Stories
Slash Boxes
Comments

SoylentNews is people

posted by LaminatorX on Monday September 01 2014, @09:04PM   Printer-friendly
from the puttering-about dept.

Phoronix has an article up about some interesting ideas of Lennart Poettering about what could be a possible future for Linux:

Lennart Poettering of systemd and PulseAudio fame has published a lengthy blog post that shares his vision for how he wishes to change how Linux software systems are put together to address a wide variety of issues. The Btrfs file-system and systemd play big roles with his new vision. Long story short, Lennart is trying to tackle how Linux distributions and software systems themselves are assembled to improve security, deal with the challenges of upstream software vendors integrating into many different distributions, and "the classic Linux distribution scheme is frequently not what end users want."

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Interesting) by sjames on Tuesday September 02 2014, @12:13AM

    by sjames (2882) on Tuesday September 02 2014, @12:13AM (#88272) Journal

    I got part way through his blog post. I had to pause a moment to go throw up.

    He claims to be concerned for the developer of embedded systems. BULLSHIT! The first step to building an embedded system image will necessarily be to rip out his nomination of an init system. Truly a mouse built to government specs. I need something lightweight that does its bit and gets the hell out of my way. That is not systemd. What I need is exactly something like Debian as it is now where I can carefully choose a minimalist set of packages that support the target application.

    He clearly hasn't a clue how the toolbox model works. It makes the exact version number of various things largely irrelevant. It's how the adventurous can use Debian Sid as a daily OS if they like. It's why you can choose to download one or two packages from upstream and compile them if you need the latest and greatest.

    That doesn't work very well with his kitchen sink dependency model which is why so many object so strenuously to systemd.

    It's also fairly clear that he doesn't run any multi-user systems.

    I strongly recommend he just borrow the Linux kernel as-is and build Poetterix to be exactly what he wants and leave the rest of us out of it. If it's really that great, people will adapt current distros to it. But what will really happen is he will use it and his 3 best friends will pretend to.

    Starting Score:    1  point
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 2) by Marand on Tuesday September 02 2014, @06:49AM

    by Marand (1081) on Tuesday September 02 2014, @06:49AM (#88390) Journal

    That doesn't work very well with his kitchen sink dependency model which is why so many object so strenuously to systemd.

    Well, he is a RedHat employee; maybe he misses the dependency hell of yore? :)

    It's also fairly clear that he doesn't run any multi-user systems.

    It's the repeating history of computing. Multi-user systems are out of fashion currently, so people that don't know any better are following the fad and ripping out multi-user bits while worshiping the gods of "boot time" and "efficiency". Later, when multi-user systems are fashionable again, they'll be scrambling to re-implement all the bits they left out. See also: X to Wayland. Rip out all the multi-user stuff because "nobody needs that" to make local, single-user use faster; realise later that people need that stuff; put it back in and hopefully it will be as usable as what you threw out.

    There are people that seem to think the current trend of excess, where everybody has at least one desktop or laptop, is never going to end, so why bother supporting multiple users on a single system? Just get another system, they're cheap! Who knows, maybe they're right? If so, losing multi-user features doesn't matter. I think it's more likely we move back into the multi-user part of the cycle for some reason. Maybe with people using tablets and laptops as cheap, long-lived thin-clients that are used to interact with powerful home-systems that do the bulk of the work; one home-server per household generally, with resources shared.

    The last part is just random imagination, but I think it's more likely than infinite single-user excess.

    He clearly hasn't a clue how the toolbox model works. It makes the exact version number of various things largely irrelevant. It's how the adventurous can use Debian Sid as a daily OS if they like. It's why you can choose to download one or two packages from upstream and compile them if you need the latest and greatest.

    Off-topic, but my desktop system is a real-life example of your statement. Bulk of the system is Debian testing, with unstable (sid) pinned at lower priority for the purpose of cherry-picking newer packages as needed, and stable also available for the rare instance when I need to go back to something really old. Plus a handful of external debian repos, and even a few Ubuntu PPAs I pull packages from. All these disparate sources, not even from (or for) the same distro, and it's all handled smartly until I try to install something like GNOME or KDE from an external repo.

    • (Score: 0) by Anonymous Coward on Tuesday September 02 2014, @07:43AM

      by Anonymous Coward on Tuesday September 02 2014, @07:43AM (#88401)

      Multi-user systems are out of fashion currently,

      Somewhat ironic, given that at the same time the cloud is all the rage ...

      • (Score: 2) by Marand on Tuesday September 02 2014, @08:17AM

        by Marand (1081) on Tuesday September 02 2014, @08:17AM (#88406) Journal

        Somewhat ironic, given that at the same time the cloud is all the rage ...

        Yeah, I considered going into that but didn't want to stray too far from the point. The cloud thing could be the first push back into the multi-user part of the cycle again, being another step toward a repeat of the thin-client model. If that's the case, it'll be funny if people like Poettering finish gutting the multi-user parts just in time for multi-user designs to become en vogue again.