> And even so, Raspbian relies on Systemd, despite the privacy fears of many.
While I do agree with Systemd bashing in general since it completely breaks with Unix design principles... this is the first time I have seen privacy as an argument?
While I do agree with Systemd bashing in general since it completely breaks with Unix design principles... this is the first time I have seen privacy as an argument?
A few concerned individuals raised their worries in a ticket and were shut down for 'tinfoil hat reasoning' or something like that.
It may be an extreme case, but many Linux users would rather not have any particular provider baked into core system services. Personally I'd rather know my DNS server is down or that I've misconfigured it, much more than I would have my system contact Google without my knowledge.
Src: https://github.com/systemd/systemd/blob/41c81c4a626fda0969fc...
As a last resort, you can always block requests to 8.8.8.8 with iptables.
It's a pretty rare phobia to be honest, and I think people who consider 8.8.8.8 a reasonable default are totally justified to. For those who don't agree there are plenty of options, that's how open source is supposed to work.
My main point here is just that open source maintainers don't owe humanity anything - they already gave it a lot of their time. If people strongly disagree with some design decisions - and it's not a backdoor or something, it's a pretty innocent design decision to rely on a highly available DNS server as a last back up - open source gives them a lot of opportunities to do their own thing.
> I doubt that a couple of requests to 8.8.8.8 will be a significant compromise to their privacy
It could be, it could be more. I don't think either of us are in a position to say exactly and likely largely depends on whether such default behaviour was changed by package maintainers in the various distributions.
> My main point here is just that open source maintainers don't owe humanity anything - they already gave it a lot of their time.
This is something we both can agree on however I tend to apply this only to those not receiving a salary for doing OSS work. I am of course not endorsing harassment or anything of the sort. People should always choose respectful conversation and debate when discussing these issues.
However I was under the impression an OSS project could use *.pool.ntp.org as their default but was preferred that they get a vendor zone.
Edit: from the other comments in this thread it seems like it it is a build time option
That said, it's mind-blowing that systemd needs a DNS resolver to start its boot sequence. It should really be able to run without one, and have this as the default. It should also not try to set the network up before it runs the network setup code, and respect all of the user's network configurations. The entire thing is completely messed up, but this one complaint isn't all that it looks like.
> I think the problem many see with systemd is that it is a very large block of code that is hard to understand and modify. This makes it possible for unscrupulous organizations to hide things in systemd for spying on users. This also makes it more susceptible to hacking and less secure. I'm not knowledgeable enough about the subject to have a strong opinion about it. I just know that many Linux users strongly oppose it.
In my view it's hard to take the author seriously with claims like this, and in the same paragraph praise for the Raspberry Pi as an open platform (it contains proprietary components).
Edit: More egregiously, the author claims that iOS and macOS are based on Linux.
I was under the impression that changing the default NTP servers was an expected part of bundling systemd in a distro?
So I had to back up, reinstall and encrypt the whole hard drive, including swap space, because that's how debian installers do it...
The anti-systemd brigade has translated this into "OMG LENNART SPIES ON MY DNS!!1!". There's cogent arguments to be made against systemd but the privacy angle is one of the weakest.