- Home /
- Moderators /
Any plan to add a layer of semantic questions for curation?
I'm not sure if this belongs on META or here but I'm confident someone will move it if I made the wrong call.
It seems like there is an ever-present flow of duplicate questions. I'm not sure the "search and destroy" method we are currently using is going to combat the that problem.
First off, I think it's a little discouraging for the poster to get his question smashed because someone else asked a question that is semantically similar but that they failed to detect in their initial search (if any).
Secondly, the answers to questions sometimes change over time. The thing that was useful three years ago might be deprecated, today, or there might just be a newer, better way to do it.
Thirdly, the answers to many questions which are related enough to seem like duplicates to some mods might still vary because the context of the similar questions varies slightly.
What if we took a completely different approach, one of curation rather than enforcement?
If mods could create semantic questions ("How do I only jump when I'm touching the floor?") and then designate certain questions as examples of that super-question, we might be able to at least start organizing the site into the user-driven knowledge-base it seems like it's supposed to be. Ironically, we might even stem the tide of duplicate questions - if those semantic questions are easy to find and tie together a bunch of similar questions, people might find the answer they seek more readily. Someone who finds what they are looking for generally is not going to then go ask the question.
Technically, this is something we could probably start to do with little or no additional functionality from the site. We could just use a process and maintain everything by hand, if we need to do that. It seems like it would work a lot better if we could get the required support together to ask for a feature like this and get it implemented.
Answer by Owen-Reynolds · Nov 05, 2017 at 06:45 PM
As long as I can remember, everyone here has always considered multi-function Qs (your jumping Q is a great example of one) to be non-useful thinking. Take a look at "multipart Qs" in the FAQ (under "How to ask a good question" section.) The only way to help someone who asks a Q like that is to explain how to break it into parts, which probably all have existing good answers.
Your point about answers changing over time - that's why we're supposed to leave answered questions open. The proper thing to do, and this occurs often, is to add a new answer to the Q, and possibly leave a comment in the old answer. I've added "this is for Unity 3.5 and earlier" to some of my answers when I saw someone added a better one (the system sends you an email, so it's easy.) Google is going to bring up that old Q anyway, so that's the best place for the new info.
Are you saying you're seeing duplicate Qs rejected (or closed)? We're not supposed to do that anymore (but the guidelines are out-of-date.) We now "reject" Qs by sending them into the Help Room. It's really the "rejected question area," but calling it the Help Room means people complain a lot less.
The problem is there's no moderator community. It fell apart years ago. There's no sense that moderators should agree on a common set of rules for the site. Take a look at last year's worth of Qs here in the Moderator area.
Heh. I guess I'll start moving duplicate and poorly-written questions to the help room.
Jumping while touching the floor was just the thing I happened to see a rash of right before I wrote this.
I guess my fundamental concern is that I don't believe Unity Answers is functioning as a knowledge base. It's more like a knowledge dumpster. Someone with sufficient context already in their head and enough knowledge of how to dig around can find a useful answer but I don't think that very describes the group of people that this site can help. I think we can do better.
That there is no moderator community might not be the fundamental problem. It might, ins$$anonymous$$d, be another symptom of the same problem.
Part of me still thinks the problem is the lack of curation. Specifically: there's no way to do any real curation... the syntax isn't even there. We can't link related questions together. We can't organize them into properly-browse-able and categorical hierarchies. We can't decouple intent from implementation. There's no good way to manage all the knowledge stored in here.
...or maybe we can do all that and I'm just too new as a mod to know how?
I only just spotted this. I've never stopped fully rejecting terrible questions and non-questions. By this I mean Sent to $$anonymous$$oderation and Rejected. Some even get the ultimate treatment of deletion and in the last month I have suspended more 'users' than you could shake a stick at (Spam bots and lost trolls, probably not 1 genuine user). Genuine but basic requests for help get sent to the help-room. Its easy for the good answers to get lost and they do get lost... buried. Buried under many tons of the same question, but unanswered like some nightmare in which youve done it before but it seemed to never happen.
The problem from way back was people didn't know when they asked terrible Qs, and couldn't understand the explanation why. They often wrote very angry complaints about how "questions about Unity" were being unfairly rejected. Part of the HelpRoom idea was not explicitly rejecting anything which someone conceivably thought was a good Q, thus avoiding all the hate mail.
Ah! From your 1st reply paragraph "someone with sufficient context .. can find a useful answer but ... ." You're wanting to help naive users.
The general feeling here is it's on them to get that sufficient context. The best thing we can do is explain game-making is hard. $$anonymous$$y version of that: https://answers.unity.com/questions/1252826/is-unity-for-beginners-if-not-what-else-should-i-k.html.
You might also take a look at an old thread: https://answers.unity.com/questions/432710/meta-unity-answers-is-degrading-.html. A lot of it is a variant of "how do we get people to read the manual before asking a Q?"
I'm not so much trying to help a specific class of user as I am groping around for the root cause of a problem. The problem I see is that we are getting a lot of questions that don't meet the standard of a good question on a healthy community-driven knowledge-base.
The most historically-successful kinds of thinking tell us that systems are easier to fix than people so, when a system that involves people is producing an undesirable result, the most effective place to look for improvement-opportunities tends to be the people.
I jumped to an implementation too quickly, which is another common problem-solving mistake. I'll read those threads (one of which doesn't load but I can find it, I'm sure.
Well, yes, THE problem is that several years ago UA became flooded with junk Qs from drive-by downloaders; the the old system of politely rejecting them wasn't working. That's the "UA is degrading" thread..
Back then, proposed solutions were 1) fix the system: improve the polite-reject process, 2) fix the people: on the Unity site, near the download button, educate naive users on what a game engine is. Curation (leaving a "bad" Q as-is and putting in in an appropriate place) was regarded as not a useful thing.
I was part of the push for a HelpRoom as a way to polite-reject. While it was being made, the old mods quit; which means the FAQ was never rewritten and there was never a push to adjust the site's mechanics. Option 2 is tough, since Unity3D Co's real customers are game developers. Teaching people how to be game devs is hard, and an entirely separate field.
Your answer
![](https://koobas.hobune.stream/wayback/20220612142442im_/https://answers.unity.com/themes/thub/images/avi.jpg)