Linked by Anonymous on Tue 23rd Sep 2014 23:02 UTC
Debian and its clones

Debian switched to Xfce as the default desktop environment back in November 2013. But that didn't last long because a few days ago, Debian restored GNOME as the default desktop, based on preliminary results from the Debian Desktop Requalification for Jessie.

According to Joey Hess, the Debian developer who performed this change, the main reasons for Debian switching back to GNOME as the default desktop are related to accessibility and systemd integration.

Order by: Score:
Correct link
by patrix on Tue 23rd Sep 2014 23:13 UTC
patrix
Member since:
2006-05-21
Makes sense...
by joekiser on Tue 23rd Sep 2014 23:34 UTC
joekiser
Member since:
2005-06-30

Gnome 3 is a very good desktop environment that gets better with each release, and XFCE development is stalled with no clear path forward.

Reply Score: 3

RE: Makes sense...
by righard on Wed 24th Sep 2014 13:33 UTC in reply to "Makes sense..."
righard Member since:
2007-12-26

Personally I don't mind the fact that the development of XFCE seems stalled. It has all the features I need, feels very coherent and with the right theme can look nice and modern.
If there'll be bug fixes, and fixes to future incompatibilities, I think I will use it for years to come.

Reply Score: 8

RE[2]: Makes sense...
by SonicMetalMan on Wed 24th Sep 2014 14:13 UTC in reply to "RE: Makes sense..."
SonicMetalMan Member since:
2009-05-25

I completely agree. I do not consider XFCE to be "stalled" it is simply a mature product that does what it's supposed to do. Too often changes are made just for sake of change and things get broken. This leads to another round of changes to fix the broken stuff that replaced the stuff that used to work correctly.

God help me but I really like LXDE too.

Reply Score: 8

good
by stabbyjones on Wed 24th Sep 2014 00:05 UTC
stabbyjones
Member since:
2008-04-15

It was annoying because I kept forgetting to change the desktop choice in advanced options before running the installer.

As much as I like XFCE/Razor-QT/Openbox style environments I just keep coming back to gnome because it includes everything I need without having to tweak anything. If gnome-shell didn't exist any more I think I'd have a breakdown.

It's been my primary desktop for about a decade now and it's always been my favourite visually and procedurally.

Reply Score: 4

RE: good
by tidux on Wed 24th Sep 2014 21:58 UTC in reply to "good"
tidux Member since:
2011-08-13

Are you a time traveller? GNOME Shell didn't exist until the 2010s.

Reply Score: 5

Comment by shmerl
by shmerl on Wed 24th Sep 2014 01:32 UTC
shmerl
Member since:
2010-06-08

Why does KDE support accessibility so poorly?

Reply Score: 1

RE: Comment by shmerl
by darknexus on Wed 24th Sep 2014 12:52 UTC in reply to "Comment by shmerl"
darknexus Member since:
2008-07-15

Why does KDE support accessibility so poorly?

Because it's not a priority for them. Simple as that. It never has been high on KDE's priority list and, while they've not exactly rebuffed the communities offers to help, they've not embraced them either.

Reply Score: 3

RE: Comment by shmerl
by Carewolf on Wed 24th Sep 2014 14:50 UTC in reply to "Comment by shmerl"
Carewolf Member since:
2005-09-08

They don't Debian just only measured accessibility based on whether the desktop supports the GNOME screen-reader API. They made no evaluation on anything else.

Reply Score: 6

RE[2]: Comment by shmerl
by shmerl on Wed 24th Sep 2014 16:13 UTC in reply to "RE: Comment by shmerl"
shmerl Member since:
2010-06-08

That's a poor metric then if KDE offers some other comparable features.

Reply Score: 3

RE[3]: Comment by shmerl
by Carewolf on Wed 24th Sep 2014 18:03 UTC in reply to "RE[2]: Comment by shmerl"
Carewolf Member since:
2005-09-08

That's a poor metric then if KDE offers some other comparable features.

Unfortunately it is not easily comparible. GNOME has better screen reader, but KDE has better accesible input settings. But it is incorrect to say that KDE has no accesibility when it has better support for many handicaps other than blindness.

Reply Score: 4

RE[4]: Comment by shmerl
by leos on Thu 25th Sep 2014 03:02 UTC in reply to "RE[3]: Comment by shmerl"
leos Member since:
2005-09-21

What's the problem with supporting Orca though? I thought Qt exposed info to the accessibility framework on linux?

Reply Score: 2

RE[5]: Comment by shmerl
by darknexus on Thu 25th Sep 2014 21:07 UTC in reply to "RE[4]: Comment by shmerl"
darknexus Member since:
2008-07-15

What's the problem with supporting Orca though? I thought Qt exposed info to the accessibility framework on linux?

Exposing is only a small part of a much larger issue. The problem is that KDE uses very few standard QT widgets. This means that, in addition to what the standard QT-AT-SPI bridge does with standard controls, special attention has to be paid to these nonstandard widgets to make exposure meaningful. Many QT apps have this issue not just KDE, and given the prevalence of custom widgets in QT apps we're a long way from having most of them be accessible on any platform to screen readers. Right now, all you get when you try to use KDE with Orca is "button, button, button" and that's just the standard buttons that are exposed but unlabeled. KDE's other controls speak nothing at all. You can try it yourself if you wish.

Reply Score: 3

MATE?
by sergio on Wed 24th Sep 2014 03:19 UTC
sergio
Member since:
2005-07-06

I think MATE makes much more sense for Debian than GNOME3.

Systemd, GNOME3... I hate the way Linux distributions are "evolving", more Windows-like than Unix-like. It's kind of sad for me.

PS: thank God for Slackware. ;)

Reply Score: 3

RE: MATE?
by Finalzone on Wed 24th Sep 2014 04:52 UTC in reply to "MATE?"
Finalzone Member since:
2005-07-06

I think MATE makes much more sense for Debian than GNOME3.


That means carrying the burden of maintaining depreciated library that MATE developers tried to get rid.

Systemd, GNOME3... I hate the way Linux distributions are "evolving", more Windows-like than Unix-like. It's kind of sad for me.

It if means to improve Linux ecosystem by removing legacy cruft and fully take advantage of untapped Linux kernel features, so be it.

Reply Score: 3

v RE: MATE?
by bassbeast on Wed 24th Sep 2014 05:55 UTC in reply to "MATE?"
v RE[2]: MATE?
by sergio on Wed 24th Sep 2014 06:56 UTC in reply to "RE: MATE?"
RE[3]: MATE?
by saynte on Wed 24th Sep 2014 07:40 UTC in reply to "RE[2]: MATE?"
saynte Member since:
2007-12-10

There are no slaves here, just people who made themselves dependent on others. You are dependent on your distro, your distro is dependent on the maintainers of the software they package.

Do you feel that those people writing init systems and those responsible for making sure they work on your distro are somehow not qualified to make decisions about init systems?

Reply Score: 5

RE[4]: MATE?
by nicubunu on Wed 24th Sep 2014 08:18 UTC in reply to "RE[3]: MATE?"
nicubunu Member since:
2014-01-08

Is not about how the code is *written* but about how it is *designed*.
When one writes a software system, before any line of code is put down, he has to imagine the usage scenarios and a target audience for it. Then if the code is written decently, users in the target audience and performing the anticipated usage scenarios will see their needs fulfilled.
On the other hand, those outside the target audience and having different usage scenarios, very likely will be unsatisfied with it.

Reply Score: 2

RE[5]: MATE?
by saynte on Wed 24th Sep 2014 09:44 UTC in reply to "RE[4]: MATE?"
saynte Member since:
2007-12-10

Yes, that is true. That is why the distributions' usage of systemd validates (at least partially) the design decisions. Distributions generally have a large target audience.

Reply Score: 4

RE[6]: MATE?
by nicubunu on Wed 24th Sep 2014 11:10 UTC in reply to "RE[5]: MATE?"
nicubunu Member since:
2014-01-08

actually you will see a lot of people unhappy their favorite distro has changed its target audience. some switch distros, some just complain.

Reply Score: 2

RE[3]: MATE?
by nicubunu on Wed 24th Sep 2014 08:12 UTC in reply to "RE[2]: MATE?"
nicubunu Member since:
2014-01-08

in other words, long term it seems the cathedral wins over the bazaar

Reply Score: 2

RE[3]: MATE?
by bassbeast on Wed 24th Sep 2014 18:31 UTC in reply to "RE[2]: MATE?"
bassbeast Member since:
2007-11-11

They can waste points modding me down but all they are doing is denying human nature...its the "busted shitter" problem.

If I ask you to paint me a picture, write me a poem or a song for free? I will get several offers, and some will probably be great...if on the other hand I ask you to fix my nasty busted toilet that I broke taking a giant dump? Most likely I will be peeing in the sink.

This is just human nature, nobody likes cleaning up others messes so these jobs DO NOT GET DONE unless pay is involved. in this case the pay is coming from a handful or corps like Red hat and guess what? They want a ROI and that ROI is THEY DECIDE the future of Linux, NOT YOU.

They can pretend that having the code magically makes it possible to have the manpower and skills required to rip out a major subsystem like systemd AND remove its dependencies AND rebuild all the programs that in some way hook into systemd but IRL that is just bullshit, reality is like Pulse and KDE 4 if the ones at the top decide "it will be thus" then that is what it will be, its like saying if the US military put the planes for the F35 on the web that means YOU can fix the flaws and have one in your backyard...YOU don't have the manpower, YOU don't have the expertise, YOU have no power in this case AT ALL because simply having source doesn't magically create 20+ men dev teams with years of low level programming exp out of thin air, this "you can fix it" mantra is just an illusion for anything more complicated than a simple text editor.

So let 'em waste mod points on delusions, we have already seen reality. Reality is that Pulse was released in a broken state and even mainline took it DESPITE what the users wanted, even the KDE devs admitted 4.0 wasn't alpha quality and the users ended up stuck with it DESPITE their protests and now no matter how much they scream they will get systemd DESPITE what they wish because users in Linux land? Have NO power, none.

Compare this to Windows, what happens when MSFT goes against its customers with its OS like Vista and 8? Sales tank, stock goes down, OEMs start demanding the previous version, bad press starts piling up which further drives down stock, so in the end they really end up with only one of two choices, they can 1.- Stay on the course and watch the company bleed to death or 2.- They can save their jobs by giving the people what they want. Now tell me, if the majority DO NOT WANT systemd how EXACTLY can they force RH and the other distros not to ram it down their throats? They can't, because they don't pay the bills, the corporate server companies do.

At the end of the day the old saying is true "if you are paying nothing you are the PRODUCT not the customer" and nowhere does that ring more true than in Linux land, where corporate interests get plenty of free labor on their server products by the community. But just as Google does not give a crap if you don't want ads these server companies are NOT beholden to the will of the users and if it comes down to them or you? Again see Pulse and KDE 4.

Reply Score: 3

RE[4]: MATE?
by saynte on Wed 24th Sep 2014 20:56 UTC in reply to "RE[3]: MATE?"
saynte Member since:
2007-12-10

If you don't like being dependent on the Linux distributions, then don't be. You can actually hurt them by moving away to Mac OS X or Windows, because they won't get (in the case of Fedora) as effective a test-bed for RedHat.

But really, can you think of a more fair exchange than the one we currently have between distribution and user?

It's funny you mention PulseAudio: I used to have a terrible time configuring ALSA to mix streams, and now I don't have to even think about it. It seems that the early testing we did actually raised my overall quality of audio-experience, so I kinda like that trade. But I admit, that it is a trade.

Reply Score: 4

RE[5]: MATE?
by bassbeast on Sat 27th Sep 2014 00:50 UTC in reply to "RE[4]: MATE?"
bassbeast Member since:
2007-11-11

So you agree you have no power in Linux land? Because you can't "hurt" them by "moving away" because you have paid zero dollars and unless you are a contributor you have paid zero hours so you might as well be dead to them, they really do not care.

Like it or not reality is you are NOT the customer, the customer is Dell, HP, and large datacenter operations. So unless you run a datacenter that is bringing in at least a high 7 figures worth of business to Red hat? You have absolutely ZERO say in what happens, you WILL take what you are given.

Reply Score: 2

RE[2]: MATE?
by Megol on Wed 24th Sep 2014 07:06 UTC in reply to "RE: MATE?"
Megol Member since:
2011-04-11

Is this a joke?

It is open source - fork it and remove whatever you want. Or (if you can't code and are too lazy to learn) start a campaign to attract people that can code and agree with your ideas...

Reply Score: 3

RE[3]: MATE?
by bassbeast on Wed 24th Sep 2014 18:40 UTC in reply to "RE[2]: MATE?"
bassbeast Member since:
2007-11-11

The planes for the MIG-15 are online, I'd like you to build me one by next Thursday...what do you mean you can't, you don't have the manpower?

DING DING DING, we have a winner Johnny! What you are describing works in the ABSTRACT but NOT in reality, because in reality you simply are not gonna be able to build up enough qualified dev people for free to fork something as huge as a major subsystem. in fact can you should me a single case of anything with over 200k lines of code that has EVER been forked by the community? Even one? You can't because reality and fantasy are two different things.

You see its basically an "is ought" problem, you are saying how things OUGHT to be, you OUGHT to be able to remove bad ideas from the tyrants at the top but reality IS they have a corp backing them and you? You have NO power, not even the power of voting with your wallet.

the simple fact is we have evidence in the past 10 years of voting with your wallet forcing a major corp to change direction TWICE, first with Vista and again with Win 8, what happened with Pulse? What happened with KDE 4? The will of the people over the will of the corp was ignored because having code isn't fairy dust and reality is if Red hat wants something and the people don't? Tough shit, they are a billion dollar corp and you are NOT their customer. the only ones that can force direction change in Linux land is corps like HP that buys RH products, like it or not that IS reality.

Reply Score: 3

RE[4]: MATE?
by johntdaly on Thu 25th Sep 2014 06:49 UTC in reply to "RE[3]: MATE?"
johntdaly Member since:
2012-07-27

Dude you sound bitter. I honestly like Linux and if you think things are goring wrong you can comfortably switch to FreeBSD. I for example don't like Windows so I use machines running Mac OS X and Linux.

Forking happens, just not by users and not by communities. Developers fork projects and new communities form around them. That is how OpenBSD and DragonFlyBSD formed. If you are sheep (a user) you can either get power (learn to develop) or vote with your feet (move to another distroy, os, whatever). Just don't expect that you can somehow spit vitreol around and organize a revoult of sheep.

Honestly you now have more choice then ever since the end of the early home computing era. Mac, Windows, Linux and FreeBSD are damn good choices and if you chose a *NIX you have more desktop choices then ever. All the complaints about anything programmed by Poettering and the changes done to the desktop by Gnome are starting to get annoying. FreeBSD with Mate would be EXACTLY the user experience you would now have on Linux none of that code ever got written. If it feels more raw then you might have hit upon the reason all that code was written to begin with.

So to reiterate. You have choice, use one of the 4 reasonably usable operating systems and if they don't suit you you can either use your money and pay developers do to build something for you (that is how Ubuntu got started) or use your development skills to build something new yourself. THAT IS IT.

I am tired of Windows users who steal all their software and run their shit on cobbled together hardware spouting off at the top of their lungs because they don't have a way to influence what is happening. Deal with it, if you want change you have to DO something, talking isn't enough.

Reply Score: 3

RE[5]: MATE?
by allanregistos on Fri 26th Sep 2014 03:31 UTC in reply to "RE[4]: MATE?"
allanregistos Member since:
2011-02-10

Dude you sound bitter. I honestly like Linux and if you think things are goring wrong you can comfortably switch to FreeBSD. I for example don't like Windows so I use machines running Mac OS X and Linux.

Forking happens, just not by users and not by communities. Developers fork projects and new communities form around them. That is how OpenBSD and DragonFlyBSD formed. If you are sheep (a user) you can either get power (learn to develop) or vote with your feet (move to another distroy, os, whatever). Just don't expect that you can somehow spit vitreol around and organize a revoult of sheep.


You cannot understand what he meant?

It means that you cannot fork an existing system overnight as a user(You have no choice by default but to use what's available). Maybe some influential developer with enough free time and resources, he can fork or create another version of systemd or audio server. But these influential or talented developers are not simply available because they have day jobs and families.

Yes fork happens, see OpenIndiana(now many years and look at them), given that these people are former Solaris developers, they cannot even release a major stable version so far. A complex software system can be forked yes, but it is not easy. This (fork it if you do not like what the project is going) thing is only possible for small projects, simple single application etc. (Try forking Blender for example, if you do not like where Blender is going) Try to do that, I challenge anyone on this thread. You can't if you do not have enough experience of doing 3D graphics programming(specific field). Talented developers cannot just simply fork a software because they were talented, it is more than the talent.

Programming is like an art, it can't be learned in an instant, you have to learn and experience it for at least 10 years at the minimum before you can claim to least fork a system like systemd.

Reply Score: 1

RE[6]: MATE?
by johntdaly on Fri 26th Sep 2014 07:30 UTC in reply to "RE[5]: MATE?"
johntdaly Member since:
2012-07-27

Systemd was already froked. I'm a dev, I know how hard it is to learn to program something worth wile and I do get his point. It's just not a particularly good point. If you can't fork systemd because you don't understand it you don't really have the right to vent so much hate in the direction of its developers, like you know better then the they do.

And the saying something as stupid as you can't influence open source but you can commercial software by not buying it is just bullshit. You can move away form software you don't like and honestly the BSDs are good operating systems and the support that FreeBSD is getting now is a hell of a lot better then the support Linux got when a lot of us started using it.

My point was simple, all of us have choices we can make. If you are a dev or a rich person you have more choices then an average user. I'm a dev, I can fork a system. He is a user he CAN move to another system, like FreeBSD. But he portrays the situation like devs can just ram whatever they like down his throat when that is just not true.

We all have different potential for changing what is going on.

Reply Score: 1

RE[4]: MATE?
by SeanParsons on Thu 25th Sep 2014 15:02 UTC in reply to "RE[3]: MATE?"
SeanParsons Member since:
2011-01-11

I can think of several significant forks that involved the community/a handful of devs to start, and then attracted more to their cause. 1) MariaDB was forked by the original developers of MySQL and seems to be doing quite well. 2) Knoppix was a fork of Debian done primarily by Klaus Knopper and it has given us the modern world of live distros. 3) Theo de Raadt did fork NetBSD to OpenBSD and then attracted more devs. 4) While it did not involve a small number of people LibreOffice was forked by a large number of developers and appears to be more popular than the original OOo.

That said, successful forks of large software projects are not all that common probably because if users dislike it enough in the free software world they will go somewhere else. They don't NEED to fork it.

Reply Score: 3

RE[4]: MATE?
by zima on Tue 30th Sep 2014 17:25 UTC in reply to "RE[3]: MATE?"
zima Member since:
2005-07-06

a single case of anything with over 200k lines of code that has EVER been forked by the community? Even one? [...] direction change

Libre Office?

Reply Score: 2

RE[3]: MATE?
by grat on Wed 24th Sep 2014 19:55 UTC in reply to "RE[2]: MATE?"
grat Member since:
2006-02-02

Is this a joke?

It is open source - fork it and remove whatever you want. Or (if you can't code and are too lazy to learn) start a campaign to attract people that can code and agree with your ideas...


I want a new rule on the internet. No one's allowed to say "just fork it!" until they've actually forked a project of no fewer than 10,000 lines of code.

Sorry, I have a job, and it's not as a software developer. If that prohibits me from having an opinion on crappy software, too bad.

Reply Score: 4

RE[4]: MATE?
by Alfman on Wed 24th Sep 2014 20:24 UTC in reply to "RE[3]: MATE?"
Alfman Member since:
2011-01-28

grat,

I want a new rule on the internet. No one's allowed to say "just fork it!" until they've actually forked a project of no fewer than 10,000 lines of code.

Sorry, I have a job, and it's not as a software developer. If that prohibits me from having an opinion on crappy software, too bad.


Having created my own gmlinux distro for my own services, I think I'd qualify, but you are right. I sought to solve a bunch of problems I was having, in many cases I achieved my goals: a tiny distro that runs well on low end servers, remote server provisioning & recovery. However the maintenance burden has become higher than I am able to commit to. I cannot afford to keep doing the work without being paid, and alas it's hard to be paid when existing distros are "good enough".

Edited 2014-09-24 20:29 UTC

Reply Score: 3

RE[4]: MATE?
by bassbeast on Sat 27th Sep 2014 01:05 UTC in reply to "RE[3]: MATE?"
bassbeast Member since:
2007-11-11

Just give 'em the same response I do, "The MIG 15 plans are online, I'd like you to build me one and have it delivered by Thursday" because saying "just fork it" is NO DIFFERENT than saying "just build your own fighter plane" because in both cases it requires a HUGE amount of resources and expertise that Joe Normal has ZERO chance of having!

So please use my little retort to point out how frankly insane the "just fork it" meme is, since we have the plans for millions of things online, from the MIG 15 to silverware, yet I don't see anybody honestly suggesting you should build your own kiln and start forging your own tableware because the ones being offered suck, do you?

What I think is ironic is nobody notices the DARK and nasty undertone of that statement because by making "just fork it" seem like a trivial thing they are saying programmers are worthless people whose time is worth nothing. You'd think that the average Linux advocate would have a little more respect for the amount of effort it takes to become a coder with the skills to be able to do such a task, but I guess not since their work is so trivial and all.

Reply Score: 2

RE: MATE?
by hussam on Wed 24th Sep 2014 07:05 UTC in reply to "MATE?"
hussam Member since:
2006-08-17

I think MATE makes much more sense for Debian than GNOME3.

Systemd, GNOME3... I hate the way Linux distributions are "evolving", more Windows-like than Unix-like. It's kind of sad for me.

PS: thank God for Slackware. ;)

This really depends on your distribution. Systemd is a lot simpler, easier to maintain and administer than the old hackish Arch Linux boot scripts which were very buggy.
Another thing is that systemd bridges the gap between distributions and eases work for upstream developers of server applications who no longer have to worry about the million ways distributions used their applications in the past.

The only issue I have with systemd is the fragile journal storage.

So yes, systemd is pure Linux-ism but so what?

Edited 2014-09-24 07:08 UTC

Reply Score: 4

RE[2]: MATE?
by javispedro on Wed 24th Sep 2014 16:49 UTC in reply to "RE: MATE?"
javispedro Member since:
2014-06-04

Hm, didn't Arch get popular in the first place because of its over-simplistic init scripts?

Reply Score: 1

RE[3]: MATE?
by hussam on Wed 24th Sep 2014 17:36 UTC in reply to "RE[2]: MATE?"
hussam Member since:
2006-08-17

Hm, didn't Arch get popular in the first place because of its over-simplistic init scripts?

It was the centralized /etc/rc.conf
Then they ended up contributing to systemd. That's partly why it works well on that distribution.

Edited 2014-09-24 17:51 UTC

Reply Score: 3

RE[4]: MATE?
by zlynx on Thu 25th Sep 2014 00:07 UTC in reply to "RE[3]: MATE?"
zlynx Member since:
2005-07-20

systemd unit files are really simple so that fits well with Arch I think.

I spent some time reading the systemd man pages and some of the existing .service files and then wrote a couple of my own.

My service files only need about 4 lines to get the job done. SysV scripts had comment block headers bigger than that.

Reply Score: 3

RE[2]: MATE?
by Lennie on Wed 24th Sep 2014 19:49 UTC in reply to "RE: MATE?"
Lennie Member since:
2007-09-22

How about a fork of systemd with a very minimal feature set which they are also porting to FreeBSD:

http://www.phoronix.com/scan.php?page=news_item&px=MTc5MzA

Reply Score: 3

RE[3]: MATE?
by grat on Wed 24th Sep 2014 19:57 UTC in reply to "RE[2]: MATE?"
grat Member since:
2006-02-02

How about a fork of systemd with a very minimal feature set which they are also porting to FreeBSD:

http://www.phoronix.com/scan.php?page=news_item&px=MTc5MzA


By "they", you mean the uselessd devs, not the FreeBSD devs, I presume.

I'm not opposed to the idea, except that by calling it 'uselessd', they're seriously lowering expectations of utility.

Reply Score: 4

RE[4]: MATE?
by Lennie on Wed 24th Sep 2014 20:07 UTC in reply to "RE[3]: MATE?"
Lennie Member since:
2007-09-22

By "they", you mean the uselessd devs, not the FreeBSD devs, I presume.


yes

Reply Score: 2

RE[3]: MATE?
by Alfman on Wed 24th Sep 2014 20:02 UTC in reply to "RE[2]: MATE?"
Alfman Member since:
2011-01-28

Lennie,

How about a fork of systemd with a very minimal feature set which they are also porting to FreeBSD:


I haven't gotten any time to play around with it at all, but it sounds intriguing. It should trim most of the bloat that makes systemd most controversial.

http://uselessd.darknedgy.net/

Whether it will ever be incorporated into any major distros remains to be seen.

Reply Score: 2

RE: MATE?
by moondevil on Wed 24th Sep 2014 07:17 UTC in reply to "MATE?"
moondevil Member since:
2005-07-08

Windows-like than Unix-like


You mean Atari-ST, Amiga, Archimedes, OS/2, Mac OS like?

A UNIX like system is by definition either a Workstation or a server.

That is a consequence of catering to the desktop. Mac OS X is full of examples more-Windows like as well.

Reply Score: 3

RE[2]: MATE?
by Morgan on Wed 24th Sep 2014 15:36 UTC in reply to "RE: MATE?"
Morgan Member since:
2005-06-29

A UNIX like system is by definition either a Workstation or a server.


Is it really, though? GNU/Linux has always been considered a UNIX-like system, and it runs on embedded machines, thin clients, network appliances (routers, APs, print managers, etc.), telemetry systems, and so on.

And then there's this, from Wikipedia[1] (emphasis mine):

A Unix-like (sometimes referred to as UN*X or *nix) operating system is one that behaves in a manner similar to a Unix system, while not necessarily conforming to or being certified to any version of the Single UNIX Specification.

There is no standard for defining the term, and some difference of opinion is possible as to the degree to which a given operating system is "Unix-like".


[1] https://en.wikipedia.org/wiki/Unix-like

Reply Score: 5

RE: MATE?
by p13. on Wed 24th Sep 2014 07:24 UTC in reply to "MATE?"
p13. Member since:
2005-07-10

I've been running linux full time for nearly 20 years now.

I don't mind systemd. I think it's a big step up from SYSV style init scripts. I don't know if you've ever had problems with boot sequences on unix systems, and had to debug it before, but it's just a huge clusterf--k of badly written scripts, stale lock files and screwy symlinks.

I have always admired the solaris and OS X way of doing things. It's only natural that linux would follow. Perhaps you consider solaris to be "too much like windows" too?

As far as your desktop choice goes ...
If you want a unix-like desktop, you are free to install twm, fvwm or CDE if you'd like ...

Edited 2014-09-24 07:25 UTC

Reply Score: 4

RE[2]: MATE?
by NicePics13 on Wed 24th Sep 2014 07:57 UTC in reply to "RE: MATE?"
NicePics13 Member since:
2009-06-08

I don't mind systemd. I think it's a big step up from SYSV style init scripts

Sure, then again Slackware solved this quite elegantly: simple BSD style init + SYSV compatibility if needed.

If you want a unix-like desktop, you are free to install twm, fvwm or CDE if you'd like

Window Maker? :^D

Reply Score: 4

RE[3]: MATE?
by p13. on Wed 24th Sep 2014 08:11 UTC in reply to "RE[2]: MATE?"
p13. Member since:
2005-07-10

I really like wmaker.

Reply Score: 2

RE[3]: MATE?
by Morgan on Wed 24th Sep 2014 15:38 UTC in reply to "RE[2]: MATE?"
Morgan Member since:
2005-06-29

Sure, then again Slackware solved this quite elegantly: simple BSD style init + SYSV compatibility if needed.


Which is why it's on the short list of GNU/Linux distros I'll use once Debian makes the switch to systemd with their next release.

Reply Score: 2

RE[2]: MATE?
by Soulbender on Wed 24th Sep 2014 10:42 UTC in reply to "RE: MATE?"
Soulbender Member since:
2005-08-18

I think it's a big step up from SYSV style init scripts.


Yeah, but that's a pretty low bar, you know. Not having anything is better than SysV style init. If systemd did that one thing, manage services, I would probably not have minded it much but unfortunately, being a product of RH/Poettering, it's a veritable kitchen sink of mish-mashed functionality and NIH syndrome.
Tools like daemontools/runit/upstart does one thing and does it well: manage services.

Reply Score: 4

RE[3]: MATE?
by p13. on Wed 24th Sep 2014 10:48 UTC in reply to "RE[2]: MATE?"
p13. Member since:
2005-07-10

It's true that they have really gone overboard with extra "functionality" that really shouldn't be implemented at that level.
Upstart also comes from a place with a lot of NIH-ness. That being said, i like upstart. It's simple enough, and it does it's job.

Quite a while ago, i was helping to beta-test a project called e-init. Now that was something!
Very fast dependency based parallel loading of daemons and services, intelligent caching, event based starting of daemons, loading and unloading of modules, etc
It was really cool. I knew one of the main devs quite well. I gave him an SGI Indigo2 ;)

Edited 2014-09-24 10:52 UTC

Reply Score: 3

RE[2]: MATE?
by grat on Wed 24th Sep 2014 19:59 UTC in reply to "RE: MATE?"
grat Member since:
2006-02-02

I don't know if you've ever had problems with boot sequences on unix systems, and had to debug it before, but it's just a huge clusterf--k of badly written scripts, stale lock files and screwy symlinks.


Only on OpenSuSE 12.3 or later... which runs systemd.

Reply Score: 2

RE: MATE?
by darknexus on Wed 24th Sep 2014 12:43 UTC in reply to "MATE?"
darknexus Member since:
2008-07-15

I think MATE makes much more sense for Debian than GNOME3.

Not if accessibility is one of the reasons for doing this. Make uses so many custom widgets and controls that none of the accessibility software can make much sense of it.

Reply Score: 3

Why?
by kwan_e on Wed 24th Sep 2014 11:41 UTC
kwan_e
Member since:
2007-02-18

Why would they change the default desktop environment from <desktop environment that I don't like> to <another desktop environment that I don't like>? They should have gone for the good option, which is <the desktop environment I like>.

Their problem is that they don't listen to users (<named me>) and they're just going to drive users (<who aren't like me>) away.

That's it. I'm switching to <not this distro>.

Edited 2014-09-24 11:43 UTC

Reply Score: 7

RE: Why?
by SpyroRyder on Wed 24th Sep 2014 12:01 UTC in reply to "Why?"
SpyroRyder Member since:
2014-08-25

This is the perfect comment template for the Slashdot version of this story. Which does make me wonder how many will read like this

Reply Score: 2

Comment by schadfield
by schadfield on Wed 24th Sep 2014 15:07 UTC
schadfield
Member since:
2012-04-16

I am running Gnome shell on CentOS 7. Very happy with it. As for systemd - I assume it is installed and is working fine. It has given me no problems I have not even had to think about it. That is the way it should be...

Reply Score: 3

RE: Comment by schadfield
by jockm on Wed 24th Sep 2014 16:18 UTC in reply to "Comment by schadfield"
jockm Member since:
2012-12-22

I assume it is installed and is working fine. It has given me no problems I have not even had to think about it. That is the way it should be...


Indeed. I personally like systemd, and wonder how much of the refusal to use it is theoretical than practical.

I have heard all the arguments against it, but I haven't experienced any of them. Nor have any of the IT people I know experienced a problem with it. They (they, you know, them) say it is un-unix like but I personally don't see it — and I have been a unix user since the 1987.

My only complaint is that it is GCC and Linux dependant, mind you these are theoretical issues and not practical ones. Right now all my *nix devices are Linux, and GCC is still the norm.

Reply Score: 4

RE[2]: Comment by schadfield
by jessesmith on Wed 24th Sep 2014 16:44 UTC in reply to "RE: Comment by schadfield"
jessesmith Member since:
2010-03-11

I think you mean it is GNU C library and Linux dependent. I don't believe anything about systemd makes it dependent on GCC as opposed to, say, Clang.

This is one reason I'm happy to see uselessd come along. It makes the systemd code more portable and avoids locking the operating system into one C library or kernel.

Reply Score: 5

RE[3]: Comment by schadfield
by jockm on Wed 24th Sep 2014 20:51 UTC in reply to "RE[2]: Comment by schadfield"
jockm Member since:
2012-12-22

I think you mean it is GNU C library and Linux dependent. I don't believe anything about systemd makes it dependent on GCC as opposed to, say, Clang.


Indeed you are quite right, thank you for catching that

Reply Score: 3

RE[2]: Comment by schadfield
by WereCatf on Wed 24th Sep 2014 18:08 UTC in reply to "RE: Comment by schadfield"
WereCatf Member since:
2006-02-15

Indeed. I personally like systemd, and wonder how much of the refusal to use it is theoretical than practical.


I don't get the problem, really. Most people seem to be against systemd only because they believe it's linux-y, not unix-y, and that's that; no other arguments, no actual problems they've encountered with it or anything, just the fact that it doesn't strictly adhere to some specific guideline these people want it to adhere to and even then only for ideological -- not any practical -- reasons.

Reply Score: 4

RE[3]: Comment by schadfield
by vivainio on Wed 24th Sep 2014 19:02 UTC in reply to "RE[2]: Comment by schadfield"
vivainio Member since:
2008-12-26

Anti-systemd arguments tend to degenerate to incoherent, long-winded fundamentalist ramblings very quickly. It doesn't really reflect well against the current Linux community ("can't these kids even agree on this simple stuff?").

At certain sw development mileage, you learn how the field works: things change, you refactor, fix and roll with the punches until the sw is "good enough" or superior alternative comes around. Rinse and repeat. Alternative is irrelevance and stagnation.

Reply Score: 3

RE[4]: Comment by schadfield
by Alfman on Wed 24th Sep 2014 19:53 UTC in reply to "RE[3]: Comment by schadfield"
Alfman Member since:
2011-01-28

vivainio,

Anti-systemd arguments tend to degenerate to incoherent, long-winded fundamentalist ramblings very quickly. It doesn't really reflect well against the current Linux community.


I disagree, systemd violates some long held principals about simple tools that individually do simple things and do them well. It's really not incoherent to speak out for "keep it simple, stupid" principals. Systemd, for better or worse, tries to be everything, overlapping with lots of other functionality (init/power management/cron/syslog/inetd/automounting/udev/etc) and is inherently complex as a result. Furthermore it depends on dbus, which is no stranger to it's own controversies.


Note that personally, I disliked the sysv style init scripts, and I view systemd as an overall improvement despite the fact that it is a dramatic departure from the old scripted init systems. However I still question whether the monolithic approach taken by systemd was the best solution, it seems to already have trouble with "mission creep".

("can't these kids even agree on this simple stuff?").


We never have and never will! I don't think this is unique to linux though. It's just that the internal developer struggles get magnitudes more public exposure, whereas closed platform development happens behind closed doors.

Reply Score: 4

RE[5]: Comment by schadfield
by jockm on Wed 24th Sep 2014 20:59 UTC in reply to "RE[4]: Comment by schadfield"
jockm Member since:
2012-12-22


I disagree, systemd violates some long held principals about simple tools that individually do simple things and do them well. It's really not incoherent to speak out for "keep it simple, stupid" principals. Systemd, for better or worse, tries to be everything, overlapping with lots of other functionality (init/power management/cron/syslog/inetd/automounting/udev/etc) and is inherently complex as a result. Furthermore it depends on dbus, which is no stranger to it's own controversies.


So aside from a philosophy issue, is that actually bad? Can you point me to some precise problem from SystemD not being implemented as separate modules.

And if someone were to arbitrarily broke systemd up in to modules would that then eliminate your concerns.

Philosophies and metaphors are great, but dogma is bad. Sooner or later every working system is going to push the limits of its underlying philosophy/metaphor at which point it has to adapt to the real world.

Reply Score: 4

RE[6]: Comment by schadfield
by vivainio on Thu 25th Sep 2014 10:32 UTC in reply to "RE[5]: Comment by schadfield"
vivainio Member since:
2008-12-26

Can you point me to some precise problem from SystemD not being implemented as separate modules.


One thing that is *not* a problem is often-alleged "running everything as PID 1". Systemd launches new processes for various things it's doing.

Reply Score: 3

RE[6]: Comment by schadfield
by Alfman on Thu 25th Sep 2014 15:16 UTC in reply to "RE[5]: Comment by schadfield"
Alfman Member since:
2011-01-28

jockm,

So aside from a philosophy issue, is that actually bad? Can you point me to some precise problem from SystemD not being implemented as separate modules.


Well, like many software design philosophies, end users don't care as long as the result is good enough. The details are "out of sight out of mind". It's mostly custom system developers who will get their hands dirty. With systemd, changes to some of the native subsystems (ie journald) will require forking/patching systemd's official compiled C binaries. Again, it's not something most users will ever think about, but it can be offputting to sysadmins who are not programmers. Hypothetically niche subsystem use cases could merged into the systemd upstream code base, but that is "feature creep" and additional complexity that will only confuse typical users.

And if someone were to arbitrarily broke systemd up in to modules would that then eliminate your concerns.


Yea, if they moved away from the monolithic approach, that would probably make systemd more ideal. However without a philosophy shift back to adopting independent daemons, I don't know if there's much chance systemd will change.

I suppose a fork like "uselessd" could make it happen, but will it reach the critical mass needed to give it widespread support?

Philosophies and metaphors are great, but dogma is bad. Sooner or later every working system is going to push the limits of its underlying philosophy/metaphor at which point it has to adapt to the real world.


Do you think I'm being dogmatic though? I'm thrilled to move away from sysv scripts. Systemd works, but it's large scope seems arbitrary and unnecessary.

Edited 2014-09-25 15:25 UTC

Reply Score: 3

RE[3]: Comment by schadfield
by jessesmith on Thu 25th Sep 2014 15:07 UTC in reply to "RE[2]: Comment by schadfield"
jessesmith Member since:
2010-03-11

I can give you two practical examples of systemd cauesing problems and one annoyance.

1. The binary logging system tends to get corrupted. When that happens there is not much you can do about it. When a text file gets corrupted, most of the log is still usable, not the case with the binary log systemd creates. The systemd developers don't want to fix this, so admins have to either deal with corrupted logs or work around systemd to get text logs.

2. systemd will monitor services and kill/restart them if they hang. Some daemons do processing of log files or scan sysetm directories and this can take several seconds. systemd will see a busy daemon as being unresponsive and kill it. Some daemons need to be patched/re-written to avoid being killed by systemd, even though they are working fine to begin with and run fine on non-systemd init systems.

3. systemd is not portable. This means one more "standard" floating around the open source world and it means more work for those of us writting low-level services as we need to test on yet another init technology.

Edited 2014-09-25 15:07 UTC

Reply Score: 3

RE[4]: Comment by schadfield
by Finalzone on Thu 25th Sep 2014 20:41 UTC in reply to "RE[3]: Comment by schadfield"
Finalzone Member since:
2005-07-06

1. The binary logging system tends to get corrupted. When that happens there is not much you can do about it. When a text file gets corrupted, most of the log is still usable, not the case with the binary log systemd creates. The systemd developers don't want to fix this, so admins have to either deal with corrupted logs or work around systemd to get text logs.

As explained on the bug report: https://bugs.freedesktop.org/show_bug.cgi?id=64116

currently, is to ignore them: when a corruption is detected, journald will rename the file to <something>.journal~, and journalctl will try to do its best reading it. Actually fixing journal corruptions is a hard job, and it seems unlikely that it will be implemented in the near future.

3. systemd is not portable. This means one more "standard" floating around the open source world and it means more work for those of us writting low-level services as we need to test on yet another init technology.

Let remember systemd was designed for Linux in mind to take advantage of untapped faetures (for example cgroups).

Reply Score: 3

Gnome should bring back the desktop.
by NewTron on Wed 24th Sep 2014 23:08 UTC
NewTron
Member since:
2012-07-27

Gnome should bring back the desktop and forget the smart TV interface of Gnome 3. The damage they have done to all Linux community is unmeasurable.

Reply Score: 1

tylerdurden Member since:
2009-03-17

The damage they have done to all Linux community is unmeasurable.


So if i5 can not be measured, why are you blaming them as if you had actually measured it?

Reply Score: 2

So... systemd does no work with xfce ??
by PieterGen on Thu 25th Sep 2014 13:13 UTC
PieterGen
Member since:
2012-01-13

".....the main reasons for Debian switching back to GNOME....are related to accessibility and systemd integration" Accessibility, that I get. This means that in Gnome it is easier for people with bad eyes to adapt the look & feel, right?

But systemd integration ??? I use archlinux (with systemd) and it runs fine with every windowmanager I tried: OpenBox, Awesome, i3WM, Herbstluft, XFCE.... So, what am I missing here? Why would systemd lead to a choice for Gnome??

Edited 2014-09-25 13:14 UTC

Reply Score: 2