Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
When (and if) Gentoo will switch to systemd?
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 16, 17, 18  
Reply to topic    Gentoo Forums Forum Index Other Things Gentoo
View previous topic :: View next topic  
Author Message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1483

PostPosted: Tue Apr 08, 2014 11:43 pm    Post subject: Reply with quote

FWIW I can't wait for them to fork the kernel.

That way, what they have is no longer Linux and we won't have to worry about it much anymore.
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Wed Apr 09, 2014 1:05 pm    Post subject: Reply with quote

Dominique_71 wrote:
RH is obviously not interested by that compatibility, and if they fork the kernel, we can be sure the first thing they will do is to remove most, if not all, of the non x86 and non amd64 stuffs from their fork.

That's not very likely at all IMO. They make good money off the work of the community, and ofc contribute code back, always have.

Also, practically every commercial distro wants to do ARM, nowadays, which is very different to a decade or so ago. Lots of money in mobile, so they won't want to lose that. I hear what you're saying about the less popular archs, but there's no real point in killing the goose that's paid their way for so long.
Quote:
It may split the community, but it can be a good thing on the long run, because today it is already spitted, and some contributors to GNU/Linux are already switching to other OS like FreeBSD or even OSX, and will instead be more receptive to stay with GNU/Linux.

That's a total non-sequitur imo. It's not already split: it's diverse, which is a great strength. As for people switching, who cares? That's the whole point of Free software: use what you want.

1clue wrote:
FWIW I can't wait for them to fork the kernel.

That way, what they have is no longer Linux and we won't have to worry about it much anymore.

Just not going to happen: they already patch it to death, and effectively have their own variant. And they don't sell Linux: they sell Red Hat Enterprise Linux.
Branding is essential, and sure they have an interest in being presented as The Linux, but that requires everyone else to be using Linux too. They also have much more at stake in terms of actually getting the work done, which requires the rest of us, as well as willing beta-testers who'll do the QA and the real work of bugfixing. But they're happy to get the kool-aid "for free" so the cash cow keeps giving.

"We won't have to worry about it anymore" is naive, afaic. I can understand the sentiment, but pat, easy solutions never come about, so why waste your time hearkening after them? About as useful as "Can't we all just get along?"
Back to top
View user's profile Send private message
Dominique_71
Veteran
Veteran


Joined: 17 Aug 2005
Posts: 1509
Location: Switzerland (Romandie)

PostPosted: Wed Apr 09, 2014 8:17 pm    Post subject: Reply with quote

steveL wrote:
That's a total non-sequitur imo. It's not already split: it's diverse, which is a great strength. As for people switching, who cares? That's the whole point of Free software: use what you want.

Or what you need. I try FreeBSD it was a few months ago, and way to much programs I need are missing, or completely outdated, to be a viable alternative for me. Also, from a technical point of view and user experience, I really prefer AROS over anything else. It is very simple (in the sens it doesn't have duplicated and concurrent efforts to do the same thing, which doesn't result into what GNU/Linux is in many points of view: a kitchen sink) and have a much faster GUI. But it have much more missing software than FreeBSD for my needs, so that's not a viable option for me.

Quote:
Just not going to happen: they already patch it to death, and effectively have their own variant. And they don't sell Linux: they sell Red Hat Enterprise Linux.
Branding is essential, and sure they have an interest in being presented as The Linux, but that requires everyone else to be using Linux too. They also have much more at stake in terms of actually getting the work done, which requires the rest of us, as well as willing beta-testers who'll do the QA and the real work of bugfixing. But they're happy to get the kool-aid "for free" so the cash cow keeps giving.

"We won't have to worry about it anymore" is naive, afaic. I can understand the sentiment, but pat, easy solutions never come about, so why waste your time hearkening after them? About as useful as "Can't we all just get along?"


You may be right on this. We will see.
_________________
[[[ To any NSA and FBI agents reading that text: please consider ]]]
[[[ whether defending the US Constitution against all enemies, ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1483

PostPosted: Wed Apr 09, 2014 9:17 pm    Post subject: Reply with quote

steveL wrote:
Dominique_71 wrote:
It may split the community, but it can be a good thing on the long run, because today it is already spitted, and some contributors to GNU/Linux are already switching to other OS like FreeBSD or even OSX, and will instead be more receptive to stay with GNU/Linux.

That's a total non-sequitur imo. It's not already split: it's diverse, which is a great strength. As for people switching, who cares? That's the whole point of Free software: use what you want.


@SteveL, I completely agree with that. I have a hard time seeing Open Source (the whole thing collectively) as a competition. Individual components within it compete (bsd/linux, or init systems, or web servers) and also these components compete with commercial offerings, but as a whole it's a big collaboration. If one were to insist that everyone pick an OS and stick with ONLY that OS then it would be different, but I pick whatever makes most sense for the specific application. But I can load apache on pretty much everything. I can load open-ssh on pretty much anything. Libre works on pretty much anything. What operating system you pick makes a difference certainly, but it's not a marriage: You can have more than one operating system, and you can work happily with all of them.

@Dominique_71, There are lots of awesome operating systems out there. Each has its strengths, and if you use them for that and write software that interconnects then everyone can be happy. Much of Open Source works on just about anything that people actually use, so how could the multiplicity of viable choices be a weakness for the end user?

steveL wrote:
1clue wrote:
FWIW I can't wait for them to fork the kernel.

That way, what they have is no longer Linux and we won't have to worry about it much anymore.

Just not going to happen: they already patch it to death, and effectively have their own variant. And they don't sell Linux: they sell Red Hat Enterprise Linux.
Branding is essential, and sure they have an interest in being presented as The Linux, but that requires everyone else to be using Linux too. They also have much more at stake in terms of actually getting the work done, which requires the rest of us, as well as willing beta-testers who'll do the QA and the real work of bugfixing. But they're happy to get the kool-aid "for free" so the cash cow keeps giving.

"We won't have to worry about it anymore" is naive, afaic. I can understand the sentiment, but pat, easy solutions never come about, so why waste your time hearkening after them? About as useful as "Can't we all just get along?"


I'm afraid you're right about that, but we can still wish for the simple solution. Them forking and calling it something else would effectively remove most of what I don't like about Linux today, and keep most of what I like. It would give them control of their kernel, and it would give me a more traditional setup that makes sense to me.
Back to top
View user's profile Send private message
Dominique_71
Veteran
Veteran


Joined: 17 Aug 2005
Posts: 1509
Location: Switzerland (Romandie)

PostPosted: Wed Apr 09, 2014 10:56 pm    Post subject: Reply with quote

1clue wrote:
@Dominique_71, There are lots of awesome operating systems out there. Each has its strengths, and if you use them for that and write software that interconnects then everyone can be happy. Much of Open Source works on just about anything that people actually use, so how could the multiplicity of viable choices be a weakness for the end user?


Take audio in GNU/Linux, it have always been a mess, and today we have different sound servers, and the only one that can interconnect with anything, pulseaudio, is not professional. That situation is due primarily to a lack of communication with the pro-audio community, which have plenty of very skilled and talented programmers. I am sure with more collaboration, it should have been possible to build on the existing pieces (ALSA + JACK) in order to get something like pulseaudio, but with true professional characteristics, and that with less efforts than what was put to make pulseaudio, and with the added goodies to transform every single audio software into a professional application.

Take also the different toolkits. They all reinvent the wheel, and the only difference for the user is that it is difficult if not impossible to get an uniform look and feel, it imply a terrible waste of resources if you run at the same times applications using different toolkits, and a developer that want to hack applications using different toolkits must waste its precious time learning different toolkits. My reference in that matter is the Amiga OS, which provide an unique toolkit that provide every thing from primitives as in X to complex functions as in Cairo, GTK+ or QT, and can be used by any language. Simple and efficient for both the users and the programmers.
_________________
[[[ To any NSA and FBI agents reading that text: please consider ]]]
[[[ whether defending the US Constitution against all enemies, ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1483

PostPosted: Wed Apr 09, 2014 11:39 pm    Post subject: Reply with quote

Dominique_71 wrote:
1clue wrote:
@Dominique_71, There are lots of awesome operating systems out there. Each has its strengths, and if you use them for that and write software that interconnects then everyone can be happy. Much of Open Source works on just about anything that people actually use, so how could the multiplicity of viable choices be a weakness for the end user?


Take audio in GNU/Linux, it have always been a mess, and today we have different sound servers, and the only one that can interconnect with anything, pulseaudio, is not professional. That situation is due primarily to a lack of communication with the pro-audio community, which have plenty of very skilled and talented programmers. I am sure with more collaboration, it should have been possible to build on the existing pieces (ALSA + JACK) in order to get something like pulseaudio, but with true professional characteristics, and that with less efforts than what was put to make pulseaudio, and with the added goodies to transform every single audio software into a professional application.

Take also the different toolkits. They all reinvent the wheel, and the only difference for the user is that it is difficult if not impossible to get an uniform look and feel, it imply a terrible waste of resources if you run at the same times applications using different toolkits, and a developer that want to hack applications using different toolkits must waste its precious time learning different toolkits. My reference in that matter is the Amiga OS, which provide an unique toolkit that provide every thing from primitives as in X to complex functions as in Cairo, GTK+ or QT, and can be used by any language. Simple and efficient for both the users and the programmers.


IMO you've hit on two of the biggest issues with Linux.

And while I would love for an audio tool suite which could be professional, I'd be even happier if it 'just worked' and had a basic interface. Logic something like this for initial configuration:

  1. Is there an add-on card or USB card? Then set that as default. Otherwise use the on-board card. Or maybe mix them all together and put all output everywhere.
  2. Set up for headset/microphone use with default settings that people actually use.
  3. Have a simple control panel with an 'advanced' button that only 10% of users need to click on.
  4. Have a variety of mixer plugins aimed at different audiences, and allow any or even multiples to occupy the 'advanced' button.


      With regards to toolkits, I have less to say about that. IMO anything that's lightweight but allows use of acceleration is fine by me, the only thing I'd like to improve is the sort of reliability and consistency required by companies like Adobe to entice them to write for Linux. IMO every GUI app should support basic X and optionally one or more toolkits. But that's pretty much what we have, other than the issues that keep graphics-intensive software from landing on Linux.
Back to top
View user's profile Send private message
miroR
Guru
Guru


Joined: 05 Mar 2008
Posts: 482

PostPosted: Sat Aug 02, 2014 1:15 am    Post subject: Reply with quote

I studied this entire topic, and Gentoo is likely really the last defence against this defeat that happened to GNU/Linux.

Since I try and maiintain a few Tips there with lesser resources of my time (the most of my efforts are with compiling my Gentoo), I hoped to get some opposition of reason there.

Hava a look:

Defeat and Hope for GNU/Linux
http://forums.debian.net/viewtopic.php?f=19&t=116472#p548548

This topic which I'm writing this post on features prominently there.

It's funny how I was thrown Offtopic so promptly, and got pronto a couple of nit pickers on me.

Compare that to the way TomWij in this topic deploys his arguments. How classy in comparison! While I don't agree with TomWij in this matter (pls. see his posts in this topic), I have to commend both the tolerance and the elegance of his writing.

Pls., let's fight this back! This tectonic-proportions defeat in GNU/Linux that systemD and all the poetteringware has brought on the Free Open Source Software.

Still reeling.

Miroslav Rovis
Zagreb, Croatia
www.CroatiaFidelsi.hr
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Sat Aug 02, 2014 1:13 pm    Post subject: Reply with quote

miroR wrote:
I studied this entire topic, and Gentoo is likely really the last defence against this defeat that happened to GNU/Linux.

Well one of them; I'm sure there are people working on at least one or two distros without systemd, at least as an option. Even on debian, you can use systemd-must-die (search that name on the debian dev mailing-list) from mirabilos, who's one of the leads on mirBSD and supports mksh which he maintains upstream, on debian.
Quote:
Since I try and maiintain a few Tips there with lesser resources of my time (the most of my efforts are with compiling my Gentoo), I hoped to get some opposition of reason there.

Hava a look:

Defeat and Hope for GNU/Linux
http://forums.debian.net/viewtopic.php?f=19&t=116472#p548548

This topic which I'm writing this post on features prominently there.

It's funny how I was thrown Offtopic so promptly, and got pronto a couple of nit pickers on me.

Well I thought it might have some relevance, but the only interesting thing I found was this (gentoo forums) post on libav which turned out to be a dud too.

You really should learn to edit your original post instead of adding a "new version" and linking to it from the OP; quite apart from you need to learn how to configure gentoo a bit more, before you start declaiming how it's all broken/that something is being "imposed" when it isn't.
It makes you look very foolish, imo, which is a shame since you're clearly intelligent and motivated to make things work better. You would do better to ask first whether your options are being restricted, or rather "how do i carry on using X?"

edit: correct link to libav post.


Last edited by steveL on Sat Aug 02, 2014 5:52 pm; edited 1 time in total
Back to top
View user's profile Send private message
khayyam
Advocate
Advocate


Joined: 07 Jun 2012
Posts: 2944

PostPosted: Sat Aug 02, 2014 1:49 pm    Post subject: Reply with quote

steveL ... you're wasting your time and inviting a deluge of noise.

best ... khay
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Sat Aug 02, 2014 5:45 pm    Post subject: Reply with quote

khay: ouch.

miroR: take the hint and stop: a) putting out long rambling posts, and b) quoting large chunks of people's posts, especially your own. If you don't like that, start another topic somewhere else to complain about it, DO NOT ramble in this one.
You really MUST take this advice or very soon no-one will want to deal with you, and rightly so.

No, we are not interested in discussing it. Just change your god-awful "style".
Back to top
View user's profile Send private message
miroR
Guru
Guru


Joined: 05 Mar 2008
Posts: 482

PostPosted: Sat Aug 02, 2014 10:02 pm    Post subject: Reply with quote

steveL wrote:
miroR wrote:
I studied this entire topic, and Gentoo is likely really the last defence against this defeat that happened to GNU/Linux.

Well one of them; I'm sure there are people working on at least one or two distros without systemd, at least as an option. Even on debian, you can use systemd-must-die (search that name on the debian dev mailing-list) from mirabilos, who's one of the leads on mirBSD and supports mksh which he maintains upstream, on debian.
...[snip]...

I really, first, thank you for that piece of advice.
That is what I asked for on Debian for, what, two days now.
I will review the rest of what you are saying.
The favor hugely outweighs the disagreeable feeling for your criticism of me, and I'll review and if I need, will admit where I went wrong, or not (as I will manage to understand, either way that it be, I'm sincere).
So, thank you again, steveL!
Miroslav Rovis
Zagreb, Croatia
www.CroatiaFidelis.hr
P.S. I'm still pondering what I said I would, but I do think it's worth considering this my new post Off the Wall:
Place for Criticism Unhelpful/Damaging to Topics Here, pls.!
https://forums.gentoo.org/viewtopic-p-7593384.html#7593384
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Sun Aug 03, 2014 11:42 am    Post subject: Reply with quote

miroR wrote:
I really, first, thank you for that piece of advice.
That is what I asked for on Debian for, what, two days now.
I will review the rest of what you are saying.
The favor hugely outweighs the disagreeable feeling for your criticism of me, and I'll review and if I need, will admit where I went wrong, or not (as I will manage to understand, either way that it be, I'm sincere).
So, thank you again, steveL!

You're very welcome.

Note however that it is not criticism of you, but of your posts: their rambling nature and their length. You really must learn to detach your ego from your output: that's essential to coding as much as it is to free-speech. I haven't said a single thing about you, apart from that you are clearly intelligent and want to help, which is why I'm spending time helping you to correct your netiquette at this stage.
Quote:
P.S. I'm still pondering what I said I would, but I do think it's worth considering this my new post Off the Wall:
Place for Criticism Unhelpful/Damaging to Topics Here, pls.!

That is such a bad move. If you have to direct what you call "unhelpful" criticism to a separate thread all of your own, it indicates you're unwilling to simply take criticism, as well as denigrating people who are trying to help you improve your output. Both a quick way to get yourself a reputation as an idiot, in whatever language you wish to dress that term up. Search "flipping the bozo bit" if you want an explanation of that process. I realise I suggested it: I didn't realise you'd take it seriously. My bad.

If I had to diagnose, I'd say it appears you're treating forums as a combination of blog, email and newsgroup, wherein you both announce/lay out your ideas and chat (and chat..) around the subject in the same post. They're none of those things, especially when it comes to a computing forum; and Gentoo users are basically some of the most knowledgeable around, and certainly have a much higher ratio of what would be called power-users in any other system. Every Gentoo user on these forums is an admin of their machine if not a network, so they tackle compilation and build problems daily; the vast majority of them work in IT, and they are typically on the older end of the bell-curve.

Know your crowd. Keep it terse in preference to chatty, especially when you are purporting to document something.

And FGS stop doing N drafts of a long document in the same topic. Just edit the OP, or I for one will simply ignore any posts authored by you if it does not stop. Sorry, but life's too short (and I'm too old) to wade thru 3 pages of noise to extract one or two paragraphs of signal.
Back to top
View user's profile Send private message
miroR
Guru
Guru


Joined: 05 Mar 2008
Posts: 482

PostPosted: Sun Aug 03, 2014 4:49 pm    Post subject: Reply with quote

steveL wrote:
miroR wrote:
I really, first, thank you for that piece of advice.
That is what I asked for on Debian for, what, two days now.
I will review the rest of what you are saying.
The favor hugely outweighs the disagreeable feeling for your criticism of me, and I'll review and if I need, will admit where I went wrong, or not (as I will manage to understand, either way that it be, I'm sincere).
So, thank you again, steveL!

You're very welcome.

Note however that it is not criticism of you, but of your posts: their rambling nature and their length. You really must learn to detach your ego from your output: that's essential to coding as much as it is to free-speech. I haven't said a single thing about you, apart from that you are clearly intelligent and want to help, which is why I'm spending time helping you to correct your netiquette at this stage.
Quote:
P.S. I'm still pondering what I said I would, but I do think it's worth considering this my new post Off the Wall:
Place for Criticism Unhelpful/Damaging to Topics Here, pls.!

That is such a bad move. If you have to direct what you call "unhelpful" criticism to a separate thread all of your own, it indicates you're unwilling to simply take criticism, as well as denigrating people who are trying to help you improve your output. Both a quick way to get yourself a reputation as an idiot, in whatever language you wish to dress that term up. Search "flipping the bozo bit" if you want an explanation of that process. I realise I suggested it: I didn't realise you'd take it seriously. My bad.
...[snip]...

No, it's not.
This is simply not good for this systemD topic. It has nothing whatsoever to do with...

systemD

My suggestion, and for this other topic here, that one by me as OP.:
Likely: musl based install on my old systems?
https://forums.gentoo.org/viewtopic-p-7593736.html#7593736
I'll reply to you in a while on that "Criticism... Here" page.
======================
EDIT: replied:
https://forums.gentoo.org/viewtopic-p-7593798.html#7593798
======================
Would have better liked to work on musl install on my old boxes, but this is now necessary.
Thank you!
Miroslav Rovis
Zagreb, Croatia
www.CroatiaFidelis.hr
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Mon Aug 04, 2014 2:33 pm    Post subject: Reply with quote

miroR wrote:
steveL wrote:
Quote:
P.S. I'm still pondering what I said I would, but I do think it's worth considering this my new post Off the Wall:
Place for Criticism Unhelpful/Damaging to Topics Here, pls.!

That is such a bad move. If you have to direct what you call "unhelpful" criticism to a separate thread all of your own, it indicates you're unwilling to simply take criticism, as well as denigrating people who are trying to help you improve your output. Both a quick way to get yourself a reputation as an idiot, in whatever language you wish to dress that term up. Search "flipping the bozo bit" if you want an explanation of that process. I realise I suggested it: I didn't realise you'd take it seriously. My bad.
...[snip]...

No, it's not.
This is simply not good for this systemD topic. It has nothing whatsoever to do with...

systemD

What you're missing is that forums are similar to email lists in that the discussion takes place where it takes place. If you get criticism for something you've posted, first you need to remember it's not criticism of you, but your words (the free speech aspect you skipped by above), and second you have to accept it's going to happen where you make the post.

You don't seem to have taken in the point about denigrating people, so I'll spell it out: you're asking them to accept your labelling of their opinion as off-topic and unhelpful. As explained that simply highlights your ignorance of the cultural norms outlined above, which makes the insult doubly so.
Quote:
I'll reply to you in a while on that "Criticism... Here" page.

I won't be responding to that again. As I said in my reply there, it's completely the wrong forum. zixnub's response, amongst others, evinces the flipping of the idiot bit previously discussed, which you apparently have no time to find out about, but do have time for a long essay I cba to read. tl;dr as you were told you'd get.
Quote:
Would have better liked to work on musl install on my old boxes, but this is now necessary.

Only because you keep responding with long essays, and far too long quote boxes (as fixed for you above), instead of taking a step back, and thinking about what's being said, preferably as you take a day or two off from hitting reply/quote.

In any event, EOD from me: feel free to carry on in OTW; I won't be responding there or here.

Again: take the time to learn the cultural norms, which includes taking time out from responding all the time, and letting your hindbrain catch up/process everything properly: this means doing something besides sitting at your computer. Or ignore me, and continue to come across as a mental defective. (If that sounds harsh, read the OTW discussion to see what I mean.)
Back to top
View user's profile Send private message
dkasak
n00b
n00b


Joined: 10 Feb 2012
Posts: 11

PostPosted: Mon Sep 01, 2014 12:43 pm    Post subject: Reply with quote

Dominique_71 wrote:
Take audio in GNU/Linux, it have always been a mess, and today we have different sound servers, and the only one that can interconnect with anything, pulseaudio, is not professional. That situation is due primarily to a lack of communication with the pro-audio community, which have plenty of very skilled and talented programmers. I am sure with more collaboration, it should have been possible to build on the existing pieces (ALSA + JACK) in order to get something like pulseaudio, but with true professional characteristics, and that with less efforts than what was put to make pulseaudio, and with the added goodies to transform every single audio software into a professional application.

Take also the different toolkits. They all reinvent the wheel, and the only difference for the user is that it is difficult if not impossible to get an uniform look and feel, it imply a terrible waste of resources if you run at the same times applications using different toolkits, and a developer that want to hack applications using different toolkits must waste its precious time learning different toolkits. My reference in that matter is the Amiga OS, which provide an unique toolkit that provide every thing from primitives as in X to complex functions as in Cairo, GTK+ or QT, and can be used by any language. Simple and efficient for both the users and the programmers.


I have to disagree with both of these examples, and the general sentiment.

I'm a user of Jack - for audio recording, production, and live work. I'm a user of pulseaudio - for simple network-transparent audio. While it's easy to claim it both of these projects should merge, there are clear reasons why this will not, and should not, happen.

1) Latency vs Compromises. A user of Jack wants low latency and zero buffer overruns AT ALL COSTS. A user of pulse wants reasonable performance, doesn't care about extremely low latency, and doesn't want their laptop or mobile device constantly allocating CPU cycles to an audio layer. Being a user of both of these systems, and being a developer, I can't see how these competing goals can be achieved simultaneously, though I admit I've not personally hacked on either. These competing goals imply the need for 2 solutions, both fit for a purpose.

2) Network transparency vs Compromises. A user of Pulse finds network transparency a great feature. I certainly do. It means I can have my home server running pulse, sitting next to my nice big speakers, and then play audio from other devices to the same big speakers without having to muck around with wires. That's a great home-use feature. That makes no sense for users of Jack, and adds complexity - bloat - to a codebase that should be concerned with remaining lightweight and low latency. The same kind of criticism of X's built-in network transparency is often heard from Linux desktop users who feel that the network protocol is increasing latency of their desktop effects. In this case, I also see both sides. As a user who admits to being fascinated by a blingy desktop ( I run Enlightenment-0.19 ), I want a low latency display. But again as a developer with a home server, multiple other desktops, remote systems to manage, I can see the value in network transparency. What's the answer? Kill X? Screw people who want a low latency desktop? Of course, we can have both. X can run in Wayland, and Wayland can run in X.

As a desktop user, I'm well aware of the fact that there are multiple GUI toolkits around. As a developer, I've coded extensively with Gtk2 and Gtk3. I've investigated QT at various points, and never liked it. I've investigated EFL and the now defunct ETK.

In the case of both GUI Toolkits and audio layers, there are very different communities developing code. The declaration that everyone should just get together and do the same thing - in the open source world no less - is just naive. Different people have different ideas. They gravitate to different projects. There are different licenses involved. There are different languages involved.

Choice is a good thing. 'Fragmentation' is the war cry of the Apple fanboi. We, the open-source community, should be embracing diversity and choice, not lecturing people on trying to develop to some lowest common denominator, in the name of having no choice but a nice dumbed-down homogeneous experience.
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3314
Location: The Peanut Gallery

PostPosted: Mon Sep 01, 2014 6:28 pm    Post subject: Reply with quote

dkasak wrote:
Dominique_71 wrote:
Take audio in GNU/Linux, it have always been a mess, and today we have different sound servers, and the only one that can interconnect with anything, pulseaudio, is not professional. That situation is due primarily to a lack of communication with the pro-audio community, which have plenty of very skilled and talented programmers. I am sure with more collaboration, it should have been possible to build on the existing pieces (ALSA + JACK) in order to get something like pulseaudio, but with true professional characteristics, and that with less efforts than what was put to make pulseaudio, and with the added goodies to transform every single audio software into a professional application.

I have to disagree with [both of] these examples, and the general sentiment.

I'm a user of Jack - for audio recording, production, and live work. I'm a user of pulseaudio - for simple network-transparent audio. While it's easy to claim it both of these projects should merge, there are clear reasons why this will not, and should not, happen.

I see what you're getting at, but you've misinterpreted "build on the existing pieces" to mean "merge", which substantially weakens your point.
Quote:
1) Latency vs Compromises. A user of Jack wants low latency and zero buffer overruns AT ALL COSTS. A user of pulse wants reasonable performance, doesn't care about extremely low latency, and doesn't want their laptop or mobile device constantly allocating CPU cycles to an audio layer. Being a user of both of these systems, and being a developer, I can't see how these competing goals can be achieved simultaneously, though I admit I've not personally hacked on either. These competing goals imply the need for 2 solutions, both fit for a purpose.

Hmm to me it's always implied the need for one layer to build on top of the other. You say a home-user "doesn't want the device constantly allocatiing CPU cycles," but quite apart from how the distro is configured, I'd say a home-user doesn't want their audio dropping out, first and foremost. That is after all the reason why video streams prioritise audio; since dropped frames are much less of a problem than dropped samples.

My point being that first off we want correct performance, only secondly do we want convenience, however we slice it, since convenience can always be built on top of a correctly functioning layer, but the other way round is a dead-end (the best you get is a "rewrite" to deliver the lower-layer after the event.)
Quote:
2) Network transparency vs Compromises. A user of Pulse finds network transparency a great feature. I certainly do. It means I can have my home server running pulse, sitting next to my nice big speakers, and then play audio from other devices to the same big speakers without having to muck around with wires. That's a great home-use feature. That makes no sense for users of Jack, and adds complexity - bloat - to a codebase that should be concerned with remaining lightweight and low latency.

It makes a lot of sense for some users of jack, and as you point out for monitoring. Though again, there's no reason why a network streamer can't use jack on the output device, and ofc it would be preferable on any capture side too, if there were one.

To see what I mean at code level, s/jack/sockets/ and s/network transparency/TCP/. It's not an either/or choice, which I think is what Dominique was getting at: a better way to implement PA "essentials" (the one that comes up time and again is in fact bluetooth headphones) is to build on top of jack as the standard audio layer. If we'd settled on that as a community instead of spending five years correcting PA, then we'd be in a much better position all-round.
Quote:
In the case of both GUI Toolkits and audio layers, there are very different communities developing code. The declaration that everyone should just get together and do the same thing - in the open source world no less - is just naive. Different people have different ideas. They gravitate to different projects. There are different licenses involved. There are different languages involved.

Sure, which is why diversity is a good thing: it means very different ways of thinking and expression can collaborate. That's why it was possible to develop PA, and systemd, using existing systems.
Quote:
Choice is a good thing. 'Fragmentation' is the war cry of the Apple fanboi. We, the open-source community, should be embracing diversity and choice, not lecturing people on trying to develop to some lowest common denominator, in the name of having no choice but a nice dumbed-down homogeneous experience.

Agreed, though that criticism is much more applicable to pulseaudio and the rest of the systemdbug-nubkit crapfest. Exactly that "reasoning" is being pushed as justification for aping Windoze, which seems to be more influential on the nu-skool than they realise. And that's why it's dangerous for the wider health of the software, and the community.
Back to top
View user's profile Send private message
F1r31c3r
Tux's lil' helper
Tux's lil' helper


Joined: 31 Aug 2007
Posts: 92
Location: UK

PostPosted: Fri Dec 12, 2014 3:54 am    Post subject: Re: When (and if) Gentoo will switch to systemd? Reply with quote

Shum wrote:
Seeing as systemd seems to be taking the linux world by storm and a lot of the other distros are adopting it, I'm wondering whether the Gentoo devs have any plans to make it the default init system and replacement for OpenRC?


When? Why? If?

If you want SystemD then use it, if you want OpenRC then use that. It is not brain science this is Gentoo you choose everything the way you want it and i do mean you choose everything absolutely.

I came over to this thread via the mention on youtube. For those that look at Gentoo and think "oh man they are just stubborn and won't use SytemD" NO just no.

Gentoo allows you to choose, Gentoo does use SystemD, it also uses OpenRC it can also use other init systems as well. You are not limited and never ever tied down to anything. The only ties you have sometimes is when using binary blobs such as 'VMWare' these binaries are compiled with specific libraries and versions against a particular kernel. VMWare rarely update the linux software that is available for download so it takes forever for the binary to be re-compiled to newer versions of the libraries they use.

The GNU/Linux software moves so fast in developments that when a binary package is made it takes only a few weeks to months for libraries to change therefore breaking the binary package. Gentoo is all source which means if a library changes that a program relies on then the program that relies on it is picked up and rebuilt against the new libraries making every element of the Gentoo system the most robust and stable.
Other distros are all pre-compiled and although they do a good job to make sure they meet all the dependencies to the binary version of a program when they compile it for you occasionally they fail. This puts the user in a situation where he has to battle the package manager to fix it, provided he even knows how.

Cut out the binary package manager and life just becomes so much simpler hence Gentoo is what it is.

SystemD for me causes issues once it has been compiled. With OpenRC i have 100% control over all the init boot stage of my system. SystemD although has a level of control it takes away some control, tells me what i can or can not run. Sometimes for good reason as so not to break the system but if I wanted to do something that might break the system that is my choice and not for SystemD to stop me.

SystemD can be very good, it can be much more than what it is but if they don't remove the restrictions type mentality in the software then it is junk in my opinion of course, i really dont care for 10 seconds faster boot time. Not everyone has this opinion so as it is Gentoo you make your own choice which you wish to use and whenever you wish to use it. Nothing stopping you using SystemD and if you don't like it moving back to OpenRC. Gentoo lets you do all that very simply.
Use flags and the options within the kernel "make menuconfig" make it very simple to switch. Make scripts were the founding installation managers of the entire GNU/Linux system from the beginning, python built emerge is just an extension ontop of that make system making it more human friendly. The make system has been intimidating to new users for a long time there is no doubt of that and emerge does a very good job of simplifying the whole process.

So hopefully this puts an end to the argument of what and why on Gentoo regarding SystemD.

Just remember Distributions are built with that particular distribution'rs (<- i know that is bad English 'rs or whatever but who cares) image of what they want the 'out of the box' experience to be like.
Gentoo is an out of the box experience set to 'the sky is the limit', what do you want it to be like? The installation is asking you the questions making it into something only you can answer.

So the when?

A:You choose it so this is a stupid question.

Why?

A:Because it is your choice alone to make and not that of a developer or some group or community imposing their thoughts on you.

If?

A: Again another stupid question as you choose it so there is no IF. Maybe this answer is answered by, Gentoo portage tree has active OpenRC and SystemD ready to be built at your disposal whatever you choose. It has always been there you just did not understand how the Gentoo model works.

Note this post is not intended to insult anyone. I just wanted to hit the questions hard with the facts so people can learn exactly what it is about. Apologies in advance if anyone became offended by anything i wrote.
_________________
When i look In-between white and black i see a rainbow of colours
Back to top
View user's profile Send private message
khayyam
Advocate
Advocate


Joined: 07 Jun 2012
Posts: 2944

PostPosted: Fri Dec 12, 2014 10:31 am    Post subject: Re: When (and if) Gentoo will switch to systemd? Reply with quote

F1r31c3r ... some friendly comments:

F1r31c3r wrote:
If you want SystemD then use it, if you want OpenRC then use that. It is not brain science this is Gentoo you choose everything the way you want it and i do mean you choose everything absolutely.

That's not entirely true, there are situations in which dependencies are non-optional yet not needed for the actual working of the applications requiring them. One case in point: net-analyzer/wireshark ... this has a hard dependency on x11-misc/xdg-utils but xdg-utils is not required for the application to function. Its a dependency so that when "help" is selected the "browser of choice" is opened with the online help. However, this can be configured via the preferences, and given this is a professional grade network analysis tool its somewhat disconcerting that the user is forced to have such a framework (and all of it its dependencies) simply so that a preference is auto-selected. So as to be able to "choose" not to have xdg-utils I've had to maintain the package in my own overlay as the bug I'd opened on this was closed WONTFIX.

So, we do not get to "choose everything absolutely", and as systemd has the stated objective of a "standard system base" and to "gently push" (sic) this "standard" there is a deeper question of who exactly is making the choices (particularly as "users" are seen more and more a recipients of the goodness bestowed upon us, and not the very community that has made, and makes, the proposition of a "free" OS possible).

F1r31c3r wrote:
The GNU/Linux software moves so fast in developments that when a binary package is made it takes only a few weeks to months for libraries to change therefore breaking the binary package. Gentoo is all source which means if a library changes that a program relies on then the program that relies on it is picked up and rebuilt against the new libraries making every element of the Gentoo system the most robust and stable.

Again, not entirely true ... gentoo isn't isolated from the wider community, nor is it free from the effects of entropy/chaos, there is always the danger that too much complexity will create an unstable system (in the entropic sense). If the wider community makes bad decisions these decisions will necessarily have an effect on gentoo ... particularly if the gentoo community doesn't thoroughly consider the repercussions, and nature of, their adoption (something which is writ large in the case of systemd).

best ... khay
Back to top
View user's profile Send private message
F1r31c3r
Tux's lil' helper
Tux's lil' helper


Joined: 31 Aug 2007
Posts: 92
Location: UK

PostPosted: Fri Dec 12, 2014 2:54 pm    Post subject: Re: When (and if) Gentoo will switch to systemd? Reply with quote

khayyam wrote:


Yes some very valid points which are much deeper and more complex workings into the not so perfect world of software.

Standardization vs the free operating system. It is true but their is also a compromise only it would seem we have not figured out how to balance it out between the two as of yet.

The moral of the story is that Gentoo allows anyone, user, developer, system admin etc to choose which init system they want to use. There is no "When will Gentoo adopt something" as such, it is more like if a Gentoo user wishes to use a certain software package or not, does that person make an ebuild of it and so on. If that person does not know how to make an ebuild then he can ask someone who does. Therefore it is not up to Gentoo maintainers to decide should they use SystemD or OpenRC, it is fair to say that it is up to them to make sure the person using Gentoo has the choice. Even if they disagree and wont put it in the standard branch then one can make an overlay to publish their source code.

the "choose everything absolutely" may have been a bit too strong wording.

For over 20 years i have never had any problems with programs that have the source code available, all be it they don't always have as much functionality and i need to contribute code to the project as to make it do what i want it to do, that can take allot of effort at times but hey it's all good fun. The programs that cause me the most trouble and always have done have been proprietary software and packages that allow binary blobs, the LGPL(SystemD is licensed under the LGPL), AMD-ATI, VMWare and others.

For a programmer SystemD can be modified, binary blobs removed as so not to run etc but for the standard user who can not read or write code SystemD could become the software that will abuse them. Not saying it will but it could.
I wonder how long it will be before those that use SystemD will see EULA agreements needing acceptance before they can use certain features?
_________________
When i look In-between white and black i see a rainbow of colours
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Other Things Gentoo All times are GMT
Goto page Previous  1, 2, 3 ... 16, 17, 18
Page 18 of 18

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum