Scott J. Remnant claimed that it’s the development community that use Debian testing or unstable. This doesn’t match my personal experience (I know many people running Debian testing/unstable who aren’t interested in Debian development), however I needed a better way to check this than just “my personal experience”. So I tried to answer the following question :
Are normal users using Ubuntu edgy and Debian testing/unstable, or do they stick with Ubuntu dapper and Debian stable ?
Simplified summary: (read below for details)
Ubuntu users:
…Ubuntu dapper (stable): 96%
…Ubuntu edgy (development): 4%
Debian users:
…Debian sarge (stable): 24%
…Debian testing/unstable (development): 76%
Detailed process:
First, I did a poll on #ubuntu@irc.freenode.net using a CTCP VERSION query, and looked at people using Xchat and Ubuntu (Ubuntu is displayed in CTCP VERSION replies). I got 288 replies from Xchat users, 197 of them were using the Xchat package from Ubuntu. 189 (96%) of them are using version 2.6.1 (the version in Dapper), 8 (4%) of them are using versions 2.6.4 or 2.6.6.
Then, I did the same poll on #debian@irc.oftc.net. I got 85 replies from Xchat users, and filtered out 15 of them because they weren’t from Debian users for obvious reasons. 17 users (24%) are using versions 2.4.0 or 2.4.1 (Debian sarge), 53 users (76%) are using versions 2.4.3, 2.4.4, 2.4.5, 2.6.1 and 2.6.4 (Debian testing/unstable – v.2.6.4, the current version in testing/unstable, has 41 users on its own). Statistically speaking, this isn’t very good, so I compared those results with another source of information : Debian popularity contest. 3002 submitters are using v.1.28 of the popularity-contest package (version in sarge) while 9682 submitters are using v.1.33 (version currently in Debian testing/unstable). That accounts for respectively 24% and 76% ! Same as above ! (OK, this is not so good, because I haven’t considered the intermediate versions here, while I did above)
Of course, those numbers are huge approximations, and are subject to a lot of hypotheses and questions, amongst them:
- Are people on #ubuntu and #debian comparable ?
- Are Xchat users or popcon participants representative enough ?
- What about users of other IRC clients ? (I can’t do it with irssi, since its version didn’t change between dapper and edgy)
- What about people who change their CTCP VERSION reply ? Or disable it ?
- This only shows the system that people use to do IRC.
- Xchat users obviously run X on the system they are using. Irssi users don’t necessarly do, and I got slightly different results with #debian and irssi : 55 users of v.0.8.10 (Debian testing/unstable), versus 30 users of v.0.8.9 (Debian stable). That makes 65%/35% instead of 76%/24%.
- This is done in the middle of the edgy dev cycle, and near the end of the etch dev cycle. This could change over time.
I didn’t run the same poll on #debian-devel and #ubuntu-devel because I received some unpleasant comments already, and don’t want to run into angry Debian or Ubuntu developers (I’m still waiting for DAM approval after all :-)
I am a developer, but I am not a Debian/Ubuntu developer. I used to use Debian unstable (because stable was just too outdated most of the time). Now I use the latest stable Ubuntu release.
I actually switched to Dapper a few weeks before it was released. Six months is about as long as I can wait before the need for fresh software overpowers the need for stability and guaranteed security updates.
This is an interesting idea, and one of very few things I wouldn’t mind a channel-wide CTCP for.
I’m “normal” Debian user. I use Unstable and like it :)
I am a ‘user’ plain and simple and know nothing about programming
I installed Debian on my workstation some years ago when I’d had enough RH dependency hell. I almost immediately upgraded to unstable and have stayed with that ever since. I don’t have any exotic hardware and simple needs as a user (primarily Family History research for which I use GeneWeb).
I recently set up a server box on which to run GeneWeb for our home network and used an Etch beta ISO. I could easily have upgraded it to unstable as that is quite stable enough for my needs but left it in testing for now.
I will _never_ use ubuntu or a derivative. Just a ‘gut feeling’ on my part but I think that in the long run ubuntu isn’t going to be of real benefit to Debian.
Note that you’re only testing “desktop” users with this method. I’m running Debian stable (and even oldstable) on a dozen systems; but I never IRC from them.
I use my development box for IRC, and unstable is the obvious choice for me.
It would also be useful to differentiate between testing and unstable. Maybe popcon could give us some more reliable results.
The “stable” version is used for 3032 submissions.
11016 come from post-stable versions, 81 from backports.org, 752 from older versions. That would make 3865/14881 ~ 26% using pre-testing versions. But again, I don’t run popcon on my servers, only on my private development boxes.
Hi,
I agree with you about many non-developers using debian testing/unstable.
I read debian-users-german on a daily basis and many people just use testing or unstable just because they want more up-to-date software and they are content with etch/sid.
Cheers,
Moritz