On 04/20/2014 04:00 AM, Dirk Gently wrote:
Extremely poorly documented. I've used Unix since 1983 and administrated it since 1993... and I don't know a single person who can make heads or tails of how to
Actually systemd is much better documented than sysvinit, much better documented than many other key UNIX/Linux facilities such as xinetd, cups and more. If you don't think so then you probably haven't bothered reading said documentation. What makes systemd documentation different is that rather than just saying what it does (the man pages) the authors have produced all the "design documentation", descriptions of what they planned, how they went about it it, their failures as well as successes. The whole philosophy behind systemd is laid open. Along the way they go into such things as the original motivations; the shortcomings in sysvinit that were impeding the growth and success of Linux; all the things that sysvinit doesn't do that a an initialization and process management system needs to do; all the performance metrics -- something we don't see for sysvinit; how-to documents on creating unit descriptors and creating generators and drivers. Its all out there and its all easily accessible and its thorough and puts many other FOSS projects to shame. The FOSS world has plenty of 'code moneys' of varying skills but few people who bother documenting their work as well as Lennart Poettering and the systemd developers. Another branch of this thread makes and issue of signature blocks and has one person pointing out that personal attacks are not permitted. If I were you, the-person-hiding-behind-the-alias-of-Dirk, I'd refrain from attacking Lennart.
SysV init was simple --- it forks off scripts.
Indeed, and some of those scripts are obscure and roccoco and they have very little documentation. While, I'm sure that the packagers of a distribution try to keep the script from getting too diverse, being scripts, they can do pretty much what they want. There is nothing to kill off rogue scripts or constrain what they can do. They all run as root, remember, and there are many opinions for 'errors and omissions', especially when custom hand-crafting ones. All is not as hunky-dory in sysvinit-land.
If you can't even understand how your system boots up, then you don't really have control of your system in the slightest.
I'm sure millions of people using Pcs and phone and tablets will be delighted to hear that. Oh, and lets add in people who fly planes, drive cars, take photographs. As Alan Cooper pointed out in his book "The inmates are running the asylum", its not a camera, its a computer with a lens; its not car, its a computer with wheels. None of these people understand how the computer that runs their ... Whatever ... Works. The reality is that computerization has often given users more control, more detailed control over their tools and appliances that the old, 'mechanical' method. That's certainly true of my camera! Its amazingly so true of my phone!
And worst of all, it has ill-defined, and constantly moving boundaries.
Ah yes, just like almost every other supported software product that has active and involved developers dealing with user feedback. What all this comes down to is this: This is FOSS. We pay a price for that. The well financed, proprietary developers and do their development and testing in-house and with selected external alpha- and -beta-testers. This means that when they release to the public they have a high degree of confidence in the product. With FOSS the developers can only take it so far then they 'release' to the community. Some of us, possibly many of us, have to do the alpha and beta testing. We're not paid for that. This is FOSS. Not so long ago we went though this with KDE-4. The initial release was terrible but the bug reports and feedback allowed the developers to make progress. Some of their ideas were depreciated, others transformed. We now have a great product, but we paid for that along the way. Also, along the way, we had the people who weren't willing to pay, who didn't want to hang in there, who didn't want to submit the bug reports for the developers to work with, and they stuck with KDE-3. They too pay a price, the effort of setting up their systems to use KDE-3 rather than the defaults of KDE-4, shifting repositories and packages and doing without many features. Just as with KDE-4, just as with the Microsoft supporters who decried Linux, we are saying and will say to them "Quite possibly you had a valid argument once, but we've addressed that, moved on. You're living in the past, arguing about things that no longer exist". -- Where we cannot invent, we may at least improve. -- Charles Caleb Colton -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org