Am Fri, 10 Aug 2012 16:33:39 -0400 schrieb Brandon Watkins <bwat47@gmail.com>:
Systemd and pulseaudio are completely different pieces of software with different purposes. Comparing them like that just because of the author is comparing apples to oranges.
Sorry, it is not. I see that PA is totally not complete and doesn't support at least half of the professional use cases. And I see that it's the same with systemd. So what's the difference? They are both developed by the same person who seemingly doesn't have much knowledge about professional computer usage and only cares about some desktop users. With PA it's currently not such a problem since I don't need to use a distro or a desktop environment which forces me to install PA. With systemd it's worse since the init system is a very serious and important piece of the system. And if this doesn't support every professional use case and isn't proved to be really reliable, it just shouldn't be made to a de facto standard. And if I can't trust PA how can I trust an even more important piece of software written by the same person? Btw., look at systemd-cryptsetup. Yes, meanwhile my use case is filed upstream and allegedly and hopefully fixed. But it shows that at least one use case was just forgotten or in other words it was not well enough thought out. The latter is the biggest problem. Like I said before, some of Lennart's ideas may, say, seem to be quite interesting, and maybe sysvinit is also not the perfect init system. But Lennart's software is just not implemented good enough. If somebody doesn't care about the professional users when writing on software, would he really care about the professional users when writing the other software? I really haven't seen so many and so long discussions and so many concerns and very negative opinions about a software than I have seen about Lennart's software. And I'm not only reading this mailing list. See e.g. pro-linux.de or heise.de (both in German). Every time when there's an article about PA or systemd a lot of people are railing against PA, systemd and Lennart. And it's definitely not only me. There must be a reason, and the reasons are always mentioned. There are bug reports upstream, but they are just ignored. Lennart mentions all those "rants" in at least one of his documentations. So he even knows about all those criticisms. What's he doing? He ignores them totally. In the same sentence he just laughs at those people, and call them so to say (not literally) stupid. Is this really a good and trustworthy attitude? I think, not. And all those comments here like "oh no, not this again", "Please guys, not again..." or "Take your concerns upstream, ...", is really not helpful. On the contrary this all is also an issue for downstream. See the ongoing infiltration of initscripts by systemd here in Arch Linux. Sorry to say that, but it's really not the best idea. Keep PA and systemd totally optional including every part of it, and everything is Ok. I'm sure nobody would mind. But as long as there are people working on making both software a de facto standard and forcing it on everybody, this discussion will never end. Not only here. Just take all those people who have a lot of concerns for some very good reasons serious. There wouldn't be so many, so long discussions every time PA, systemd or Lennart Poettering is mentioned if this all was such a very good, perfect and professional software. If this was the case then I'm sure that everybody couldn't wait to get it and a lot of people would ask when it will be available. Instead a lot of people on the web rail against them. So think about that, and think long and good. Maybe there are a few use cases for which PA is working and for which PA makes sense. But there are a lot of use cases for which PA does not work. The same for systemd. So think about the use cases for which they don't work. Btw., someone else here on this mailing list has mentioned a lot of software which, as he said, do the same as systemd does allegedly better than sysvinit, but on top of sysvinit and in a more UNIX like way. There came not even one word, one short discussion about those suggestions. It was not considered if those software could be the better alternative. Instead the systemd fanboys kept on hyping systemd. If you buy a book at Amazon e.g., what do you read? Only the best 5-star reviews or also the 1-star reviews? I tell you something. Not always but a lot of times the fewer 1-star reviews are the better and more realistic ones. Heiko