Since the Vi and Emacs proposals entered commitment so close together, it's inevitable that they would be compared to each other. But internally the conflict had nothing to do with the editors themselves1 and everything to do with Stack Overflow. Back in July and August, Shog floated a Stack Overflow Academy proposal. We are well aware that our flagship site is far too large to be a friendly place for people who don't know the culture. To us, that's a problem because we believe the site should be a long-term resource for the next generation of programmers. SOA was a vehicle to explore that issue.
On the topic of splitting support channels, I wrote:
We [community managers] are currently divided on the wisdom of launching this proposal. All of us have at least some reservations....We shut down 1 in 5 sites that reach beta. If this becomes a clone of MSO (or the help center), we'd hardly keep investing in it.
In other words, splitting off a topic only makes sense if the new site has something unique to offer. A couple of weeks later, we shut down our proposal. Despite really liking the idea, I just couldn't make an argument that it would be anything other than a small clone of Meta.SO.2 At best, it would be a "nicer" copy of the content on the meta site.
Shog recently mentioned in chat:
The best argument for splitting off a topic is that there's no single site where the entirety of it is already welcomed.
The next best argument is that there's a unique audience that isn't welcomed on any existing site.
The Stack Overflow Academy proposal failed to pass either test because we just couldn't come to grips with how such a site could work without an identical topic and the audience as our current Meta Stack Overflow.
When I first spotted the Emacs proposal, I was immediately inclined to close it on those principles. Over the weekend, the Emacs proposal passed from definition to very near 100% commitment. We added it as an emergency agenda item to our Monday meeting. My notes from that meeting:
If this had been language split (Ruby, Haskell, Lisp, etc.) we would have shut it down immediately. But Emacs sits in a fuzzy boundary between language, platform, software package, and lifestyle. (I'm only slightly exaggerating on the last point.) As such, Emacs questions (especially those that don't have much to do with programming) are underserved on our network.
If the site mostly consists of Elisp questions or is duplicating content on Stack Overflow, we'll close it down in the private beta.
We aren't interested in starting dedicated sites for each and every text editor out there. ... I'd be surprised if we got an Eclipse or TextMate site any time soon.
The next day, I wrote up a more extensive summary of what we'd decided. But notice that we didn't have any hard numbers yet. In essence, we were making decisions from our gut, which isn't a good idea as a rule. So I put some time into answering the question Does it pay to spin off sites? I wanted to know if questions got more and faster answers, and more views when asked on a site other than Stack Overflow. This is the baseline:
Tag/Site questions views score closed answers accepted answer TTA
median avg % avg avg % *
------------------- ----- ----- ----- ------ ------- -------- ------ ---
Stack Overflow 8032148 248 1.6 4.5 1.7 57.5 88.4 24
And here are the Vi/Vim and Emacs tags:
[elisp] 2884 160 3.2 1.9 1.9 77.8 97.1 55
[emacs] 11505 227 3.6 3.3 1.9 68.9 93.5 68
[vi] 1087 289 11.1 6.9 2.9 71.7 96.2 9
[vim] 14637 260 5.1 4.4 2.1 73.6 96.1 22
The astounding thing to me is that by nearly every measure Vi and Vim questions are in better shape than the median question on SO. If you ask a question about vim, 96% of the time you will get an answer with a non-negative score and 74% it will be the answer you were looking for. Half the time you'll wait 22 minutes or less to get an answer. Half of vim questions will be seen by 260 people or more. Compared to many topics (including emacs) these results are exceptional. Shortly after that, I wrote on this proposal:
How many non-programmers use vi or vim? Or to put it another way, can we expect this proposed site to serve a broader audience than folks who already know and hangout on Stack Overflow?
When we looked at the example questions, it was difficult to find questions that could not be asked on Stack Overflow. The few off topic for our programming question site are on topic for one of the handful of operating system sites (particularly Unix/Linux). Vi[m] is a hardcore programmer's editor; the learning curve is too steep for mere mortals. Splitting off the topic could only make it harder to find the answers to most user's questions.
Because of the ancient editor wars the expectation was that if we launched Emacs, Vi/Vim would follow quickly. And there's nothing I can say or do that will make y'all feel like you got a fair shake. Maybe the silver lining can be that Emacs, as it is wont, got a bloated Q&A site while Vi/Vim just works in a handful of Stack Overflow tags. More relevantly, we think the future of Stack Overflow involves the formation of real tag-based communities. In the years to come, we expect that creating a separate site, as we just did with Emacs, will have no particular advantage for a community of users.
I once wrote:
For the purposes of full disclosure, I must say that I am an XEmacs user. It stands at one of the twin peaks of Unix editor evolution. (On the other peak stands Vim.) With a little work you can make it do nearly anything, including sort mail. It's also darn good at editing documents.
This was shortly after what might be the final official release of XEmacs. So I might not be the best at picking winning horses.
Shog's not kidding about team meetings, by the way:
Everyone involved on both teams has very strong opinions about how these sites should operate. So when a new feature is proposed - even very trivial changes to how some bit of the system works - it is extremely likely that the need for the change will be challenged.