In the very beginning, as you found out, SF was a site created by devopsy people as a way to create a community that didn't have all the problems of those that came before it in very much the same way that SO did the same for programming. It worked at first, as has been mentioned a lot.
The scope originally included what both RobM and syneticon-dj have pointed out: interested amateurs who may be professionals in other IT-areas asking questions.
As memory serves, by August 2009 we had a sense of community and enough standards to be indignant about them. I wasn't a mod then, but it was around that time that SO/SF/SU started getting some developer focus on mod-tools. We didn't get our very own meta until 2010, so the meta.so record is light; people who found SF post-beta and didn't know about SO simply didn't know about meta.so as the place to talk about things.
Enough history
ServerFault is not the first place on the Internet purpose-built to be a home for professional systems administrators to talk amongst themselves without having to deal with the clueless hoards who just want a step-by-step HOWTO written for deploying a Wordpress site on a Drupal instance running in an Amazon AWS micro instance. Oh no, we are not the first.
We may be the first open access such site. In elder years we had Usenet (barrier to entry: knowing it was there) and IRC (barrier to entry: knowing it was there, and putting up with the culture), neither of which was terribly discoverable. In the web era there have been several closed-access attempts to do just what we do, but they all died out or are very obscure because they lack discoverability.
As the creators of SO found, SO is a very big Google-magnet when it comes to Programming related things. And when people noticed that a lot of good answers started coming from that one site, they dropped in themselves directly to ask questions. And a community was born. SO became the canonical programmer's resource because it was:
- Discoverable (which maximizes exposure to potential target market)
- Open Access (which minimizes the barriers to entry, so new people can participate much faster)
- Concise (which means people don't have to wade through pages of drivel to get to the good stuff)
The closed-access sites that came before us completely failed the discoverability test and partially failed the open-access test, which meant that they grew by word-of-mouth not word-of-Google. That's why so few people knew of them.
I give partial credit to Experts-Exchange for being discoverable and concise. We've all heard of them, they show up in search results often enough that such is pretty much guaranteed.
Being open access and discoverable exposes us to this problem:
Surly target market http://sysadmin1138.net/mt/blog/2012/09/08/surly-target-market.png
Note: Log-scale used.
For SO, this isn't a problem. People at the very front, the complete newbs, are actually welcome. Their questions may get dup-closed within 20 minutes, but they probably won't get snarked at to hire a professional.
For SF, this is a problem. There are more, a lot more, interested amateurs looking for tips/howto then there are actual professionals in the field. If they've done some research and tried a few things before coming here, we'll welcome them (only the exceptional will fit).
Back in 2009-2010, our discoverability was actually pretty low. I remember cheering when I started seeing SF questions in search-results because it showed we were doing our jobs. Because of that, the ratio to clued-to-unclued askers was very good and we grew a good community of answerers. SF was doing it's job! We had a good site!
We still lost users due to perceptions to too-many-newbs, but we were gaining more people than were leaving in disgust.
Fast-forward to now.
We are discoverable, as proven by the Quantcast traffic numbers (click 'all' to get a nice chart of traffic to the site as it has increased over the years).
The ratio of leaving-in-disgust to attracting-good-answerers has now tipped in the wrong direction in very large part because of the open access policies.
Systems Administration, like programming, is balkanized into domain-knowledge silos. To answer questions relating to Cisco router/switch configuration you need a good population of experts to provide those answers. The same goes for Fibre Channel maintenance, Exchange issues, highly transactional database infrastructure building, large scale video transcoding, and thousand-node-plus configuration-management issues.
You don't get to the point where you can answer questions in things like that without being in the industry for a while, which is why attracting the clued is so important. This is where in a professional capacity comes into play, as it's our attempt at a filter for attracting just that kind of user.
ServerFault has never had true critical mass for some of the deep-knowledge areas, we had a very small handful of people in these tags who did nearly all of the heavy lifting and if a question came in when these people were in bed they just went ignored. Because of that, we didn't get a reputation for:
You know, ServerFault is a great place to talk about Storage Area Networking.
What mass we had has been dribbling away over the last year as the perception of, ServerFault is a place for professionals has been eroded away by the combined forces of the increased numbers of un-clued and the perception by the clued that questions on deep-knowledge areas would go unanswered.
In a professional capacity is a very much needed filter for keeping these deep-knowledge experts around and fed with enough quality questions to answer they like being here. That's why it's there. It's just not working so good anymore.
