On Fri, Apr 26, 2013 at 1:38 PM, Alan Mackenzie <a...@muc.de> wrote: > 'afternoon, Canek!
Hi Alan. > On Fri, Apr 26, 2013 at 12:02:38PM -0500, Canek Peláez Valdés wrote: >> On Fri, Apr 26, 2013 at 11:29 AM, Alan Mackenzie <a...@muc.de> wrote: >> [snip] >> > Somebody reported that pulseaudio is an absolute requirement for Gnome >> >>=3.8. That may not be 100% of users, but the "forced" is certainly >> > there. > >> No one is forcing nothing on anyone, since nobody is forcing no one to >> use GNOME, Gentoo, or Linux for that matter. > > That's a strawman argument. Anytime a free software project drops > support for something, it forces its users to make choices. Yes, force. I don't think that's true, since we are not paying anyone to do the work (well, at least for sure I'm not paying anyone to do anything). They (the developers) don't own us *anything*. >> The developers of any project can always decide the dependencies of a >> project. If you are not a developer, you simply have no vote in the >> matter, although you certainly always have voice... that they can >> choose to ignore. > > Free software developers, having got people to commit to using their > software, have responsibilities, albeit moral ones. If you want to get into morals, this will become a religious argument, and sorry but I'm not interested in that. > The prime one is to support their users. No; the prime one is to do their jobs. Most of them are employed by several of the available Open Source supporting companies; their responsibilities is to do the job they are being paid to do. If they are hobbyist, then their prime "responsibility" is to do whatever the hell they want to (and gets accepted in a community project). > You'll surely have noticed that what gets up the > noses of people on this mailing list most is when support for reasonable > configurations gets dropped. Witness all the recent trouble over eth0, > for example. What problem? I use NetworkManager in desktop and laptop; there is no problem there. I read the instructions in my media center and servers: no problems there. I don't particularly like the new funny names, but I don't write the code, and the fruits from it I get for free, so I don't complain about it. >> > There's a difference between a "default choice" and an absolute >> > requirement. > >> Yeah; and the decision is for the developers to make. > >> >> Basically there's a bunch of vague criticisms of unnamed systems where >> >> "they" force stuff on "all users" for "no good reason". Nevermind that >> >> we can actually state what the reasons are. Fingers in the ears. >> >> neener neener. > >> > Please feel free to state those reasons, which as far as I can see, >> > nobody has done yet in this thread; "they" being the gnome team, and the >> > reasons being for the forcing, not for a non-existent "default choice". > >> If GNOME has to support PA and non-pa systems, they need to code, >> test, support and bug-fix 2 different sets of of systems. If they need >> to support ConsoleKit and logind, the number grows to 4 (PA/ck, >> PA/logind, non-PA/ck, non-PA/logind). With 3 different optional >> requirements, it's 8 sets of systems. With 4, is 16. With n, it's 2^n. > >> That's exponential growth, which in CS is always no-no. > > WADR, that is simply false. With features which interact chaotically > with eachother, yes, you have exponential growth. With distinct, > self-contained features, each one is merely an incremental test effort. > ALSA and pulseaudio are self-contained, and are also well tested in their > own right. Only integration needs testing. OK, I exaggerated a bit; but who is going to do the integration testing? You? Because the GNOME developers have no interest in doing that, and I support their decision. > If you were serious about this exponential growth, how on earth could, > e.g., the Linux kernel or Emacs, both with thousands of options[*], > possibly get tested anywhere near acceptably? > > [*] 12,666 in Linux 3.7.10, 7,510 in vanilla Emacs 24.3. Because they have enough integration testers. They have enough interested users to do the required testing; the kernel and Emacs is oriented towards technical apt users. The stated goal of the GNOME project is that even my grandmother could use it. >> Who is going to code, test, support and bug fix all those possible >> configurations? You? > > No. The gnome developers. Why? Because you say so? Do you pay them? > I test and support all reasonable (and many > unreasonable) combinations on my own free software project. Good for you: that's your call. It's not your call to say what the GNOME developers should use. >> The GNOME developers simply cannot support all different sets of >> possible configurations, and PA covers the sound needs of *ALL* users >> (doesn't matter if you like it or not), even the simple cases. > > What about the needs of those high-end audio users, for example, who need > jack? There are several success stories about mixing PA with Jack; you can Google them. I don't see the problem. > What about those, like me, with audio problems, where the need exists to > strip a system down so as to isolate those problems? As I said below: if PA has problems, they need to be fixed. Did you report the bugs? >> If PA has bugs in some configuration, those bugs need to be fixed; the >> solution (in the GNOME developers view) is not to "remove PA", since >> the goal of the project is to cover *ALL* use cases. > > pulseaudio is a server component - gnome is an application. They are at > different levels of the system hierarchy, just as a mail transport agent > and mail user agent are. The maintainers of mutt don't force the use of, > say, postfix. By long tradition on *nix, sysadmins configure their own > systems, selecting those components which best fit their needs. gnome's > decision to mandate pulseaudio interferes with this tradition. IMAO, > this is a Bad Thing. GNOME is a desktop environment, and it wants (from some years now) a vertical integration from kernel to the last userspace application. I root for that. And I have been using Unix since 1996, and I don't care about what *nix "long traditions" are. I want a Linux system that works from my cellphone to my big iron server, and everything in between. I don't even care about *BSD; I don't wish them any ill, but I don't care about them. If you don't agree with that, that's fine; but if a big enough set of developers thinks similarly, several projects will move in that direction. It's already happening. >> But hey, the source is there; feel free to patch whatever needs to be >> patched in GNOME (and probably GStreamer) so it doesn't require PA. >> Just be certain that those patches will be rejected by upstream, for >> the reasons stated above. > > Making minor changes to free software is impracticable on a casual basis. > Only forking a project can do this. You know this full well. Well, fork away then. It's in your freedoms when it comes to free software. Look at the success stories from MATE, Cinnamon, the Trinity Desktop Environment and eudev. >> And by the way, this is also true for Gentoo: it cannot support all >> different sets of possible configurations, no matter how hard they/we >> try. > > It come pretty close. :-) http://article.gmane.org/gmane.linux.gentoo.devel/85223 "I'm not sure about the future of the core of OpenRC: Upstart & systemd have some clear architectural benefits, despite their implementation shortcomings (either upstream or per-distro). The /usr merge is inevitable, as is the integration of other components into the init system (udev, dbus, ...). What has become dis-integrated instead is the configuration: lots of hardware ships specific udev rules with few problems." That's a Gentoo developer talking. Many of them think like that; many of them oppose them. But the truth is, Linux distributions are moving to a vertical, tightly integrated OS. There will always be niche distros for the people that don't like this (and hey, that's the beauty of using Free Software), but the big distros are going that way. It's possible that Gentoo will follow (as did Arch, as Debian-based Tanglu and Gentoo-based Sabayon are doing). Greg Korah-Hartman will integrate an incarnation of dbus (or something really similar) into the kernel: https://plus.google.com/u/0/111049168280159033135/posts/BDgg5DKVjkX and then the only required dependency for systemd will be Linux. Man, that will be a nice day. Regards. -- Canek Peláez Valdés Posgrado en Ciencia e Ingeniería de la Computación Universidad Nacional Autónoma de México
