Wow, it's been quite a while since I've been here.

Bit quieter than I remember as well.

Hey

Allo

Hello!

How have you been?

I'm okay

Busy

I saw a Steam sale that made me think of yall and I wanted to drop in and say hi.

I feel you. I've hardly had any time to do anything this week.

Awwww. :)

@Summer 🙂

I've been in a funk. I have a lot to do but I haven't actually been doing any of it so it gets worse and I get more overwhelmed.

I have a project I can work on at my leisure and get paid for it when its done but I keep dicking around instead :(

@Summer Basically me this month

How about you all. What have you got going on for you?

I made my first app that I got paid for. Although it was crypto, it's still something at least.

Congratulations.

Congrats indeed

That's a huge first step.

I'm working on a research paper about platformer level generation

ohh this looks cool

@Summer Thank you! I’ve looked and apparently it’s still being used. Now I’m just porting it to WPF, because WinForms is a pain.

Is this part of the platform stuff you shared with me earlier?

Yeah, building on that.

Oof. Good work slaying the dragon of WinForms and managing to ship something with it! WPF is also a pain, but at least a somewhat more modern pain.

Pretty cool idea. I'm surprised it hasn't been done before. Just from reading the abstract it seems like such a potentially simple aproach.

I say that with very little background in level design or procedural generation tho

It probably has, but I haven't found any papers on exactly this approach yet. Outside of one for Lode Runner, but there the character controller behaves a bit more like a cellular automaton.

I remember lode runner

@DMGregory Yep. Another tool in the belt. Or something like that. I forget the phrase.

my god that was a long time ago

I decided to port to WPF after getting multiple reports that it just didn’t run well (or at all) and after hearing that WPF was sorta easier.

I've also been dabbling in some HTML and JS. And also C. ;)

@DMGregory People have complained about this issue over and over and SE has made it clear they don't want to change anything. And chat matters even less. They serve to extend people's patience a bit because they might feel like waiting around a couple of months in the hope that their complaints are heard and changes are made, but it hasn't happened and I don't think it will.

@nwp I'm not talking about SE as a company, I'm talking about us as a community. You mention not getting traction on changes — what concrete changes do you think could help improve the situation?

I'm too lazy to find the meta posts, but there is a famous one titled "Fastest gun of the West" or something along those lines.

There have been many suggestions on how to fix that. For example taking away reputation gained by answering duplicates or even turning that negative so people make an effort to check. Also rewarding reputation for correctly closing questions as duplicates.

There also needs to be a mentoring mechanism. Something where you can take someone to the side and explain to them patiently how to do things.

Oh, and the other very famous one about "Help Vampires" and the one about the endless summer.

I haven't kept track of meta drama and maybe it's a bit different on gamedev, but the issues go deep and SE makes no attempt at fixing them. At this point I'm sure they have strict orders along the lines of "The formula is working, do not change it."

I'd say we have a decent rate of dupe-closing on this SE. Have you been finding particular dupes slipping through here recently?

I'm not at all active on gamedev.se, so no idea about rates. That's also besides the point. The point is that SE actively rewards dupe answering and punishes dupe closing.

I'm 100% with you on the need for a mentoring mechanism. We limp along with chat but it's certainly not ideal. The Q&A format is a good fit for folks who already know the lingo, but it doesn't work well for folks who need hand-holding.

Well, not really punishes. You just miss out on the reward and make a target of yourself.

I'll confess, I haven't seen that dupe-farming dynamic or ire toward closers driving behaviour in my experience here. Rep's easy enough to come by via other means.

I suppose it doesn't work so well on gamedev. On stackoverflow you just keep answering the same question about pointers/printf/arrays for an hour a day or so to get to rep cap.

To be fair, "make a target of yourself" is also a bit exaggerated. People give you annoyed comments, that's about it.

I definitely think the volume they have there exacerbates a lot of issues that aren't so bad here. Though the volume of users can also be helpful — less risk of questions languishing without attention, or burning out the few active users trying to coach folks.

Also Tyyppi 🤣 I do wish they'd find that one in search for themselves.

Just ... the feeling of noticing a severe issue, then being told that the issue has been there since the beginning, that it resurfaces every couple of months and that there will be no attempt at fixing it ever is just disheartening.

I totally understand that, at a corporate level. I'm curious if there's anything we can do as a community here to mitigate some of the problems on this specific SE (I don't think I can fix StackOverflow, but I might be able to help make GDSE a little better!)

That is not how it should go.

With duplicates for example, we've had good success by noticing certain patterns of questions that come up a lot — like nulls in Unity, or getting roll in an FPS camera — and making canonical Q&A to link them all to. That makes dupe-closing much easier.

There are many solutions. Mentoring is one, having proper close reasons is another.

Though the latter actively goes against what SE is trying to do. They have dismantled all the effective close reasons such as "Lacks minimal understanding".

We do regularly tell users to behave better, so that's not why the comment was deleted. Just phrasing it more constructively is all it would take to keep comments like that around.

We do have custom close reasons we can refine if we feel they're not serving our needs well

Another thing that would massively help is some sort of MCVE requirement, properly explained and with detailed instructions on how to make those.

I think we'd need a larger user base to effectively do mentoring. Right now I think there's just 3-4 users who will regularly make time to coach a beginner, and we get way more beginners than those users have time/patience. I wonder if there's some support that could make that easier on them.

The main problem with that is that it reduces engagement. People just want to dump their question, not read instructions. If they were willing to read instructions they wouldn't need to ask in the first place.

We do have a "lacks MCVE" close reason that links to such instructions. Maybe we should make some customized instructions for the Unity case though, since that makes up the bulk of our questions lately.

Haha, that's a point. It would be nice if there were a filter where we could find out if a user was willing to engage in collaborative problem-solving before sinking time in an unfruitful comment thread.

Not sure how such a filter could operate though. 😕

This is actually one of the really difficult problems. I don't think there is a way to educate people on how to ask a proper question. They are not interested in that, they want to make a game. I understand the sentiment.

Maybe having a chat bot do the part of asking for missing information until a reproducible problem emerges would work, if we could make something that sophisticated.

I kind of wish, instead of closing, we had a "beginner/student gamedev" SE or other help site that we could migrate those questions to. Though maybe that's just kicking the can down the road. 😅

Still, it's my best guess. Somehow automating the infuriating, boring and stupid work involved is most likely to make all members of the community happier.

Also a sign from above that SE at least kinda sorta cares and stops treating it as a finished product that reeks in money.

Oh yeah, making the search less terrible to make finding good questions easier is also frequently requested.

Camping the main question page with some selected tags is not that great, but I don't have a concrete suggestion for how to make it better. I'm sure they have some in the meta posts.

I was thinking more about being able to maintain two different sets of community norms. You have the beginner site where you can expect to be coached by other students, and the pro site where the expectation is that you know how to search and read docs and prototype/profile. Right now we're trying to serve both masters and it's a serious tension.

I don't think that's viable. The 2 different communities is people who want their question answered with minimal effort and people who want to hone their skills by answering interesting questions. Separating them does neither one a favor. You really need to get the first group to somehow ask proper questions and the latter to have tools to deal with bad questions in order to make it work.

There is nothing wrong with wanting your question answered with minimal effort by the way, as infuriating as it is for people looking for good questions. We should try to minimize the effort required to ask a good question (and not reward people for asking bad questions).

I don't know why I said "we" there. I'm thoroughly put off by previous bad experiences alone that I will not put in any effort.

You're putting in effort in this conversation, and I appreciate your insights. ☺️

So this isnt 100% related to gamedev but you guys are pretty c++ experienced and probably still can help me... Im on windows and i use mingw. In one of my projects im using "filesystem" which was introduced in c++17. It recognizes the includes but cant build the program. However the error message is pretty weird, mostly telling me that stuff like "string_type" were not defined ( gonna attach it in a sec ).

People told me to update mingw which i did using their ui tool... however that didnt helped. Anyone got an idea why i get this error ?

pastebin.com/3ej6UJqs

I've had problems with std::filesystem not working properly too, but I don't know why that happens.

@nwp Glad to see im not the only one ^^

@Tyyppi_77 Already tried the experimental channel... and updated mingw... didnt worked either. Im downloading msys right now, gonna test this out

@doppelgreener the little dragon flourish was a nice touch. "BURNINATED!" 😆

lol

it's routine for the room but yes, extra good this time

@Tyyppi_77 To the newest version ^^ Yeah but the path didnt really changed... I have now installed msys2. And installed opencv via the msys console. However vscode tells me that it doesnt know the include path.

Doesnt matter anymore, i already removed mingw completly... but i just downloaded and installed the newest available version from their site ^^

Looks like it was 8.0 but i have no clue, i just read update mingw and did that. However im now using msys2 and hope to get it running now. But vs code cant find the opencv headers... and i updated the include paths already. Damn next time i just use linux for this shit

Well i just tested it again... it neither can find opencv... and it cant build and compile properly because it doesnt know opencv :/

So looks like it isnt intellisense only

My config looks like this :

The task for compiling

pastebin.com/v9zhqxWu

Intellisense ( if i understood that correctly ) : pastebin.com/T72PhHbd

Both are referencing msys2 and mingw in it... and also the opencv4 directory

Oh look like its opencv4/opencv2 where the headers are... but if i include them via <opencv2/header> it should still find them properly or ? prnt.sc/20wz2bz

gcc complains about "no such file or directoy"

Regarding the include path

<opencv2/opencv.hpp>

@Tyyppi_77 Thanks ! Such a stupid mistake ^^ This already solved the invalid path issue

However... one issue remaining... this time it cant bind the libs correctly : prnt.sc/20wz4s4 ^^

Which makes sense... because there no libs in that place for opencv. But where does msys places them ? Any ideas ? :o

Oh wait

It worked... the opencv version was different and required some other libs... however thanks for your help ! It finally runs :)

Thinking of canonical dupes, I wonder if it would be useful to make a canonical Q&A to link up all the "I need to make a formula that gives me these values" type questions - giving strategies for how to find formulae in general, teaching to fish rather than giving a fish.

@Summer Aha, looks like someone has done it before. If I'm understanding this ProcJam 2015 talk correctly, it sounds like Tommy Thompson is describing something very similar to what I do (though more sophisticated):

I'll need to add that to my references/prior work. But I think my approach might still have some degree of novelty in terms of allowing back-tracking, and being driven by the controller physics (not sure if the "Action Generators" he describes adapt automatically to the controller, or if they need to be specially coded with knowledge of the character's physics).

Aha! Later in the video he talks about "can I parametrize actions based upon the character you're using" - so it looks like that's something the generator he was presenting at the time didn't do, while mine does. :D (Though to be fair, he probably solved that problem in the intervening 6 years better than I have)

Taking a look at a paper he co-authored the next year for more details...

Ahh, so here the "actions" are more high level/semantic than mine. They represent specific level design modules, effectively, that will induce a particular pattern of player actions to navigate them. So it looks like my approach of modelling player action first is still distinct.

("Distinct" is not the same as "good", but for the sake of originality for a school assignment it's helpful. 😉)

Nice