Distribution quote of the week

[Posted January 9, 2019 by ris]

I've said this before, but I think it's an important reality check: it doesn't matter nearly as much who uses Debian, or how many people use Debian, because we are not a company or a product, we don't sell something, we're not trying to make a profit or maintain some growth curve, and we're not part of this capitalist system. We are building a Linux distribution, to a very large extent, for each other, and delightfully other people also find it useful. Sometimes those people even join us! Which is great!

— Russ Allbery

User count in distros

Posted Jan 10, 2019 11:09 UTC (Thu) by meskio (guest, #100774) [Link] (4 responses)

Interesting quote while Fedora is toying around with the idea of counting users (and as a side effect tracking them):
https://fedoraproject.org/wiki/Changes/DNF_UUID

User count in distros

Posted Jan 11, 2019 0:14 UTC (Fri) by pabs (subscriber, #43278) [Link]

Debian counts systems too, except it is opt-in and sends a lot more info than it sounds like Fedora will:

https://popcon.debian.org/
https://qa.debian.org/popcon.php?package=apt

User count in distros

Posted Jan 11, 2019 4:14 UTC (Fri) by mattdm (subscriber, #18) [Link] (1 responses)

For what it's worth, we're probably going to go with Lennart's suggestion of using a client-side "countme" variable rather than using the UUID (an idea I got from looking at what openSUSE does).

Also, "as a side effect tracking them" is... not quite right. The plan was actually to make the server side infrastructure so it didn't do that. But, of course, that does require you to trust that we wouldn't _actually_ lie and track you for the one month proposed lifetime of a UUID. Since the goal has never been to track in the first place, and I think the proposed idea will work just fine, we're looking at implementing that instead.

I have a lot of respect for Debian and for Russ Allbery. I do think that having more insight into actual use really does help us make something better.

User count in distros

Posted Jan 12, 2019 4:41 UTC (Sat) by rra (subscriber, #99804) [Link]

I completely agree that it's good to capture some amount of volunteered statistics (popcon in Debian is useful for us). It provides some (rough and flawed) statistics on what packages are particularly important, for example, and can help answer questions like "if we remove this buggy orphaned package instead of fix it, how many people are likely to be upset."

But every once in a while, someone will say "if this keeps up, I'm going to stop using Debian" in much the tone that one threatens businesses one is purchasing products from, and, um, please use whatever operating system makes you happy! I want you to be happy. I like having more people use Debian in large part because it translates into more people supporting and developing Debian, but I think it's more important that you be happy and use the operating system that's right for you, and no operating system is going to be right for everyone.

They can all co-exist. Diversity is great. Also, that way we can all make different mistakes, which makes the world far more exciting.

User count in distros

Posted Jan 11, 2019 5:51 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link]

I suggest the Debian method - introduce a flaw into SSH key generation. Then wait for a few years and then count the publicly accessible vulnerable servers as a proxy.

Distribution quote of the week

Posted Jan 10, 2019 13:45 UTC (Thu) by miquels (guest, #59247) [Link]

At the Open Source Summit Europe in Edinburgh last October, Greg K-H said (I'm paraphrasing here) "I will tell you something most people do not talk about. The world runs on Debian. Beneath almost every service, cloud provider, you name it - it's Debian. But noone admits to it".