon for daring to express a contrary opinion in a thread calling for government violence against Sun and Oracle shareholders...  That is something I do not ever forgive!      (Ian is an exception I guess.)  The sad truth is that BSD projects get even more government funding than GPL ones, and their culture reflects that quite a bit...  At least OpenBSD is a bit farther away from Berkeley and a bit less likely to get government grants.

I don't mind the BDFL / "hero worship" if the guy deserves it and isn't a jerk, because in a Copyfree project people are always free to fork it.  From what I gather Linus is an awesome guy, but sometimes he's too nice, and that leads to bloat - kernel configurability has made it a non-issue, but there might be scaling limits to that as well.  Theo does deserve the recognition that he gets, and he has good reasons for excluding some technologies that people who know what they're doing can do without.  I disagree with his inclusion of perl and more userland GNU shit than he had to (especially the super-fugly FVWM), but aside from that he's kept things pretty lean.

OpenBSD's security is aimed at a particular set of applications, or it otherwise becomes a joke.  For example, its inferior package repository and lack of binary updates for -stable leads a lot of people (including myself) to put their "faith" in a lot of side-projects, like that Chromium binary I've just mentioned, or www.openbsd-stable.org (which I even mirror).  There are also half-dozen other packages that aren't in ports so I have to fetch and compile them myself.  If any of those sources ever becomes malicious, or is hijacked by a third party, a lot of OpenBSD machines are going to get seriously fucked!  (Especially if people running OpenBSD on the desktop are storing ssh keys to other OpenBSD servers.)  Even on Windows you can usually limit your downloads to just Microsoft and one trustworthy freeware downloads site!

Regarding a "copyfree desktop" - that's an aging paradigm.  The Web browser is the new desktop!  OK, you still need xterm with lots of command line / textual interface apps (including for P2P downloads, chat/IM, etc), mplayer for video, an IDE (if you can't do everything in vim), and a tiling window manager like wmii.  This stack still relies on a few GPL'ed components, but their number will shrink eventually.  I even install Gimp when I need it (and uninstall it when I don't).  However I think in the long run everything will be doable from inside the browser, including tab / tile / window management, command line interaction, etc, etc, etc - even high-end multimedia editing and 3D games!  And it wouldn't necessarily be "trusting the cloud" because the server-side components can run (and/or cache stuff) locally as well.

[Zhwazi]

I have doubts about the feasibility of a browser replacing native applications. The responsiveness just isn't there and computation can be done much faster in native, machine-and-OS dependent code. There's definitely a lot that it can do, but we aren't there yet.

As it currently stands there aren't any serious operating systems that don't have a history of either corporate or government sponsorship, directly or indirectly. Minix I don't believe gets government grants but it's not quite a serious OS for general use yet. I think this is more an issue of, with all the government induced market insanity, this is how it gets done. I'd take a government grant as well, they'll tax me out of it in the long run anyways. OpenBSD did get military grants but as you said, probably not a lot more on the way.

By the way, does OpenBSD still have a giant kernel lock or is its SMP reasonable yet? Last I used it (4.5 I think?) the kernel was single-threaded and the amd64 edition saw only 3.2 out of 6 GB of RAM. If they've gotten beyond this I'd like to give OpenBSD another shot but I dont have any systems that don't have multiple cores anymore and I don't see much sense in running an OS that can't use them fully.

[Alex Libman]

Replying to the above paragraph by paragraph:

I'm talking about the coming decade.  Google's Native Client will go a long way in bridging the in-browser performance gap.  Even the local operating systems will eventually be 99% managed code, so it doesn't really matter if that code is accessed from the hard disk or via HTTP (although the security policies would obviously differ).

There's nothing wrong with corporate sponsorship.  I'd be running Windows right now if it offered me more as a Web-centric developer.  It certainly does the best job in terms of Web browser selection and performance - I need to remote-access / emulate a Windows box anyway, because most clients would be rather offended if their sites didn't work right on IE.  New technologies like AJAX, HTML5, and HTTP(P2P) might actually be a Linux killer, because the one area where Linux dominates, servers, will diminish in importance, while the importance of hardware acceleration and compatibility will increase.

Yeah, OpenBSD performance still sucks, but if you have enough CPU power it's not so bad.  It is a cilice that I wear in the name of freedom. 

[Zhwazi]

What I would be hoping for instead of AJAX interfaces and such would be something more like a client-server model to most applications where zero latency between cause and effect isn't a priority, like transmission-daemon or quassel or mpd, where you could theoretically use any interface you like including web interfaces or native UI widgets to communicate with a server process that doesn't necessarily have to be on the same system. Web browsers might be a universal client with their own, standard, cross-platform, easy-to-learn language for building interfaces, but it seems silly to think that this will be the farthest extent of it, especially as new types of interfaces like touch, and voice control likely to become more mainstream as voice recognition software continues to improve. Your standard AJAX page in a web browser may become the least-common-denominator for development of UI's with native UI clients likely providing better usability.

[Alex Libman]

I definitely agree about app client-server model with multiple types of clients being able to communicate with the server API.  This API could be accessed by scripts, by command-line clients, and by AJAX interfaces (which could operate in several modes - optimized for phones, optimized for touchscreens, etc).  I just don't see much point in using any existing native widgets anymore (ex. Qt, GTK, wx, etc).  Maybe I'm biased because of the licensing, but it seems that anything native widgets can do, HTML5+ / AJAX will soon do just as well, providing an easily accessible and consistent interface across all platforms.  It's also a lot easier to get users to try an app by simply clicking a link rather than having to install something locally!

Of course there are other in-browser technologies in addition to HTML5, but if even Microsoft is downplaying Silverlight in favor of HTML5, I think JavaFX, AIR, and even the almighty Flash will someday follow.  Something like HTML6 (although they won't use this versioning scheme anymore) can combine HTML5 with what Google is doing with Native Client to provide pre-optimized binary sites, complete with standardized logical objects that are no less accessible than any native widget toolkit.

BTW, here's another point of view in an article once recommended to me by Sprewell (the aforementioned Hybrid Source / Chromium porting guy) - A New Thin Client.

[Zhwazi]

I expect competition from browser based UI's to push native-execution UIs to be better. They're different tools, AJAX UI's will be the general purpose ones and natively executing UIs will be for applications where sophistication and responsiveness matter. Neither one is going to completely supplant the other, AJAX UI's are just new so their growth hasn't approached their own limiting factors yet.

[TimeLady Victorious]

Stallman pisses me off.

But I do what I want cuz a pirate is free

[TimeLady Victorious]

Although seriously,  when I get the next computer I have - when I get around to _building_ a computer - it's probably gonna be dual-booting a flavor of BSD and a flavor of Linux.