Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Comment: Re:Discussion is outdated (Score 4, Insightful) 460

by geoskd (#48899947) Attached to: Ask Slashdot: Is Pascal Underrated?

to be FAR more true for C than Pascal.

On an instruction cycle basis, there are orders of magnitude more C code executing in the world today that any other language. The worlds single most executed pieces of code are all written in C. People should keep that in mind when they make proclamations about C's demise. C will be with us for many hundreds of years for the same reason COBOL wont die, only far more so.

All android devices run Linux at their core which is C. Further, a large swath of embedded devices that don't use android are written in C anyway. These language popularity counting systems always neglect the jobs that require C, but don't explicitly state that. Jobs like hardware designers that need to provide hardware access code. Virtually nobody writes embedded code in Java, Perl, Pascal, COBOL, or any of hundreds of other languages. The few that have made some inroads into embedded systems are rare, and have not made much progress. If you want to understand the Internet of things, you'll only ever get halfway there if you don't know C.

Comment: Re:I agree with Lennart (Score 1) 551

by geoskd (#48838021) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

Why would LibreOffice or GIMP ever be dependent on systemd? They have nothing to do with the startup or shutdown of the system - they are plain vanilla applications (same most likely goes for JBoss and KDE, though they may provide some 'system-like' services). Seriously, folks. It's just this kind of hyperbole from systemd haters that makes me think it must be good...

I wondered the same thing, and both Gimp and LibreOffice work just fine without Systemd. They both make use of Systemd features if its available, but they do not require it.

Comment: Re:Ohreally (Score 1) 551

by geoskd (#48837999) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

Sometimes they can, sometimes they can't, this has been addressed elsewhere in the thread so you should not be bringing it up here.

I have read (and posted) this thread extensively, and There are no truthful examples of a portion of Systemd that cant be replaced except udev which has to be an integral part of the init system.

There are several examples in this thread, and there have been several examples in prior threads, and there are several examples in various blog entries which you could find with google if you knew how to internet, bro.

Every time I ask for examples, I get a response just like yours: "Look around, examples are everywhere!". They are not. Even in this thread, there are no examples whatsoever. The closest I have seen to examples are either lies, or half truths, like the statement I found indicating that Systemd replaces ntpd (it does not, but it does offer an alternative to ntpd, which is not required). I have only been able to conclude that people are opposed to Systemd for personal reasons which I can only guess at the reasons for. Those reasons seem to have no technical foundation, but mostly appear to stem from a long working understanding of xyz init system, and a healthy fear of the new and unknown.

My personal suspicion is that there are a lot of Sysadmins out there who have been working with sysvinit like systems for a very long time, and are comfortable with the various scripting languages involved, but with Systemd being written in C, it scares them to think they would have to understand C in order to debug their boot-up problems in the future. I can fully understand this particular bit of paranoia, but it is simply paranoia, nothing else.

No, balls is taking away someone's lunch and replacing it with a shit sandwich,

TANSTAAFL. Linux doesn't belong to you. You haven't paid for it, you didn't build it, so don't whine when it isn't what you want. Furthermore, no one is taking anything away from you. You are still free to continue using whatever distro you use today. It is yours as long as you like, but no one owes you free updates, in fact no one owes you a damn thing, so get over your self righteousness, because you haven't the right.

Comment: Re:The Dangers of the World (Score 1) 779

by geoskd (#48835239) Attached to: Parents Investigated For Neglect For Letting Kids Walk Home Alone

Children invariably treat their own parents far worse than strangers. My own kids are no exception. I suspect its because with other adults, they dont know what to expect, so they err on the side of caution. I have the same issues with my own kids. Around strangers, they are mostly polite and obedient, but once its just us, I get a lot more pushback. Generally, I don't get a lot though, as my kids learned at a very early age not to push me too hard. My two are 4 and 6, and I haven't had to so much as slap them on the wrist in over two years.

My wife on the other hand has never even so much as laid a hand on the kids, and she cant get them to obey any commands. The only time she gets any respect from the kids is when they know I'm in the house and she threatens to "get daddy". In public they tend to behave better, and at school, I'm told they are both pictures of perfect behavior.

Other parents we talk to tell pretty much the same story, with whichever parent administers the discipline getting the lions share of the good behavior and the other parent getting $#!7 on regularly.

Comment: Re:Ohreally (Score 1) 551

by geoskd (#48835159) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

This thread is full of evidence of both. Don't be deliberately disingenuous, nobody likes a liar.

This thread is full of the following:

Systemd is bad because it is monolithic. (Systemd may be monolithic, or it might not, but that is completely irrelevant to its usefullness or fitness for use)

Systemd is bad because it replaces X (While Systemd does offer replacements for many other components, the alternatives can continue to be used as before)

Systemd is bad because it might break something (These comments piss me off because no one can give me *any* examples of anything that Systemd breaks, just a bunch of what ifs, and maybes )

In the end, I see a bunch of BS FUD, and nothing of substance from the anti-Systemd crowd. At the end of the day, I can easily see how init was broken, and pottering stepped up and made a solution. The rest of the people who are whining need to either a:) provide a solution of their own, b:) point to an alternate that works or c:) STFU. end of story. Nothing pisse me off more than people who are getting something for *FREE* and are whining that they are not getting what they want. That is the definition of balls my friend.

Comment: Re:The Dangers of the World (Score 1) 779

by geoskd (#48830613) Attached to: Parents Investigated For Neglect For Letting Kids Walk Home Alone

Sadly by this going public I expect CPS to be up in this family's affairs for years to come. They may not actually lose the children, but that won't negate the pain and frustration of having to deal with CPS with "if you don't do this, we'll take the children" being their goto response for any question.

I'm afraid you're right. I understand where the concept of CPS comes from, I have to state categorically that the cure is worse than the disease. I find CPS to be unqualified arrogant assholes who have seen so much of the worst of society, they assume its there in every visit, and if they cant find it, they just haven't looked hard enough. I have not had personal run-ins with CPS, but I know people who have, and that "save-the-children" attitude is a large part of the reason why we see rising obesity, falling competitiveness of our children in the academic world, and the abysmal work ethic of the why-bother generation. When the only safe thing for a parent to let their kids do is rot in front of the TV, or play video games all day, its small wonder that kids end up the way they do these days. The whole thing is a fundamental failing of our leadership on both sides of the aisle. Its time we take our country back from the save-the-childrens brigade.

Comment: Re:Just keep it away from Gentoo and I'm good (Score 2) 551

by geoskd (#48830351) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

So you know nothing about them, but you know systemd does everything better in every conceivable way. Such as binary logging that you probably not be able read if you have to boot from disk. The scope of what you understand and what you think the system needs to be able to do is limited, and therefore your capability to answer reasonably is as well.

Interesting assumptions. What I do know about Systemd vs upstart vs sysvinit is the algorithm that each uses to solve the same basic problem. As someone who can understand algorithms, and who understands the basic problem that needs to be solved, I can tell you that the algorithm that Pottering (and others) claim that Systemd uses is smarter, faster and conceptually better in every conceivable way to upstart and sysvinit. What I dont know is the specifics of how to configure upstart, sysvinit or Systemd, as I have never had to work with any of them before. I have worked with a proprietary init system I was told was a distant cousin of sysvinit, and it was not what I would call a good experience.

Then why did it not handle it before, assuming you mean it has to handle it as journald does now.

because syslogd cannot handle early boot reporting. By definition Systemd is the first thing to run, so any logging related to this time in the boot sequence has to be handled by it. Since it has to do it anyway, make the utility available for general use. Why have another utility duplicate the functionality when Systemd has to have it for early boot time logging.

Why does init have to handle ntpd?

Unless you read something I didn't see, there is no reason in the world it has to replace ntpd, but the option is there if you want it. On a more serious note, I would conjecture that there are daemons that rely on ntpd, that could be started sooner if there was a version of ntpd that could run before the network was up. why should a program that requires accurate time (hence ntpd) wait for the network to come up in a system with a real time clock? a version of ntpd that would come up to a running state without depending on the network being up would save a large amount of time in the boot sequence. Though your system might not care about boot time, there are many that do, and if the new version of ntpd does everything else ntpd does, why not use it. Pottering could have volunteered a patch for ntpd to add the functionality, but he chose to rewrite it. If you use his version, the system boots 3 seconds faster, if you use ntpd it takes 3 seconds longer to boot. In all other ways its the same...

Comment: Re:Ohreally (Score 1) 551

by geoskd (#48829971) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

If he listened to actual users, (old-time sysadmins who know what they're doing) then he'd know systemd is unwanted and dangerous. It's too all-encompassing and introduces way too much risk into the server infrastructure.

First, Systemd is neither unwanted nor dangerous, until and unless you can give me a specific example.

Second, Any new technology does not belong on high reliability servers until they have been thoroughly vetted. No one is putting Systemd into stable releases yet, its still going through the vetting phase.

Third, are you running Upstart? That was a new technology once. It also had to be vetted, but You would be laughed out if you referred to Upstart as unwanted and dangerous. All told, Systemd can claim a great number of actual performance and ease of use improvements over most init systems including Upstart and sysvinit.

Fourth, Neither pottering nor anyone else has the kind of clout it would take to force these distributions to do anything (Possibly excepting Linus). The dastardly way Pottering got all of these distros to switch to Systemd was to present it on its merits! (I can see why that kind of foul play would be wholly unacceptable to you.)

At the end of the day, The anti-Systemd crowd fails to offer any concrete examples, and relies on rhetoric, which is why they are failing to stop the spread of Systemd. Systemd is winning, and quickly, because it is a technically superior solution to a problem that every operating system has to solve. The wonder isnt that Systemd is being adopted so quickly, the wonders are that it took this long for a good solution to appear and that so many people who are otherwise qualified administrators would choose to oppose it. I guess it just goes to show how many people in IT really can't hack it.

Comment: Re:Fork it all (Score 2) 551

by geoskd (#48829743) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

But what has the rabble up in arms with systemd is that that particular "freedom" means basically having to rewrite the entire operating system.

No it doesn't, grab any distribution you like and fork it. Just don't get mad when others refuse to do maintenance work on your fork. If your solution is good enough, you will get enough people interested in helping you to maintain it. The fact that the greater majority of people who actually write the code are switching to Systemd should be a sign that it is technically superior.

If five years from now, as you suggest, Systemd is an integral part of every mainstream distribution, and you wont use any mainstream distribution because of that, then I would suggest the common point of failure is obvious...

Comment: Re:Just keep it away from Gentoo and I'm good (Score 3, Insightful) 551

by geoskd (#48829635) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

Systemd does multiple things and does them poorer then what they replaced, therefore it does not do one thing,

Citation needed. Please elaborate on any things that Systemd does worse than something that it replaced. Specifics would be appreciated.

I know very little about init, or Systemd, but what I do know is the very basic idea of how they both work. Init works, just like the documentation says, by starting system components in the order specified by the init scripts. These scripts are structured such that when a particular component is done starting, it will then trigger the start of another set of components. I can tell you that this is a hideously malformed way to start a system. If I have a configurable system and want to bring up component x, but it requires a list of 50 components are up first, then i need to look through all of that configuration to figure out which components should trigger mine to start up. This is backwards from how it should be.

Systemd works the problem backwards. I tell it what things my component needs to have running first, and Systemd figures out what order to boot things. It is smarter, faster, and conceptually better in every conceivable way. The particular implementation may or may not be good, but the algorithm is far superior, which is why it is being universally adopted. Much of the rest of Systemd is a result of the fact that any init system will have to touch on all of these other areas of the system. Init has to handle logging. It also has to handle system configuration (udev). It will also have to touch upon process management, and resource allocation. It is also the logical place to put any kind of hot swap functionality for devices, as the init system will largely be handling hardware interaction modules, and many of the system component dependencies are hardware dependencies and not just software dependencies. In all the problem is very complex, and I'm glad to see someone actually tackling it intelligently.

As with all things Linux, if you don't like it you're free to present an alternative of your own, but don't get huffy when people choose Systemd over what you recommend, because they are choosing for them. You are quite free to choose for you. In the end, if you don't like it vote with your money, and choose a different distribution, or OS.

Comment: Re:Fork it all (Score 1, Troll) 551

by geoskd (#48829453) Attached to: Systemd's Lennart Poettering: 'We Do Listen To Users'

Because somewhere back down this little road between windows and Linux I got the idea that Linux is about choice and freedom. Crazy concept, I know.

Linux is all about choice. You're just confused about the choice to be made. Linux was never about the choice in what software others make for you. It is instead about the choice to make the software for yourself.

In other words, if you dont like it, you're free to rewrite any part of the software you want. You have the freedom to replace any part of the system with anything you like, but there is no good reason in the world anyone else should have to put the system together for you the way you want.

Comment: Re:Any actual examples? (Score 5, Interesting) 598

by geoskd (#48738047) Attached to: Tumblr Co-Founder: Apple's Software Is In a Nosedive

He makes the claim that their software quality has taken a nosedive, that they're introducing tons of bugs and functional regressions, but he doesn't give a single example of any of that. He just makes the unsubstantiated claim.

across 4 different iPhones and three PC's, I have never once had an iOS update that didnâ(TM)t brick the phone. For the last two, I made the appointment at the Apple store before I attempted the update. The last one was done in the apple store with the whole store watching. Bricked it so badly that the phone had to be replaced under warranty.

Comment: Re:I disagree... (Score 1) 181

by geoskd (#48731391) Attached to: Indiana Court Rules Melted Down Hard Drive Not Destruction of Evidence

I don't know what this guy is playing at because I seriously doubt that this day and age he doesn't know about such services but ...

If you keep backups, like any sane person would, throwing away a failed drive is the normal option.

I had a drive fail once, and didn't have backups. The drive failed in such a way that I was quoted $2,000 for an attempt to recover the data, and no guarantee the procedure would work. I decided I didn't really have any valuable data on that drive after all... If I'd still been working at Kroll, I might've been tempted to try to get one of the lab techs to throw the platters into the test rig, and pull my data, but it just wasn't in the cards...

Comment: Re:Hello microwave (Score 5, Informative) 181

by geoskd (#48731335) Attached to: Indiana Court Rules Melted Down Hard Drive Not Destruction of Evidence

From what I understand, all it takes is one knock with a hammer to ensure platters are bent and can never be used again. And there is no known practical method to restore data after a single overwrite with 0s. Everything else is pure paranoia.

Overwriting with 0s will not perfectly overwrite the tracks. There is some slop on the read head positioning that will normally allow enough data to be recovered that the ECC can be used to rebuild the full data set. This can only be done by taking the platters out and putting them into a machine that is purpose built for the task of recovering such data. Having worked on part of the design of such a machine, I can tell you two things about it: First, it exists, and second, it is not cheap. I was told the cost of the one I worked on was $12M USD.

Physically deforming the platters will make the job much more difficult (expensive), but not impossible. A combination of clever chemistry and an electron microscope can usually recover data off a mangled platter. This process sucks because it has to be done by hand, but if its worth enough to you, there a couple of labs around the country that can handle that level of data recovery. A better way to ensure total destruction is to melt the platters outright. not only does the melting deform the platters, but sufficient heat will thoroughly destroy the data, even if the platter does not completely melt.

If A = B and B = C, then A = C, except where void or prohibited by law. -- Roy Santoro