I have some ideas but you are not going to like them...
Please see my bit of amateur anthropology to get an idea of the framework I'm going to construct this around. I really believe that many of this site's issues discussed on Meta are really the byproduct of the lack of consensus between the ServerFault community and the StackOverflow community about the scope and purpose of this site.
Ban the Noobs!
Yeah, yeah, yeah. ServerFault is full of a bunch of elitist pricks (and prick-ettes?). But you know what? That's OK. Everyone's forgotten that the ServerFault community has an explicit requirement for some level of ability that we are calling professional capacity. Everyone seems to have forgotten that StackExchange is supposed to let communities have some kind of ownership of the scope and content of their respective sites ...we empower our communities to curate [content]. And everyone seems to have forgotten that all the closed questions belonging to those poor, angry and alienated folks that just got here yesterday not only are likely to get better answers on other sites but that those sites are explictly open to questions from everyone. It's like a bunch of weekend warrior Home-Depot-DIYers showed up an IBEW convention and then got upset that they were told in tones of increasing frustration that neither was this the appropriate convention for them nor that their contributions were welcome or relevant.
I'm exaggerating a bit for comedic effect so please don't take the above paragraph too seriously. You also really need to have read my previous answer about the "culture war" between the ServerFault community and greater StackExchage community.
I just waded through the Review Queue. Here's what I found:
- 4 questions from SO users with at least 1000 rep that were either off-topic or lacked sufficient information, background and details to meet the professional capacity requirement. I'm an idiot and if I can figure it out so can these guys.
- 8 questions from low-rep users with less an 100 rep on ServerFault that were "I CAN HAZ CODE?" questions, off topic, unintelligible or otherwise lacked sufficient information, background and details to meet the professional capacity requirement.
- 1 question from a brand new user to StackExchange that was completely off-topic. If you can't or won't read the About/FAQ then you certainly are not meeting the professional capacity requirement.
- 7 questions that I voted to leave open as they more less looked on-topic and met the professional capacity requirement. Many of these were in the queue as duplicates but I'm generally willing to give people the benefit of the doubt unless I have a deep enough understanding of technology to know if their question really is a duplicate.
60% of those questions don't meet the "curate the content" model being advocated by a portion of the ServerFault community. 60% of these questions should of been stopped at the Help Center. A quick glance through the front page reveals a similar ratio. How do we stop them and completely invert our signal to noise ratio?
It's heavy-handed but I think this would do the trick:
- Anonymous users can't ask questions
- StackOverflow users don't automatically get 100 rep points when cross-joining sites. 30% of our review queue is composed of SO users that don't bother to read our About/Help.
- Users need at least 100 rep to ask questions.
The bottom line is for years ServerFault has asked that people comply with our discretionary policies for topicality. We have tired all the manner of cajoling, FAQ-rewriting, transfer-pages without success. As the site becomes more popular the signal to noise ratio decreases accordingly.
If this site is to survive it is time to implement mandatory polices. We asked nice. Thrice. If you can't provide a few decent answers that deserve those rare and tasty upvotes then you certainly don't get to spam your help vampire questions all over the front page. Mean? Elitist? Yes. I suppose but what else are we supposed to do when 60% of our questions don't meet the topicality standards?
I've talked about this before and so have other folks notably Ward. The last few weekends I have been looking at the top voters. My estimate is that we only have about 50 users that vote at all during a week period furthermore there are only have about 20 users that cast more than a dozen votes. This is absolutely terrible considering how many views and how many questions we get a day.
I used to think that if we all buckled down and just voted a bit more we could reverse the signal to noise ratio. I now think differently. The engaged audience here is really small - I'm guessing a dozen at best. There's just not enough people interested anymore to make voting work.
However - that's the gamble. If we implement the "mandatory policies" and they actually work, we have a decent signal-to-noise ratio again, and much like animals returning to a restored ecosystem, we can re-engage or retain those people currently engaged and get some voting going.
Why is voting important? I have said this so many times but voting tells users that the community values their contribution. This has immense psychological value. I tried to put on my optimist hat at the beginning of the year and try for the rubber ducky challenge. I'm giving that up. Now my answers aren't stellar but I gave it my best shot and managed to come away with about a 1000 more rep points in six months. No way I'm going to add 10K in a year. In fact, I don't think we're going to see any more users breaking the 10k rep barrier in a long time. What I'm getting at here is without voting there is no signal to any new people in our narrow target audience that anyone is even listening. It's not surprising in the least we don't retain any new talent. If you spent an hour on a decent answer and only get one upvote would you come back?
Embrace the Subjective-Good
This is where ServerFault can really shine. I have also talked about this before. We have an incredibly wide scope in terms of content but because of our ability requirement the scope of questions that get posted is actually pretty narrow. This seems contradictory but let me explain.
Questions that lack depth are almost always closed (with increasing vitriol). Whereas people used to consider leaving these basic questions open so that new users would have something answer now we pretty much vote to close anything that looks like it could be answered by doing a bit of research or a thorough reading or re-reading of the manual. This change makes sense because there are no new users to chew through these basic questions anymore.
Questions with sufficient depth to be interesting and on-topic are rarely answered or occasionally closed as To Localized due to the deep-knowledge issue addressed by sysadmin1138 in the Why "professional capacity"? question. This again makes sense, we just can't seem to attract and retain our target audience, either because of reputation of being a bunch of jerks, our low signal-to-noise ratio, the difficulty in searching through the "noise" questions looking for "signal" questions and lack of participation (lack of voting). What deep-knowledge professional wants to come here, struggle through the low rep stage collecting those rare votes here and there and finally spend all their time complaining in Chat and closing questions. Deep knowledge professionals have no use of this site any more.
This leaves us with a pretty narrow group of good questions that we can expect to see. To increase the size of that group we should really advocate for Subjective-Good questions. Despite the small interested audience there is an absolutely mind-boggling amount of expertise on this site. For someone like myself, a lowly junior in the midst of figuring out what to do, I can only hope that ServerFault can help me and others propel our career in a similar manner to @MDMarra's.
I'm talking about things like (these are just off the top of my head, naturally to make good questions they need some refinement):
How can I backup, secure, and support non-standard, non-network PCs used for SCADA systems?
How can I collect metrics that will let me make good decisions about PC refresh cycles?
How has the prevalence of cloud applications changed the in-house vs. MSP-managed equation for small/medium business?
What criteria should I use to evaluate the suitability of cloud-based applications?
Is it safe to put corporate secrets into cloud applications? How can I evaluate and ensure that the vendor is meeting their promises with regards to security?
There's a lot of gold to be mined here but it is getting harder and harder to get at.
Where do we go from here?
Honestly. I don't know. If this site isn't useful for me and isn't useful for someone like MDMarra anymore I can't be optimistic about its future. I know I have proposed some drastic, maybe even draconian measures, but desperate times and all that...
Personally. I think it is the end. I think we all know it. (DRAMATIC MUSIC!)