On 9/3/2012 5:34 PM, Bryen M Yunashko wrote:
On Mon, 2012-09-03 at 22:47 +0200, Dimstar / Dominique Leuenberger wrote:
On Mon, 2012-09-03 at 16:24 -0400, Felix Miata wrote:
When decisions to change important paradigms are made it's generally known in advance to be change that requires a disproportionate amount of resources to implement. Complaining about "limited" resources is out of place - the problem was expected at the outset.
The fact that systemd becomes a rather big defacto standard was also to be expected... we could have missed it entirely.. and be way out of the loop now
New does not automatically equate to better. Too often opposite turns out to be the case when hindsight proves implementation turned out to be premature.
Not having resources to invest, and pulling at two ends, surely does not help the situation, does it?
Dominique
I find the arguments somewhat curious here on at least a couple of fronts.
Resources: I find the argument against systemd because of limited resources to be tenuous at best. Linux and open source in general is known to be a collaborative environment with shared resources across the board. We collaborate and she share fixes that benefit upstream as well as mutual distros. We learn from each other and we implement the lessons we learn. If we stay with sysinitv, then clearly we would end up with less resources than what we have now. It stands to reason that the definition of "resources" in an open source world is global rather than internal. So how would staying with sysinitv ensure that we would be better off?
Number of people concerned: I see the argument that openSUSE is simply not ready for systemd. Mainly I see it from a handful of people and while I do not dispute the legitimacy of your concerns, why aren't the concerns more widespread when there are thousands of openSUSE users and servers out there? One argument against this group has basically been "because you don't know how to use systemd." Is this a fair defense in favor of systemd? Is there a difficult learning curve that needs to be addressed for a few people? Are there legitimately bugs that can have widespread effect? Are the bugs more corner-case? What are you doing differently from the rest of the world, so to speak?
I don't know, but the fact that so many thousands are using and appreciating openSUSE with systemd speaks volumes from an observational standpoint here. And I think that if the sysinitv holdouts want to win their case, they have to frame it in the context of how it is/could be negatively affecting so many people rather than how it affects just them and their unique implementation of openSUSE.
Considering the massive adoption rate of systemd across the board, I would assume that systemd being dysfunctional would mean there would correspondingly be a mass defection from Linux in general, and I'm not seeing that anywhere.
Bryen M Yunashko openSUSE Project
The main problem is using popularity and reported problems as meaningful or valid info. Systemd probably works fine for many people in most cases. That doesn't mean it's not a downgrade for many, where they gain systemd features they don't care about, but lose features that are less "cool" but more critical to actual work. Unix isn't supposed to be a limited platform that only supports the common case. Many of the people who will have the worst problems are simply avoiding it still, so there are no reports of incompatibilities from them. They have better things to do, they have work to do, so don't make their lack of trying and reporting their fault either. Many of the people who will have the worst problems may simply consider themselves forced to waste money, equipment, and time in order to stay compatible with whatever the OS does, so no compatibility problem report from them either. Some doctors office doesn't know they didn't really need to replace their server because it had some odd hardware, or replace some commercial proprietary no-longer-supported software, or replace some $3000 appliance. Their computer guy said so so they swallowed it. No report from them either. Unless the default systemd implementation is really a drop-in replacement, where packages with init scripts install with no problem, and commands all work the same way, and booting and syslog and dmesg and the console including the serial console are all left working exactly the same as before I will have a problem. Maybe not large, maybe large. I'll deal with it, but don't expect thanks for making me have to, because I haven't seen one systemd feature that I give a crap about, since my servers are not cars, but I will have to spend effort to deal with it, so it's a net loss for me. The real fix would be if the systemd authors just weren't such inconsiderate dicks that insist working with systemd has to come at the expense of working any other way, instead of being flexible to work however the user requires. Left and right they are telling everyone else, who predates them by decades, that they are all doing things wrong and need to change. I vastly prefer the way an init script can do _whatever_ I need it to do, instead of a faster and more easily managed systemd config file that can only do that which systemd happens to provide an explicit feature for. If systemd vs sysv init were like xinted vs inetd, I would have no problem. -- bkw -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org