8
$\begingroup$

I'm posting this as a place to share thoughts - clearly it is not our decision but it's good to have an understanding of what factors may affect that decision, especially as that decision has been delayed until next week to give us a last chance to prove that our dip in questions is temporary.


Obviously we all hope that we will be granted a public beta - that's why we signed up. A separate site feels positive and independent and gives a sense of identity. However, none of these is the reason for being a separate site, they just provide us with motivation.

I was very interested by the bullet points in Robert Cartaino's answer to Apparently this is a duplicate, so I'll reproduce a few of them here for discussion:

In order to split off a site, we would need to see a clear demonstration that the topic will likely do much better on the new site (presumably without gutting the existing one). We would be looking for things like:

  • An ecosystem of questions that are not already in scope on the existing site(s)
  • A show of numbers that clearly demonstrates the parent site is failing to attract that audience ← this is a big one
  • A weighty argument that a new community exists that does not identify themselves as part of the larger subject (Ubuntu versus Unix)

My own thoughts on this have ended up longer than I'd expected, so I'll post them as an answer rather than cluttering the question.

$\endgroup$
15
$\begingroup$

Although computer graphics can be seen as part of programming, or part of game development, there are aspects that most programmers will not necessarily be good at or even familiar with, and there are areas of computer graphics not used by most game developers, with conflicting requirements and priorities. Computer graphics questions can be answered elsewhere, but the answers received may not be good ones. This is worse than no answer at all.

  • If you ask about offline rendering on Cooking, you will get close votes and no answer.
  • If you ask about offline rendering on Stack Overflow or Game Dev you may get answers that get upvotes but are misleading or wrong.

An example: Perlin noise questions

I asked about a common technique which seemed counterintuitive when I examined it. This technique is a commonly suggested approach which can be found in various places online and could easily be mistaken for best practice or a standard method. This gives the potential to provide a Stack Overflow answer that looks right and gets upvotes. However, the answer on our site shows that this commonly suggested approach is not only far less efficient but also noticeably visually inferior to a much simpler approach.

Expert voters

We need not only expert askers and answerers, but expert voters otherwise voting is swayed by what sounds plausible and rewards the wrong answers. Asking on Stack Overflow may give a non-expert's answer next to a graphics expert's answer. However, it is voting that decides which goes to the top and seems plausible. If the voters are not graphics experts the non-expert answer may end up at the top. The Perlin noise example shows that even backing up with links to an established method elsewhere doesn't guarantee the right answer. So we need a community of graphics voters.

Is computer graphics handled well elsewhere?

Graphics is a broad area with only some of it being on topic on Game Dev. Stack Overflow seems more likely to contain the scope, but are questions there answered well? Looking through, my impression is that the answer is "sometimes". There are graphics questions answered very well, and there are others that have sat unanswered or with just a link to questionable or unclear material.

Even though our question rate is low, questions here are seeing more voting and better answers than some of the similar questions on the same topics on Stack Overflow. Note this only applies to some - we can't compete with the questions that got popular there. This also can't be a meaningful comparison until we get enough questions.

This is relevant to the second bullet point about a parent site failing to attract that audience. There are questions that are on topic on Stack Overflow, but they don't get much attention there due to being swamped (not every question that is on topic here would be tagged "graphics" on Stack Overflow, so graphics experts on Stack Overflow won't necessarily see them).

Our wider scope

We also seem to allow questions here that help with understanding, rather than just fixing code. It's been interesting to see the scope develop (and I've been testing the boundaries with borderline questions), and it seems focused but broader than I'd expected. I think this matches the more specific range of our still very broad subject. We cover more aspects of a narrower field than Stack Overflow. We have become a graphics community which answers more than just Stack Overflow questions about graphics.

Combining two communities

In addition to feeling that we benefit from separating our graphics community from the vastness of Stack Overflow, I also feel there is a lot to be gained from having a site that caters to both realtime and offline graphics. These are two quite separate communities (for example realtime games programmers and offline animation programmers) but despite taking different approaches and having different priorities, there is so much overlap in expertise and perspective that I expect each to benefit the other. I want to be able to ask questions that are exposed to experts from both sides, rather than just getting the game developers' side. Also, topics gradually drift from offline into game dev as software and hardware improve, which is another reason to bring these two communities together.

Our little community

All this being said, our community (active users) is small, even for a private beta, and our question rate is low (although far better than our previous private beta, even if you exclude my questions since I've been much more question focused this time).

We have a huge number of users who committed and signed up, and are on our user list, but have not asked, answered or voted even once, either on main or meta. These ghost users suggest that we made it to private beta artificially, our numbers buoyed up by people who are interested in a computer graphics site existing, but not interested in contributing to it (even by voting).

My impression is that there are good reasons for having a computer graphics site rather than just asking on other Stack Exchange sites, but before we can make it happen we need a community big enough to keep up the momentum. What we have at the moment is high quality but slow. Faster than some of the other failed private betas, but not quite enough questions to seem sustainable.

Remember that you can invite new members to join us and help with our question count, and you can ask more questions yourself (especially if you've found yourself focusing on answers so far).

Experts and askers - a false divide

I think part of the problem may be that our user list seems to have divided into experts and askers, whereas ideally the experts would also be asking. Without those questions that the experts currently think are too advanced for this site, the site will never become a place that seems advanced enough. Please don't restrict to only asking questions you know the answer to. If we get some unanswered questions, that's life. If we don't get enough questions, that's site death.

I'm guessing that the final decision will be affected by not only how many questions per day we manage, but whether we start to see our experts contributing questions rather than mostly answers. We obviously need the answers too, but without expert questions we're shrinking not growing.


I think the claims I make here sound intuitively plausible, but what is needed for decision making is numbers to back it up. With the current size and activity of our community, I don't see how we can provide that. If we want this to be more than just guesswork, we are going to need more members and more questions.

One thing that gives me guarded optimism is that despite most of our questions coming from just a few people, recently there have been a number of first time questions from new people, suggesting that we could be beginning to expand. We'll need to prove that quickly if we are to make it to public beta though.

$\endgroup$
  • 4
    $\begingroup$ i think you are understating the fact that simply many of our questions would not be on topic in SO. And could not be. $\endgroup$ – joojaa Aug 20 '15 at 18:30
  • 1
    $\begingroup$ @joojaa I definitely agree with that, but unless we get a flow of questions going now, we won't get a chance to prove it. $\endgroup$ – trichoplax Aug 20 '15 at 18:33
  • 1
    $\begingroup$ @joojaa I hope that didn't sound dismissive. I wondered if it would be worth adding an extra answer to list examples of questions that would only fit here, but I don't think that is the main obstacle to reaching public beta. If we had plenty of questions each day and showed we were a self contained community, it wouldn't matter if all our questions were on topic elsewhere. However, if all our questions are unique to our site but we only have a handful of people asking, then that's not going to be enough. $\endgroup$ – trichoplax Aug 20 '15 at 19:44
  • $\begingroup$ I notice that you never mention Computer Science. Does the answer change, considering that many if not most (?) non-programming questions about computer graphics are probably ontopic on Computer Science? $\endgroup$ – Raphael Jun 12 '16 at 15:16
  • $\begingroup$ @Raphael My points apply similarly for those sites I haven't mentioned. Similarly to Stack Overflow, Computer Science would accommodate some but not all of our on topic questions. Perhaps Computer Science would perform better than Stack Overflow in the example that I gave of a poorly answered question, but I still think that over the range of questions we receive we benefit from expert voters so that votes reflect specific domain knowledge. $\endgroup$ – trichoplax Jun 12 '16 at 15:35

You must log in to answer this question.

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