18

As part of the process to create new sites, the Community team reviews proposals to make sure they would be good additions to our existing set of sites. One reason we close proposals is when they significantly overlap with existing sites. Starting a new site is a major investment of time and effort, so we'd prefer new topics go to existing communities whenever possible. That way allows users to to take advantage of existing moderation teams, experienced users, tagging schemes, meta policies, and so on.

One good reason for splitting a site is when the culture of the new topic conflicts with the culture of the existing site. An ideal case study is the Unix/Ubuntu split. Despite extensive overlap in topic space, the split has largely been successful. Before today, I didn't know much about elementary OS. It looks like a community that shares a lot with Ubuntu, but also distinct from it. In other words, a good potential split.

However, when I look at the proposal commitment metrics, I see a potential problem:

Commitment stats for Elementary OS

In order for this proposal to reach private beta, it will need more users who have experience with other sites on the network. It's important for a portion of early users to know how the various reputation, voting, and privilege systems work. Therefore, I suggest that those who have committed to this site but haven't used Stack Exchange in the past, head over to the [elementary-os] tag on our Unix & Linux site. That way, you will:

  1. Learn how Stack Exchange sites work,
  2. Increase the commitment progress of this proposal, and
  3. Get answers to you existing elementary OS questions.

In terms of the metrics I use for evaluating spinoffs, here is the baseline for Unix.SE:

N     median_views avg_score closed_rate avg_answers accepted_rate answer_rate median_tta 
----- ------------ --------- ----------- ----------- ------------- ----------- ---------- 
61205          335       3           6.7         1.6          51.6        85           36         

And here are the same statistics for the [elementary-os] tag:

  140          295       0.8         4.3         0.8          37.1        64.3        261    

Overall, I think this proposal is a good candidate for a split if people are ready to use it on day #1.

Proposal: elementary OS

6

I'm in two minds about the creation of a site about Elementary OS. I generally oppose splitting off sites about a specific distribution. Ubuntu is an exception: I supported and continue to support Ask Ubuntu as a separate site, because it has an audience that is rather different from Unix.SE.

  1. Ubuntu is the distribution with the largest user base.
  2. Ubuntu favors a standardized user interface.
  3. Ubuntu cares a lot about users who don't care what's under the hood.

Most other distributions fail these criteria. Even Ubuntu users who do tinker and who do not use the default user interface may be more comfortable on Unix.SE. The difficulty that casual Ubuntu users may have with Unix.SE is that they're likely to be confronted with questions like “what window manager are you using”, and answers that say “run this command” without explaining things like opening a terminal and switching to the intended directory. This isn't to say that you have to know what a window manager is to ask questions on Unix.SE, but we tend to make the default assumption that people do know, which is a poor fit for people who only know and care about a single environment.

How does Elementary OS fare with these criteria? 2/3:

  1. Not a huge user base.
  2. Yes, that's the point.
  3. Yes.

That's different from, say, Debian, Fedora or Arch, all of which fit comfortably on Unix.SE.

With respect to your criteria:

  • You don't have a measure of the number of active users. The tag [elementary-os] on Unix.SE does not mean “I'm using Elementary OS”, but “my question is specific to Elementary OS”. While the relative popularity of distribution tags is somewhat correlated with the popularity of the distribution on the site, a lot of Elementary OS answers are to be found under other tags such as [ubuntu], [apt], [bash], etc.
  • On the other hand, I think the low scores and low answer rate does indicate a misfit between Unix.SE's primary audience and Elementary OS's. The difference is not as great as the numbers might seem to indicate — all distribution tags have lower answer rates and higher TTA than average (I think this is because many of the questions with a distribution tag are questions that should not have a tag but that few people were interested in and nobody bothered to retag). But Ubuntu, Debian, Arch, Fedora, CentOS all have >1.5 average score, ~80% answer rate and <120 median TTA, which is far better than Elementary OS.

So if Elementary OS can muster a large enough community on Stack Exchange, I think it will fare well on its own site.

  • Well, that's what the different stages are for. Right now the site is only 18 users with +200 rep away from going live. How large a user base a distro has is not as important as how committed that user base is to helping others. Also "You don't have a measure of the number of active users", does Ubuntu? If so, how? – Gabriel May 20 '15 at 19:14
  • 1
    @Gabriel What do you mean by “does Ubuntu”? I don't think Ubuntu has a measure of active Elementary OS users on Unix & Linux, but why does this matter? What does matter is how many Elementary OS users are active on Unix & Linux, and what I'm saying is that the popularity of the tag is a poor indication of that. – Gilles 'SO- stop being evil' May 20 '15 at 19:31
  • I'm asking what measure of active users Ubuntu has. Since you request one from elementary, I wonder what measure Ubuntu has that could provide such an answer, since this is the distro you are comparing elementary to. – Gabriel May 20 '15 at 19:32
  • 1
    @Gabriel I'm not requesting a measure of active users from Elementary. Jon is making one and using that as part of his decision to support the site or not; I'm informing him that his measure is not accurate. – Gilles 'SO- stop being evil' May 20 '15 at 19:34
  • I don't see where Jon is talking about active users. He's simply providing the stats for the elementary-os tag. You criticized this stating You don't have a measure of the number of active users. The tag [elementary-os] on Unix.SE does not mean “I'm using Elementary OS”, but “my question is specific to Elementary OS”. So I requested how you would measure the "active users" number for Ubuntu, since tags are not a proper way to do it. – Gabriel May 20 '15 at 19:40
  • @Gabriel I'm sorry but I still have no idea what you're requesting from me. The number of active users comes in because that's one of Jon's points when considering whether to spin off sites). I don't see why it would be useful to measure the number of active users of Ubuntu on Unix.SE, since the creation of an Ubuntu site is not at stake. – Gilles 'SO- stop being evil' May 20 '15 at 19:43
  • Thanks for the link, that is an interesting discussion. My point is simple: you are using Ubuntu as an example of a distro whose spin-off is warranted, at the same time you're stating that we do not have the number of "active users" for elementary OS and that the [elementary-os] tag is not a good measure of it. So my question is: how would you measure the "active users" for Ubuntu, ie: for a site whose spin-off is functioning as intended? How do we compare the "active users" in a spin-off that is working as intended with this one? – Gabriel May 20 '15 at 19:51
  • @Gabriel The relative popularity of distribution tags (looking only at distribution tags, not, like Jon is doing, at all tags) gives an idea. But then you need to fudge because of derived distributions, e.g. many questions applicable to Elementary OS but not to Linux in general are tagged [debian] or [ubuntu]. But in any case, as I explain in my answer, I don't think it matters much. – Gilles 'SO- stop being evil' May 20 '15 at 20:03

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .