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 ... 15, 16, 17, 18  Next  
Reply to topic    Gentoo Forums Forum Index Other Things Gentoo
View previous topic :: View next topic  
Author Message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3436
Location: The Peanut Gallery

PostPosted: Fri Mar 28, 2014 8:31 am    Post subject: Reply with quote

1clue wrote:
most people just took the default for whatever distro, or installed a few and played with all of them.

That ofc is what lies behind the push for Gentoo to switch its default to systemd; if people use it by default they won't even know about alternatives, much like the impetus behind subsidising OEMs to pre-install Windows, or indeed the browser anti-trust cases.

We're practically the only hold-out amongst the well-known distributions, and that bites when you want to be able to say "every major Linux distro uses systemd, so rewrite your apps to depend on it."

OFC that completely ignores the realities of building a distro from the ground-up. As discussed ad-nauseam, Gentoo is built from GNU software, which means component-based, since GNU originally had no OS, and the tools were written to install practically anywhere, piecemeal. That also happens to chime with the ethos and craft of software-engineering, which is why it's been so successful. Someone working on find, for example, does not need to worry about what the GIMP is doing or needs, and vice-versa.

Yet people are advocating exactly the equivalent of that for our core system, then getting irritated and resorting to personal attacks (such as "fusty traditionalist" or "hater") when more experienced heads all say "no way." If you haven't learnt the basics, and that means seeing why they're the basics in real-life, which takes at least a few years of making those mistakes in real-world development, then don't expect us to take you seriously.

And don't for one second think we're going anywhere. We're not going to "fork Gentoo" so you can claim the putsch as your victory for systemd, just as we won't surrender the Linux userland to a Windows-clone so RedHat can make more money off consumers-not-users.

We don't stop you doing what you want, so kindly do the same for us. Or y'know, get thee to /dev/null ;p
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 2588
Location: Dallas area

PostPosted: Fri Mar 28, 2014 9:39 am    Post subject: Reply with quote

steveL wrote:
We don't stop you doing what you want, so kindly do the same for us. Or y'know, get thee to /dev/null ;p


:lol: Indeed

1clue wrote:
I've used blackbox off and on, and recently discovered i3 as a more modern alternative


I've never used i3 but I prefer a composting wm vs a tiling one (just personal preference)

General:
I use lxde (openbox under it) but I really only use two parts of lxde, the panel and the image viewer,
so I could go with just a wm and add the few other pieces that I need, no matter which wm.
I certainly don't need or want a full blown DE with lots of things that I will never use.

And because of that I don't need or want anything more than openrc as my init system.
I've recently switched from an old version of udev to eudev and it's working fine for my needs.

Once again, I say, choice is good. I could care whether some corporation likes my choices, they're mine. :wink:
_________________
Asus m5a99fx, FX 8320 - amd64-multilib, 3.15.9-zen, glibc-2.19, gcc-4.9.2, eudev
xorg-server-1.16, openbox w/lxpanel, nouveau, oss4(2011)
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Fri Mar 28, 2014 3:12 pm    Post subject: Reply with quote

Anon-E-moose wrote:

I've never used i3 but I prefer a composting wm vs a tiling one (just personal preference)


That's got to be the funniest typo I've seen in awhile, and strangely appropriate.

You mean compositing wm, compost is a pile of ... biological refuse ... being maintained under certain conditions until it turns into fertilizer. This one is full of glee in the context of this thread's discussions.
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 2588
Location: Dallas area

PostPosted: Fri Mar 28, 2014 3:20 pm    Post subject: Reply with quote

LoL, yep, that's what I get for not checking my spelling :lol:

Quote:
This one is full of glee in the context of this thread's discussions.


Agreed.
_________________
Asus m5a99fx, FX 8320 - amd64-multilib, 3.15.9-zen, glibc-2.19, gcc-4.9.2, eudev
xorg-server-1.16, openbox w/lxpanel, nouveau, oss4(2011)
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Fri Mar 28, 2014 3:58 pm    Post subject: Reply with quote

Spell checker won't get you out of that one, it's a real word. Composting means either the act of making fertilizer out of biological waste or a machine which can do so.

Frankly I like it better the way you typed it.

Getting back on the off-topic window manager topic, I like compositing wms better too, but I also like the completely keyboard-driven feature set of i3. Which is why I went back to fvwm, but as yet haven't configured it at all.

Now with respect to "composting" window managers/desktop environments, the two that come to mind are gnome and kde. KDE at least doesn't leave ... biological waste ... sitting around infecting the rest of your system even if it's overly huge and ungainly. But gnome seems to be not very effective at composting because once you install it you can never be entirely rid of its effects. The ...biological waste... will sit around in odd corners of your hard disk, causing grief whenever it can.

I'll try to cut back on the glee. This is way too much fun.
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 2588
Location: Dallas area

PostPosted: Fri Mar 28, 2014 4:12 pm    Post subject: Reply with quote

I liked xfwm well enough when I was using xfce.
But it became clear that they wanted to keep on hanging onto being gnome-light
and I went seeking other non-gnome things.

I'm happy enough with openbox though it's only does stacking,
I like the ability to configure lots of keyboard shortcuts,
and have added compton for transparency effects.

I haven't used fvwm since an old 1.something version.
I was just looking at some screenshots for the 2.* version and it looks like it's been modernized
I especially like the patched version for rounded windows.
_________________
Asus m5a99fx, FX 8320 - amd64-multilib, 3.15.9-zen, glibc-2.19, gcc-4.9.2, eudev
xorg-server-1.16, openbox w/lxpanel, nouveau, oss4(2011)
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3436
Location: The Peanut Gallery

PostPosted: Fri Mar 28, 2014 5:27 pm    Post subject: Reply with quote

1clue: I hope you're looking at fvwm-crystal (it's in portage.)
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Fri Mar 28, 2014 6:45 pm    Post subject: Reply with quote

Yeah, fvwm-crystal looks nice but it loses a few features I really like. Like the window dimensions when you resize, or position when you move it.

When I get around to it, I'll add crystal. Maybe.

Thanks.
Back to top
View user's profile Send private message
mv
Advocate
Advocate


Joined: 20 Apr 2005
Posts: 4602

PostPosted: Fri Mar 28, 2014 7:33 pm    Post subject: Reply with quote

1clue wrote:
Yeah, fvwm-crystal looks nice but it loses a few features I really like. Like the window dimensions when you resize, or position when you move it.

In the thread Fvwm-Crystal, the return I described in some posting how to change this and a few other things I did not like, too, in the beginning. After a few such configuration changes, I can warmly recommend fvwm-crystal. Be sure to read the follow-ups to this posting (in particular, to install the latest version because of a few security fixes). Unfortunately, the latest version is not in the main tree (but e.g. in the mv overlay.)
Back to top
View user's profile Send private message
Dominique_71
Veteran
Veteran


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

PostPosted: Fri Mar 28, 2014 10:27 pm    Post subject: Reply with quote

I am an old GNU/Linux user back to my 386. I try all the wm/desktops and well, fvwm-crystal is the best for me.

My first decent PC was an Amiga 2000. It showed me what was a good OS is. One of its main feature is that even the toolkit is integrated into the graphical server. The result was a preemptive and multitasking OS with wonderful audio and video capabilities for that time with 1MB RAM for the A2000 and 500 ko for the A500. It was also a very simple but powerful OS. What I think about that "One True Desktop" thing is they must try the Amiga OS, or its free AROS implementation, if they want to learn about how to do it in an efficient way.

Hosted AROS into fvwm-crystal is much fastest than native fvwm:
http://s30.postimg.org/6bt0axsip/screenshot_20140328_232145.png
_________________
[[[ 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
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 33113
Location: 56N 3W

PostPosted: Sat Mar 29, 2014 9:06 am    Post subject: Reply with quote

Dominique_71,

Efficiency no longer matters. Its easy to add RAM and CPU power today.
With the Amigas, it had to work in the 1M because adding more was out of the question.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Dominique_71
Veteran
Veteran


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

PostPosted: Sat Mar 29, 2014 10:52 am    Post subject: Reply with quote

NeddySeagoon wrote:
Dominique_71,

Efficiency no longer matters. Its easy to add RAM and CPU power today.
With the Amigas, it had to work in the 1M because adding more was out of the question.


With such a biased and wrong design choice, GNU/Linux can only become the same shit than windows on the long run, or even worst:

Quote:
- Oh man, look at my computer, it run the GUI.
- Nice, and what are you doing?
- I just said you, I run the GUI!


It was possible to expand the RAM on the A2000, which was a must to be able to do some serious work with it. It is peoples that want a GUI that is not in their way, that in order to do some work too. And the Amigas was providing it. When I read statements like yours, I am worried about GNU/Linux. Why should GNU/Linux doesn't matter about efficiency any more? Because our whole society is about is about competition, which is a waste of resources at the first place? And we must reproduce it into GNU/Linux? I don't buy it.
_________________
[[[ 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
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 33113
Location: 56N 3W

PostPosted: Sat Mar 29, 2014 11:09 am    Post subject: Reply with quote

Dominique_71,

Let me temper my statement that "Efficiency no longer matters."
I learned to program on a computer with 16k words of storage. If your program would not fit into that, it could not be run.
This was 1963 and the machine was an Elliot 503. This taught Efficiency in a way that has been lost today with the exception of a few embedded systems.

There is a similar Efficiency in the real time world. If you are calculating video fames at 50Hz, thats one every 20ms, there is a hard limit at 20ms.
That too teaches other aspects of Efficiency but to a different set of constraints.

-- edit because my keyboard stopped working --

Changes an hardware have made Efficiencies like the above examples unnessacary. At that time, harware floating point was not even an option, so floating point arithmetic was never used.
Fixed point is a wonderful thing but a dead or dying skill. Likewise, if you needed sin or cos in real time, it was lookup table driven.

Efficiency today is just different.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.


Last edited by NeddySeagoon on Sat Mar 29, 2014 11:33 am; edited 1 time in total
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 2588
Location: Dallas area

PostPosted: Sat Mar 29, 2014 11:23 am    Post subject: Reply with quote

Early systems, cp/m, 1st dos, amiga OS, etc were wonderful systems for the time.

Yes they were efficient for the time, but that was because they had no other choice,
hardware constraints being the biggest one.

I started out on systems where programs were input by way of a card reader.
I much prefer using a keyboard and editor vs a punch-key machine and the quirks of a jammed card reader.
I also started programming in assembly lang,
but now I much prefer to use C or perl or other things that make it easier to program.
And the same with practically everything else on a computer.
Yes the trade offs for ease of use is sometimes a loss of efficiency.

If you can make a program, small, efficient and fast then do so.
If you do a good enough job others may see it and try and emulate it.
_________________
Asus m5a99fx, FX 8320 - amd64-multilib, 3.15.9-zen, glibc-2.19, gcc-4.9.2, eudev
xorg-server-1.16, openbox w/lxpanel, nouveau, oss4(2011)
Back to top
View user's profile Send private message
krinn
Advocate
Advocate


Joined: 02 May 2003
Posts: 4647

PostPosted: Sat Mar 29, 2014 11:28 am    Post subject: Reply with quote

You cannot really compare the value of 1M of ram for a CISC and RISC cpu.
Nor you could even compare the value of 4M of ram for a RISC running 32bits and the same one running 64bits.

For the amiga, 1M was really more than a decent value of installed ram. You cannot expect todays cpu to do the same with only 1M as everything is higher... But by that time, it was twice what other computers have generally.

If as of today Windows need 8M (wow don't jump on me, i don't know, it's pure supposition), you could expect a user with 16M do wonder with Windows too, even if that would just mean 8M of cache.
I think efficiency and optimization are down because of complexity, you can have a guy review and optimize 1k of code, it's not sane to ask him to do that on 100k lines. CPU gets more complex and powerful, code grow in complexity, but human ability remain the same. Must be why human use helpers now, like apple code review or gcc optimization of the code for you.
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Sat Mar 29, 2014 2:05 pm    Post subject: Reply with quote

Dominique_71 wrote:
... Because our whole society is about is about competition, which is a waste of resources at the first place? And we must reproduce it into GNU/Linux? I don't buy it.


This is so extremely wrong on so many levels I can't believe you said it.

Without competition, how can you know if your solution is efficient or not?

At the simplest level, if there is only one known solution, of course that one is the most efficient, and the concept of efficiency is irrelevant.

Get into it a bit more, let's say there are 10 solutions to a problem. Not all perform equally, and while efficiency might be a good place to start for some use cases, reliability or speed might be most important in another case.

Let's use window managers as an example since we've already been playing in that sandbox. Gnome and KDE are two of the most popular window managers out there because they have all the features, but I like keyboard-driven everything and care not one bit about fancy widgets, so I've gone to fvwm. It's good enough for what I need to do, it has all the features I want and it's very light weight by modern standards. You like your Amiga-based setup. More power to you. You and I are obviously atypical.

When it comes down to it, efficiency DOES matter, but only to a point. You have to weigh efficiency against reliability and cost and a host of other variables.
Back to top
View user's profile Send private message
Dominique_71
Veteran
Veteran


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

PostPosted: Sat Mar 29, 2014 4:51 pm    Post subject: Reply with quote

For me, the problem begun at the hardware level. The processors have grown up in power and complexity, and with the multi-core machines, that hide a very bad resources utilization already at the hardware level. At the same time, the compilers have been greatly improved, and it is possible to make a better optimization of the assembler code using C than using direct assembler coding in most cases, and that even for processors like DSP.

For calculations, the DSP algorithms have proved to be much more efficients than any general core, even with DSP instructions. A DSP architecture allow parallel operations and an efficiency which are not possible with the architecture of a generic core.

That imply, in order to combine the advantages of all these technologies, it should be possible to design mixed-cores processors with one low power and low frequency core for things like stand-by, resume to disk/RAM, and so on, one dedicated core to handle the hardware, on generic core for the GUI, and one DSP core for the calculations. And I really hope that will be a trend for the coming years, because it will make possible to combine energy saving with efficiency and the increased complexity already at the hardware level, resulting into a paradigm shift from efficiency or complexity, to efficiency and complexity.

It it possible to make a high end tower PC that use less than 1 watt of power in stand-by, but not with the actual generations of processors and power supplies. The high-end TV sets are doing it from years, and their complexity is not less than the complexity of a PC. And they can do so because they made other technical choices with specific stand-by processors and supplies, as well than other design choices like that.

That will of course imply to do some work on the compilers and the applications. Maybe that move will be driven by the mobile market because due to the hardware limitation of current multi-cores, they are reaching the limit of what is possible to archive. Some work in that direction is already made by the OpenCL community, but to reach its full potential, new hybrid processors have to be developed and put in production.
_________________
[[[ 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
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3436
Location: The Peanut Gallery

PostPosted: Sat Mar 29, 2014 6:21 pm    Post subject: Reply with quote

1clue wrote:
Dominique_71 wrote:
... Because our whole society is about is about competition, which is a waste of resources at the first place? And we must reproduce it into GNU/Linux? I don't buy it.

This is so extremely wrong on so many levels I can't believe you said it.

No it's not. If you truly think we don't waste resources on a massive scale, despite the hegemony of "competition" then I can't believe you're that naive.
Quote:
Without competition, how can you know if your solution is efficient or not?

At the simplest level, if there is only one known solution, of course that one is the most efficient, and the concept of efficiency is irrelevant.

Get into it a bit more, let's say there are 10 solutions to a problem. Not all perform equally, and while efficiency might be a good place to start for some use cases, reliability or speed might be most important in another case.

You're conflating competition with comparison, or evaluation. We're not in competition with each other: we're in collaboration, always have been. "Competition" is just a pretext used to keep us divided and thus conquered, along with insinuations about "crabs in a bucket", and other propaganda so we don't question the fact that the wealthiest in our society collaborate with each other to defraud us all.

The conceptual problem is the confusion of the "market" with "the world", combined with a real paucity of true education, as opposed to conditioning. The goals are quite different to the confused ideas though, and very clear thinking lies behind them. It's just not discussed, as to do so would be to reveal the "realpolitik" which is just another word to hide the truth behind. Cognitive dissonance means people would rather resort to the lazy thinking of prejudice and received "wisdom", than accept that we live in a pervasive kleptocracy.

Which is why USian "IT people" tend to get very uncomfortable with the idea that all modern computing derives from the work of Noam Chomsky, ime. "Apolitical" just means "abdicated self-determination" but geeks are typically uncomfortable with even usual social interaction, let alone standing up to emotional manipulation and conditioning. Lots of talk about "individuality" from people who do their damnedest to conform is the paradoxical, and sometimes amusing, result.
Quote:
When it comes down to it, efficiency DOES matter, but only to a point. You have to weigh efficiency against reliability and cost and a host of other variables.

Well if your software doesn't perform correctly, the efficiency doesn't matter: it's borked. Once it's doing the correct thing, reliably, then you can start to talk about efficiency (where you likely have surprises coming.)

At least, if you're expecting to do this professionally, that's how you should approach it. We do seem to get a lot of muddle-headed ideas presented nowadays though; democratisation of computing is good on so many levels, but not for overall quality if the basics are not well understood. It's like the democratisation of medicine or plumbing; no offence but for all the books out there, I'd still rather employ someone qualified, with a track record of reliable implementation. And no way would I employ an electrician on the basis of "shiny-shiny."

I take my computing just as seriously; I have to, as I work in a legislative society.
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Sat Mar 29, 2014 7:48 pm    Post subject: Reply with quote

Wow, that's quite the response!

steveL wrote:
1clue wrote:
Dominique_71 wrote:
... Because our whole society is about is about competition, which is a waste of resources at the first place? And we must reproduce it into GNU/Linux? I don't buy it.

This is so extremely wrong on so many levels I can't believe you said it.

No it's not. If you truly think we don't waste resources on a massive scale, despite the hegemony of "competition" then I can't believe you're that naive.


Yes, it is. There ARE resources wasted on a massive scale, but competition is not the cause of it. Competition means that the job must be done correctly, on time, under budget or somebody else gets the next job. If the U.S. government had competition for its space program then the same amount of resources (money, materials, etc) would have gotten 10 times as many rockets in the air.

Quote:

Quote:
Without competition, how can you know if your solution is efficient or not?

At the simplest level, if there is only one known solution, of course that one is the most efficient, and the concept of efficiency is irrelevant.

Get into it a bit more, let's say there are 10 solutions to a problem. Not all perform equally, and while efficiency might be a good place to start for some use cases, reliability or speed might be most important in another case.

You're conflating competition with comparison, or evaluation. We're not in competition with each other: we're in collaboration, always have been. "Competition" is just a pretext used to keep us divided and thus conquered, along with insinuations about "crabs in a bucket", and other propaganda so we don't question the fact that the wealthiest in our society collaborate with each other to defraud us all.


Comparison is a small part of competition, where the potential buyer evaluates available options and then chooses one. Are you saying Ford and Chevrolet are not competition, when a customer walks on the lot, compares two items and chooses, and one side gets money and the other side does not? You're using some pretty big words but if you think that's not competition then you need to learn the meanings of these words before you use them.

Collaboration happens when a group of interested parties (a potential market) gets together and defines an open standard for a product, and then when manufacturers have products which meet those standards they choose between the competing products. For example, motherboards made to a specific form factor have standard dimensions, placement for mounting holes, location and type of power connectors, etc.

Likewise in software you have standards. A system logger perform a specific set of tasks. If somebody doesn't like the available options they may choose to write something else, and as long as it performs all the necessary tasks others may choose to adopt the new product. Or, <drum roll...> an init system!

If you're thinking that Open Source projects are not in competition with one another, just start looking at old projects which have been abandoned because something better came along. If you think market influence makes no difference here, then look at systemd and its obviously inferior design, taking over through brute force propaganda.

Quote:

The conceptual problem is the confusion of the "market" with "the world", combined with a real paucity of true education, as opposed to conditioning. The goals are quite different to the confused ideas though, and very clear thinking lies behind them. It's just not discussed, as to do so would be to reveal the "realpolitik" which is just another word to hide the truth behind. Cognitive dissonance means people would rather resort to the lazy thinking of prejudice and received "wisdom", than accept that we live in a pervasive kleptocracy.


Let's at least try to stay on topic.

Quote:

Which is why USian "IT people" tend to get very uncomfortable with the idea that all modern computing derives from the work of Noam Chomsky, ime. "Apolitical" just means "abdicated self-determination" but geeks are typically uncomfortable with even usual social interaction, let alone standing up to emotional manipulation and conditioning. Lots of talk about "individuality" from people who do their damnedest to conform is the paradoxical, and sometimes amusing, result.


I don't know what they teach now, but Noam Chomsky was credited when I was taking my compilers class.

Quote:

Quote:
When it comes down to it, efficiency DOES matter, but only to a point. You have to weigh efficiency against reliability and cost and a host of other variables.

Well if your software doesn't perform correctly, the efficiency doesn't matter: it's borked. Once it's doing the correct thing, reliably, then you can start to talk about efficiency (where you likely have surprises coming.)

At least, if you're expecting to do this professionally, that's how you should approach it. We do seem to get a lot of muddle-headed ideas presented nowadays though; democratisation of computing is good on so many levels, but not for overall quality if the basics are not well understood. It's like the democratisation of medicine or plumbing; no offence but for all the books out there, I'd still rather employ someone qualified, with a track record of reliable implementation. And no way would I employ an electrician on the basis of "shiny-shiny."

I take my computing just as seriously; I have to, as I work in a legislative society.


I AM a professional programmer. Here's what we've discovered:

You code for efficiency and spend a long time doing it, and then a crunch time happens. You slam it together. After the crunch time, you run diagnostic tools on your code to find bottlenecks and realize that the carefully crafted code is no less likely to have a bottleneck than the slammed in code.

You decide to slam it all together, get it working and then use benchmarks to figure out what to improve, and you get a lot more done, more effectively.

Is the slammed in code less efficient than the carefully crafted code? You could argue both ways, only in terms of economics the carefully crafted code is less efficient because it cost more to make it for no measurable result.

Addendum:
I find it much better to code for easy readability and modularity than for efficiency. Easy maintainability is definitely something that reduces costs in the long run. Modern system architectures are hard to predict. Back when I was in college computers had a single processor with a single core, there were only a handful of CPUs out there and you pretty much coded for the target hardware and the target operating system. You COULD optimize for speed and get good results.

Now you don't know how many cores, how many sockets, whether hyperthreading exists, and you don't even know if it's cisc or risc. You can afford to assume multiple cores because even phones have multiple cores now. As you mentioned, modern compilers are better at optimizing than humans are, and to some extent external tools are helping out too. The Java VM optimizes the VM in such a way that common current practices are made to be faster. I'm sure that other platforms have similar optimization where possible.

So coding for easy maintainability and easy testing is much more important than the first-pass optimization.[/i]
Back to top
View user's profile Send private message
ArneBab
Guru
Guru


Joined: 24 Jan 2006
Posts: 380
Location: Graben-Neudorf, Germany

PostPosted: Thu Apr 03, 2014 10:30 am    Post subject: Reply with quote

steveL wrote:

You're conflating competition with comparison, or evaluation. We're not in competition with each other: we're in collaboration, always have been. "Competition" is just a pretext used to keep us divided and thus conquered, along with insinuations about "crabs in a bucket", and other propaganda so we don't question the fact that the wealthiest in our society collaborate with each other to defraud us all.

The conceptual problem is the confusion of the "market" with "the world", combined with a real paucity of true education, as opposed to conditioning. The goals are quite different to the confused ideas though, and very clear thinking lies behind them. It's just not discussed, as to do so would be to reveal the "realpolitik" which is just another word to hide the truth behind. Cognitive dissonance means people would rather resort to the lazy thinking of prejudice and received "wisdom", than accept that we live in a pervasive kleptocracy.

I like seeing how the question about systemd really gets the discussion to core questions instead of just hovering at shallow details.

And we can actually pull this back to systemd: With systemd, collaboration is between hard and impossible, because it imposes its own standards on everything it touches. If a new system comes along, it has to compete with systemd at every level.

With a nicely separated system, someone with a new idea for some subtask can come along and can get help from distributions and other developers, because doing that task more efficiently does not break other stuff. The developers of the old tool might actually go to the new one and say “that’s a cool idea - we’ll help you on the corner-cases of our API if you need to”. They might actually join the new one if the tool has enough value for them, because it’s likely that they are no longer working on the old tool (after all it does the job it was built for) but have switched to using it in their projects.
_________________
Being unpolitical means being political without realizing it. - Arne Babenhauserheide ( http://draketo.de )

pkgcore: So fast that it feels unreal - by doing only what is needed.
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Thu Apr 03, 2014 3:39 pm    Post subject: Reply with quote

ArneBab wrote:

I like seeing how the question about systemd really gets the discussion to core questions instead of just hovering at shallow details.

And we can actually pull this back to systemd: With systemd, collaboration is between hard and impossible, because it imposes its own standards on everything it touches. If a new system comes along, it has to compete with systemd at every level.


No, it doesn't. All the things systemd does is fine conceptually, but there's no reason it couldn't be multiple projects which COULD collaborate with other projects which might be present on the system.

The ENTIRE hoo haa as far as I'm concerned is that all this stuff is glommed into an init system. Systemd is poaching in other territory. More than half of the systems I deal with neither need nor want what systemd has to offer. I generally want an extremely simple, small system that does one thing. Often with a limited amount of space.

Quote:

With a nicely separated system, someone with a new idea for some subtask can come along and can get help from distributions and other developers, because doing that task more efficiently does not break other stuff. The developers of the old tool might actually go to the new one and say “that’s a cool idea - we’ll help you on the corner-cases of our API if you need to”. They might actually join the new one if the tool has enough value for them, because it’s likely that they are no longer working on the old tool (after all it does the job it was built for) but have switched to using it in their projects.


There's no reason why anyone should have to code for systemd. An init system is an init system, they should be interchangeable with respect to anything else not dealing specifically with init systems.

If others want to improve systemd then by all means go ahead. Just don't tell me what I have to run on my hardware. I'll choose what's appropriate based on the tasks the system needs to perform.
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 2588
Location: Dallas area

PostPosted: Thu Apr 03, 2014 3:51 pm    Post subject: Reply with quote

1clue wrote:
All the things systemd does is fine conceptually, but there's no reason it couldn't be multiple projects which COULD collaborate with other projects which might be present on the system.

The ENTIRE hoo haa as far as I'm concerned is that all this stuff is glommed into an init system.


I agree as a general rule and made similar points in a post there.

Init should be simple. It should start something like openrc which should start other processes.
The only other things that init needs to do is watch for runlevel changes, including shutdown, and harvest and dispose of orphans.

Init should not be concerned with cgroups, logging (I might not even want or need logs), logind (desktop logins), etc.
But I also pointed out that the reason that sysd isn't designed that way is that they do not want people to swap out pieces.
They want everyone to take the whole package. They are aiming to be a windows clone, which is all you will ever see.
And if you want something done on your system then you must do it their way.

If people could swap out cgroup manager, or login component, etc then there would be no lockin to their way of doing things.
_________________
Asus m5a99fx, FX 8320 - amd64-multilib, 3.15.9-zen, glibc-2.19, gcc-4.9.2, eudev
xorg-server-1.16, openbox w/lxpanel, nouveau, oss4(2011)
Back to top
View user's profile Send private message
steveL
Advocate
Advocate


Joined: 13 Sep 2006
Posts: 3436
Location: The Peanut Gallery

PostPosted: Thu Apr 03, 2014 9:46 pm    Post subject: Reply with quote

1clue wrote:
steveL wrote:
If you truly think we don't waste resources on a massive scale, despite the hegemony of "competition" then I can't believe you're that naive.

There ARE resources wasted on a massive scale, but competition is not the cause of it. Competition means that the job must be done correctly, on time, under budget or somebody else gets the next job. If the U.S. government had competition for its space program then the same amount of resources (money, materials, etc) would have gotten 10 times as many rockets in the air.

Perhaps you could consider the duplication of effort which that "competition" results in, and the insanity of forced competition between people who actually want to collaborate, like nurses or teachers. Further, as I said, we've only ever achieved anything by collaboration, not through competition. If we weren't kept divided and conquered, we'd both do a lot better and waste a lot less.

Competition has not addressed the "waste of resources on a massive scale": in fact it fosters it through massive duplication of effort, as parties try to obscure the background knowledge shared by all that led to their "innovation", and you end up with time, effort and resources spent on trying to bring the other down, instead of on actual progress for everyone. Economies of scale apparently only work in the "private"-sector, which just happens to be massively subsidised by the rest of us; the so-called "free market warriors" are the biggest welfare recipients of all.

As for the space program, istr a quote from one of the first astronauts, asked what he was thinking, replying how reassured he felt that the components used in the craft had all been bought at the cheapest possible price.
Quote:
Quote:
The conceptual problem is the confusion of the "market" with "the world", combined with a real paucity of true education, as opposed to conditioning. The goals are quite different to the confused ideas though, and very clear thinking lies behind them. It's just not discussed, as to do so would be to reveal the "realpolitik" which is just another word to hide the truth behind. Cognitive dissonance means people would rather resort to the lazy thinking of prejudice and received "wisdom", than accept that we live in a pervasive kleptocracy.

Let's at least try to stay on topic.

Oh please, that is the topic you brought up. That's exactly how competition is used, and for what: to keep us fighting each other while the wealthy collaborate to rip us off. I noticed you skipped past those links: you should definitely read the latter, before you claim to understand how things work. We live in a kleptocracy. Period.

I know that's hard to accept, but it is the truth.

Yes, competition for resource is very much part of our nature, so indeed part of our psychology. It hasn't exactly led us to very good results though, has it? We're kept in competition with each other, and our baser natures are stimulated so we don't even think clearly. But meanwhile war is inevitable, and so is poverty, and since life isn't fair, we don't have the choice to be either. Or something.

In fact everything we've ever achieved, we've achieved via collaboration. Perhaps collaboration forced by the struggle for survival, but collaboration nonetheless. It's only when we work together that we make any sort of progress.

And much of the competition was totally unnecessary, forced upon us to satisfy the ego of what are at best sociopaths.

What really gets me is that the competition which does work, ie friendly rivalry, is subsumed into this insane variant where we have to do each other down.
Back to top
View user's profile Send private message
1clue
Veteran
Veteran


Joined: 05 Feb 2006
Posts: 1484

PostPosted: Thu Apr 03, 2014 11:32 pm    Post subject: Reply with quote

Maybe a friendly moderator would split off this discussion between SteveL and me and put it in a more appropriate section?

This is the sort of discussion I could spend months on, but it has absolutely nothing to do with the systemd topic, and FWIW I still think this is the most rational gentoo+systemd discussion to date.

Call it something like "Competition vs Collaboration"
Back to top
View user's profile Send private message
depontius
Advocate
Advocate


Joined: 05 May 2004
Posts: 2809

PostPosted: Thu Apr 03, 2014 11:58 pm    Post subject: Reply with quote

Before splitting off, and since this is getting (too?) philosophical, I'll have to put in a point for competition, and use the free market economy as an analogy.

Nobody has a monopoly on doing it right. Even if someone is right this time, or N number of times in a row, that's no guarantee that they will be right next time.

The real point (IMHO) behind free market economics is just this. Nobody is guaranteed to be right, next time. You may guess which will win, but you never know. The economy that has viable alternatives capable of emerging for any given marketplace wins, in the long run. Many use this as an argument against central planning by the government, or indeed any interference by the government. Personally, I believe it's an argument against central planning by anyone, including dominant marketplace players or coalitions thereof.

Bringing it back to the topic... If we really believed competition was a waste of time and cooperation was the one true way, we'd all get behind systemd. Some of us "systemd nattering naybobs of negativity" happen to believe it's going down the wrong path, and taking everyone it possibly can with it. Personally, I believe within the next few years systemd will either become moribund under its own weight, complexity, and entangledness or be the attack point for a security nightmare, or both. I don't say "SysV is the one true way," I merely believe that diversity must be kept viable.

Competition should be more about finding a better path than about duking it out for market share. At work I've been both CAD support and leading-edge CAD user. Better tools win over users. From what I've seen, forcing users into a CAD tool seldom makes things better.
_________________
.sigs waste space and bandwidth
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 ... 15, 16, 17, 18  Next
Page 16 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