launchd and systemd

Status
Not open for further replies.
I guess I'm just really bad explaining myself. This area is for feedback I think. I really love FreeBSD and I don't want it to change. This is my worry. Since Debian changed I worry that FreeBSD could easily change into something similar. Only a handful of "distributions" now have init anymore. This is a huge change and it came out of the blue.
 
binary semaphores

Binary is really just short for "binary numeral system" it's nothing but a different way to present numbers. Computers use binary because it's result is always 1 or 0. It's because computers use electric currents. Computers only understand yes or no. You can translate binary, hex, or decimal, very easily using the Windows calculator. It can translate the most used systems within the standard positional numeral systems. Coincidentally they are used in computing science a lot. Converting binary to hex, binary to decimal, hex to binary, hex to decimal, so on. My point was that I do not want to use Linux nor Windows systems at home. I have no time for that. If you are kind enough, you would listen to what I have to say and perhaps make a good judgement based on a real thought.
 
But IMHO, systemd is the wrong answer to the correct question. We really NEED some reliable management and control over system processes.

Oddly reliable management and control has happened for over 40 years. We do need something like that in Windows, that rules most of the world.
 
Binary is really just short for "binary numeral system" it's nothing but a different way to present numbers.
Just read the papers ... and the rebuttals. Binary seems utterly easy -- it is NOT as soon as multiprocessing comes to the table.
 
An operating system developed without concern for administration doesn't sound possible to me.
 
Being a system administrator was part of Ken's degree. Why would they continue to make operating systems comfortable for admins unless they were one?
Indeed no one is saying that Ken and Dennis didn't have admin knowledge. ;) What we are pointing out is that admin skills are the minimal part of the process to build an OS.
 
An operating system developed without concern for administration doesn't sound possible to me.
Yup, it would be like building a car without controls. :) However, controls are the final part. You need engineers to make the engine first. ;)
 
Yup, it would be like building a car without controls. :) However, controls are the final part. You need engineers to make the engine first. ;)
To make an analogy to the "you need administrators" point - imagine a F1 race car designed and build by the drivers. And only the drivers. Yes, you need to know how to drive the thing if you design it, but you do not need to be a total expert at it. You need to know how that thing works when you are an expert driver, but you do not need to be an expert mechanic or engineer. So in designing an operating system, I would take administrators into the team, as well as hardware engineers and architects. But I would also make sure that none of them would have the last say in any matters.

And as far as systemd on FreeBSD might be an issue, it will never be able to run on the kernel. The people designing the thing felt it was a good idea to make absolute sure that thing would only run on Linux, using a GNU user land. I would doubt it could work with anything non-glibc. So the problem we face is not that systemd suddenly comes to town, but that all applications upstream fall into "all the world is a VAX" traphole and take systemd for granted everywhere. That would seriously impact the available applications. And maybe that, among other reasons, is the key point RedHad is making with pushing this. Making sure that nothing but Linux will have good support for Gnome[3+] for example, and that they be the one with the best support for their abomination of an init system.
 
Oddly reliable management and control has happened for over 40 years. We do need something like that in Windows, that rules most of the world.
Not really. What's almost completely missing is monitoring of the processes. BSD init offers /etc/ttys (the name already suggest it's intended just for tty-controlling processes), SYSV init offers /etc/inittab. Both are useless for anything a bit more complex, so most of the process launching is done from scripts. This offers all the flexibility you need, but monitoring gets out of scope. systemd tries to solve this shortcoming, which would be nice, but ... it tries to solve tons of utterly unrelated (or even non-existent) problems as well, tightly integrated, and, as if this wasn't bad enough already, it does it in a "linux-only way". That's what I meant when I said it was the wrong answer to the right question.
 
I guess I'm just really bad explaining myself. This area is for feedback I think. I really love FreeBSD and I don't want it to change. This is my worry. Since Debian changed I worry that FreeBSD could easily change into something similar. Only a handful of "distributions" now have init anymore. This is a huge change and it came out of the blue.

I hope that never happens either, but if those people won't come to a realization of how they ended up in that situation to begin with, then it is hopeless and we will see the process repeat itself.

That means recognition that it wasn't just someone else's fault, but likely theirs too.
 
Competing with Windows, and trying to beat it, is not FreeBSD's goal like it is for Linux.
There are people still claiming this isn't Linux' goal, either, and IMHO they are right, but we will see if they somehow manage to be heard again ....

On a side note, with all the strategic changes happening over there at Redmond, maybe all this "competition" thinking could be leveraged a bit soon ...
 
I guess I'm just really bad explaining myself. This area is for feedback I think. I really love FreeBSD and I don't want it to change. This is my worry. Since Debian changed I worry that FreeBSD could easily change into something similar. Only a handful of "distributions" now have init anymore. This is a huge change and it came out of the blue.

Only constant thing in life is change.

Take a lead and steer the way you would want your platform to evolve. Or someone would pick it and run with it.

FreeBSD has potential for so much.

If you think it is the perfect system and nobody needs to tinker with it then maybe it has reached its evolutionary dead end then it will be ousted by those that emerge as a fitter alternative.

However, not everyone sees FreeBSD as you or many with similar thought process see it.

Many would like to see features added to see FreeBSD go places hitherto not thought of.

As someone has posted about the way Microsoft is moving (this thread and other thread), should show us the trends. OS's themselves will be of lesser importance in future. They will just be the staging grounds for consuming products and services hosted elsewhere. What's on table from FreeBSD?
 
All this stargazing depends on the people who will be affected. Trends are followed by those who with to follow, the rest does not.
What's on table from FreeBSD?
For me, a fat big crate, peacefully humming away in the evening, doing the computing, archiving and processing I do not trust the klaut* with. Privacy has a value, and thus a price has to be payed. Sadly it is payed to keep it, not when you get deprived of it. And anyway, it is payed in the wrong direction.

* a pun, the german word is spoken as you would say cloud, but it means theft. Which applies to the data you store there.
 
Only constant thing in life is change.

Take a lead and steer the way you would want your platform to evolve. Or someone would pick it and run with it.

FreeBSD has potential for so much.

If you think it is the perfect system and nobody needs to tinker with it then maybe it has reached its evolutionary dead end then it will be ousted by those that emerge as a fitter alternative.

However, not everyone sees FreeBSD as you or many with similar thought process see it.

Many would like to see features added to see FreeBSD go places hitherto not thought of.

As someone has posted about the way Microsoft is moving (this thread and other thread), should show us the trends. OS's themselves will be of lesser importance in future. They will just be the staging grounds for consuming products and services hosted elsewhere. What's on table from FreeBSD?

Well I won't be cloud computing and neither will many others here. My files are mine and will be stored on my hardware. If this project ceases serving my needs and meeting my expectations it will be dumped and that you can count on.

You can time travel back to the 1970's and live off a mainframe, the rest of us will not.
 
"You can time travel back to the 1970's and live off a mainframe, the rest of us will not."

Office 365 (cloud) is a mainframe. We have OneDrive, Google Drive, iCloud at least.

I'm an admin in Finland and we have pretty much everything on OneDrive. Everyone is freaking out, for a reason.

I don't know how you understood this, but schools in my country are storing everything on OneDrive despite legislation. It's easy, they've been told to do it, they have no idea that it's illegal.

We have 2 ways to deal with this:

1. Trust no one
2. We are all one

I support #2 and I'm really happy that you are so kind and understanding. Kaya.
 
We have 2 ways to deal with this:

1. Trust no one
2. We are all one

https://en.wikipedia.org/wiki/False_dilemma

Incidentally, OneDrive, iCloud, and Office365 can only be used with specific proprietary operating systems that---by sheer coincidence, I'm sure---were developed or approved by the same companies offering those services. If I didn't know any better, I'd think the exceedingly wealthy and powerful business interests that created those services had something other than a platform-agnostic future of mutual respect in mind...

EDIT: If you want a better, FreeBSD-related example of what might be, look at Tarsnap. Founded by a FreeBSD developer, but usable by anyone running any Unix-like operating system, and leaving the client/customer in complete control of their own data.
 
We have 2 ways to deal with this:

1. Trust no one
2. We are all one

I support #2 and I'm really happy that you are so kind and understanding. Kaya.

I'm not sure if trying to say you're for, or against the cloud computing business. But the mainframe statement was not directed at you.

However the only way I'm surrendering privacy, control and autonomy is by my cold dead hands.
 
I heard about another linux distribution without systemd, void linux. Did not test it myself... yet. :-)

I've been using this for the last few days. It uses runit, which is basically an updated version of djb's daemon tools. Here's a startup script that I wrote:

Code:
#!/bin/sh
exec /usr/local/bin/dnsblock server -v 2>&1

systemd, for all its claims of "simplicity", looks like:

Code:
[Unit]
Description=dnsblock
Wants=local-fs.target network.target
After=local-fs.target network.target

[Install]
WantedBy=multi-user.target

[Service]
Type=simple
User=root
Group=root
ExecStart=dnsblock
KillMode=process
Restart=on-success
PrivateTmp=true
StandardOutput=syslog

My system boots in ~6 seconds. About the same as systemd.

Other than this, it's a very "DIY" system. Which is good in some ways, but not so good in others. For example I spent some time mucking about getting the fonts to look nice (I eventually ended up copying /etc/fonts/conf.d from an Ubunty system) and getting things like pulseaudio running (currently running it as the unsupported "systemwide mode", since the "user mode" doesn't work properly, but is required for Firefox to work properly).

I also installed the musl libc version, which may have been a mistake. One of the reasons I run Linux in the first place (and not OpenBSD) is so that I can run Linux binaries. However, "Linux binaries" are pretty much always linked against GNU libc which is not binary compatibly with musl libc :-/
 
Not sure where this thread is going, but good scientists make things possible. Good engineers take those things and make them practical.
At a first look such sentences are agreeable when reading.

"Good scientists" developed atomic and biological weapons, list extensible.
"Good engineers" made cheating software possible, list extensible

A discourse about what can be considered as "good" often unveils different camps of interests.
Even what is considered as progress can be disputed.

Regarding Systemd I hope that FreeBSD can be defended.
Otherwise I hope for a fork "SystemdFree".
 
Arguing about whether "good" means "competent" or "morally acceptable" indicates that this thread has run its course.
 
Status
Not open for further replies.
Back
Top