Potential Changes

[garrynewman]

Internally we talked about some changes that we think make sense long term after seeing this weekend's activity. These are open for discussion, so please give any feedback - either in support or not. It's useful to hear about the way you work.

Erroring Games

If a published game errors, we should disconnect from the server and close the game on the first error. We shouldn't be allowing developers to tolerate errors.

We should delist games that are erroring consistently and alert the developer to fix the problem.

One Game Project

We're going to be changing the editor so you can only have one game project at a time.

No Menu In Editor

We don't love that people are playing in the editor. It's cool for your own games, that you're developing, but we don't think it's a great long term strategy to let people basically play the entire game in the editor (because of overhead, content collisions etc).

We'd like to remove the main menu from the editor and only have it available via the game exe.

Game Addons

Right now when running the game (not editor), it'll load all the addon projects too. We're going to stop that, so when you open the game you know it's raw, it's what everyone else is experiencing.

[TheBoxTrot]

Personally I like the ability to be able to jump around between different projects at a whim and having to relaunch the game could get annoying. But I assume this would speed up start up times too?

[TheBoxTrot]

Thinking about this more this would prevent cross contamination of assets from different projects

[BlakeWinser]

Would it be possible to replicate Unity's error system by freezing the game once an error occurs while also being able to unpause if warranted?

[Xenthio]

I feel like delisting games and disconnecting people when an error happens is harsh, it would probably be nicer to have a nag popup for each error to make them more annoying, but not game stopping annoying, plus it might encourage people to purposely try and do things like malformed console commands or other things to sabotage a game and potentially forcefully attempt to falsely delist it

[Remscar]

Getting stack traces in asset party would be nice too. Just seeing that there was a null reference is absolutely useless.

[TheBoxTrot]

With the game addon section does this mean the game won't load local projects when starting a game? Cause I believe that would be a good change as I often forget to disable them in the editor and then I have to restart the game to disable them so people don't have to download all that from me.

[yuberee]

I agree with the last 3, it's probably going to fix some issues I've had myself but could never pin down.

But for the game errors I hoped it could get explained more, are we talking a single NRE disconnects you from the game?
Not throwing shade to the developers, but everytime I see a TF:S2 playtest I see errors, but that doesn't stop the player from playing the game. If the server disconnected every time you get an error, playtests would basically be useless.

[CarsonKompon]

I agree but maybe there should be a threshold. Maybe if a game throws 10 or so unique errors then I agree, but if only 2 or 3 are being thrown, it could be the fault of a new s&box update/api changes (or just a bug that wasn't caught before production)

[yuberee]

Yea unique errors over multiple of the same.
I say it's better to notify developers of their errors more aggressively.

[peter-r-g]

Errors really shouldn't be a thing on publically accessible games. Do internal tests to find errors and save public tests for tweaking your gameplay. If you struggle with NREs you should probably enable Nullable reference types.

[NotGaylien]

I mean.... most games spit out errors or crash and so on, it cant be perfect.
Plus we are people doing this in our free time, why should we be put up to high standards,
if i wanted to have a playable game thats a erroring buggy piece of shit, and people wanted to play my game thats an erroring buggy piece of shit, i think thats totally fine and we should be able to do that.
we are stuck releasing our games on s&box and its the only way for people to access them, so i just dont see how its fair on us.

If its really an issue on s&box and how it looks, just make it so recommended, trending and so on cause games spitting out errors to naturally end up lower down in the lists.

[DelcoGoblin]

Are public playtests just not going to be possible then?

[Gmod4phun]

Stopping the game on an error is probably not a good idea, you don't want to stop a server just in case something occurs that hasnt't been thought of and only happens during very specific conditions. It also would be easier to replicate the error if the game just went on as normal. If there will be errors happening all the time, people will just disconnect on their own and not play the game.

Iirc there is already a system that tracks errors, so there could be a menu warning that the game you are trying to join can potentialy be broken or hasnt been updated in a long time (if there was alot of errors recently for that game).

[Xenthio]

One game project also sounds really annoying, removing functionality just for the sake of it is kinda stupid, it’s nice to hop back and forth between projects, same with the main menu, sure you don’t want people playing the game in the editor, but you don’t have to remove functionality for the sake of “I don’t want people doing this and that”, it’s also important for the settings menu to exist, perhaps making the main menu show local gamemodes only in editor mode as the main page is the way to go?

[kurozael]

You could run multiple instances of the s&box editor, one for each project you're working on. This fixes a lot more issues than it creates.

[CarsonKompon]

Since I'm developing my game Home in a way where the games on the arcade machine also double as standalone s&box games, I'm jumping from game-to-game constantly. Having the menu on the left to instantly launch into any of my games at any time is almost a necessity for me 😅
Launching multiple instances of the editor sounds like a lot of overhead that we don't need.

[matekdev]

Can we get the ability to view the stack trace of an error while in "game" mode? I don't believe this is currently possible. I imagine it'll be needed as games won't be played in tools mode as much.

[handsomematt]

It's always been in the log file afaik

[matekdev]

It's always been in the log file afaik

I feel like it'd be nice to be able to click on it in console and have it display somewhere. Just my two cents.

[TheBoxTrot]

Perhaps allowing games to state what stage of development they are in such as play test , beta , full release. If a game is stated to be full release maybe then you can shut the server down on error

[codecakegit]

Yeah send every game that errors over a threshold straight to the shadow realm, get it out of here! If it can't even compile at the least then kick it off the menu.

[TheBoxTrot]

As far as I'm aware the game has to compile to even get on asset party

[yuberee]

You can't upload it in the first place if it doesn't compile, so only way this happens is if there has been a pain day in between and those usually delist games that fail already.

I'm thinking, and what others seem to think, is that every error that happens during gameplay is going to disconnect the player, even minor ones like NRE.

[codecakegit]

Maybe instead of removing games they get a Beta flair added to it, So people know they are going into a potentially buggy game, they could even sort by ect.

[NotGaylien]

My main opinions

Erroring games

Personally i really hate the idea of kicking people out the game or. hiding the game if the game is erroring.
I rather let the player keep playing and leave the decision for them to stop playing because of errors up to them,
since they might not care, let them carry on if they want, and deal with shit thats erroring.
I like to carry on playing games if the error isn't something major.

If its really an issue maybe have a WARNING this game is producing a lot of errors before people play.

one game projects

This i overall think is pretty useful,
although in some cases it might be a massive pain aswell.
since i have occasionally wanted to take something from another project and i quickly test the 2nd one to make sure it works as i remember before pinching that bit of code.
But other than rare situations like that it makes a lot of sense.
so i agree.

No menu in editor

yeah this makes sense,
i personally have exclusively been using editor for everything including playing games,
thats a really bad idea and makes it super easy for people to fuck with things in games to cheat.
so 100% agree

[NotGaylien]

Quick super important note with

one game projects.

What will happen with content project type, i imagine i will still be able to access them no matter what game project is open.
since as a artist i use multiple content projects alot. and content across multiple projects.

[trundlr]

There is an option for no game project

[NotGaylien]

No i mean like.

Lets say i got a no game project (used to be called content) and i fill that with common assets i use across multiple game projects.
(models, farticles, maps and all that)
I want to always have access to that and edit that no matter what game im working on.

And not have to close my game project, go into my content project, edit the asset i want, upload to asset party, go back to my game project and use it from asset party from there.

Thats a really horrible and kind of make re-using your own assets and editing them a horrible process.

will i no longer be able to do that.
or will they be a thing i can easily include into my game projects and still be able to share across them.

[handsomematt]

Content projects will float around and always be mounted

[NotGaylien]

Ok sweet im onboard with that idea!!!

[TheBoxTrot]

With no menu in the editor it makes it a bit hard to test add-ons for live versions of other people's games

[gio3k]

don't do that error idea - it'll just be jarring and weird especially for small issues that wouldn't really affect the game anyways

maybe just move the red exception notification out of game land into tool land - non editor should just log the error in console and not show anything to the user

[NotGaylien]

Oh also more on what i hate that error idea.

I understand why, having people come into s&box, have lots of erroring games that are broken or barely playable all over the place.
its not a fun time for the players and makes s&box look like a hub for awful erroring games.

But, Publishing games on asset.party and playable in s&box is my only way for people to play my game right now, i dont really have any other options,
if i want to let people play my game even if its buggy or whatever then i want to do that,
having the ability for people to play my game taken out of my control just feels awful, its my game i made, even if i made it in s&box, i still coded it, made the assets an all that (i make all my stuff from scratch in this case lets say) and i want to control when people play it, and when they get to play it no matter what state it is in.

[rileyzzz]

Erroring Games

Maybe this is a hot take but different behavior depending on whether your game is published or local doesn't sound like a good idea.
Especially since addons are a thing. IMO if you install a fucked up addon to give everyone pumpkin heads but it hasn't been updated in 3 years, and it just happens to create bugs when used in conjunction with Annoying Orange Simulator v1.4.23.2, that shouldn't kick you out. This only becomes more likely because of how abstract the s&box API is.

Asset.party isn't Steam. When players open up S&box they aren't searching for some life-changing gamemode, because all the games are free. It's kind of the halfway point between the GMod Workshop and Steam, because yes while you can make more fully-fledged experiences than you ever could in GMod, you can also make hot garbage and get away with it.
Shitty games are inevitable - I hate to say it but sometimes part of the charm is how god-awful broken UGC can get.
If you kick the player out as soon as things go horribly wrong, yes it acts as a deterrent for devs to keep their games not-shit but it doesn't save the experience in any way. It just makes the end-user mad at Facepunch for immediately kicking them out and ruining the fun, instead of being mad at the dev for making a god-awful gamemode that falls apart.

I think once release comes around, for the first few months you're gonna have a pretty stark dichotomy of games: 95% shitposts and 5% passion projects.
No shade towards Ohio RP, but a low effort game that spews console errors is not gonna have the time put towards it to make sure it's error-free or that it runs smoothly on all hardware. IMO the console spew only makes it funnier.
If you're a crap developer and the errors are rare enough, chances are you'll straight up ignore them anyway, and shrug it off when your game gets delisted.

For developers, especially with huge projects like TF:S2, this sounds like a nightmare. "Tolerating errors" is pretty much 70% of the job description of a developer.
Errors and crashes are kind of inevitable with large codebases. The bigger the codebase, there's always gonna be a percentage of people who are catching rendering bugs, have missing files for some reason, have old or corrupted local/org save data, etc.

My only other thought here is that more logging severity levels might be nice here. i.e. Log.Verbose, Log.Success, Log.Fatal. Potentially user-defined channels too, like rendering, audio, gameplay, physics, analytics, player movement, music, etc. If errors are gonna mean the difference between my game getting delisted or not, I'm gonna want to be able to color code/sort my own logs so I can figure out what was happening that led up to it. Compared to what we have now, VConsole seems like a sports car

One Game Project/No Menu In Editor

Fine by me, matches Unity, Unreal, Roblox, etc. If it ain't broke...
The only potential question here is what you'll put in place of the menu when you aren't actually in the gamemode. Obviously I don't want to restart sbox-dev.exe each time I make a code change that can't hotload/requires a reload.
I guess just throw the game's custom menu/lobby up. Or if you're feeling really wild, throw a limited hammer viewport in there or something.

[rndtrash]

Removing erroring games from the public access sounds like a great idea, and it will improve the image of s&box for a general consumer, until someone figures out how to send fake exceptions to blackmail competing gamemodes. How are you going to check if the error is real?

[SupinePandora43]

One can intentionally create addon for causing errors in the competitors' game

[TheBoxTrot]

I assume that would count on the addon not the gamemode

[Phyksar]

I think erroring games topic is like peeling an onion, it has many bad things most of which were already mentioned. I have to add also that this decision may lead to a tendency that some developers will start to "hide" errors, well at least as much as they can, so no Facepunch employees would ever know about them. And this is many times worse. In such case you may not be able to tell if the game is buggy until you spend enought time on it.

[SupinePandora43]

A zero-error tolerance strategy is good for a "production" environment, but not for development.
Maybe move high bug rate games down in the search results and/or warn about instability?

[dch-GH]

I like these ideas. I think the erroring out of a gamemode is a step in the right direction. Not sure how addons would be factored into this though. Would be nice to avoid the "Something is creating script errors" this time around. I'm not sure what the best way of thresholding this system overall would be, but I think moving towards it is a good idea in the long run.

[wheatleymf]

TL;DR about error policy. This policy would probably work best for standalone games IMO, not for simple games inside s&box, especially now when many gamemodes are mostly prototypes or simple games. People are still figuring out s&box and it's too early for any restrictions, and it might be too early even after the initial "full" release. Getting kicked after specific errors sounds frustrating - just let developer decide on their own which errors should do it. Every gamemode is different, you can't know which error can be ignored during gameplay, and which one would require restarting the game.

---

I feel like such changes should not be enforced anytime soon - s&box is still a (technically) new platform and with new invite system there are many people that are still learning the API. I'm pretty sure some of them will release their own gamemode at some point - and there's a chance something will be broken, even if they do a number of private playtests before going public.

Rust has Staging Branch that anyone can access and help Facepunch finding & reporting bugs - and it's quite helpful because lots of bugs end up being fixed before going public, indeed. However, every time new update comes out, I still see some new bug reports anyway - just because some issues can't be found during tests with a small audience, even when Rust team is much larger than 5-6 years ago. Yeah, Rust is also much more complex than any gamemode in s&box, it's a very poor comparison, but even on smaller scale, how do you expect solo devs or small teams to release a game without potential game breaking bugs? No one is making gamemodes for s&box for a living (yet?), everything is just a hobby and not all people would have time to find & fix any massive issues ASAP.

This system might punish new developers just for lack of their experience - their work should not get unlisted (or straight up removed) just because they're still trying to figure out things. Maybe mark these games as "unstable" or warn the player upon launch.

[tylerlowrey]

The stance on erroring games might make sense for a curated selection of games. But I think it doesn't make sense for all games. In my opinion, the player should be able to choose what experience they want. I think that by default the front page that the player is met with should present them a list of curated games that are popular and are in good standing error-wise. I haven't kept up with the details regarding games making money but the FAQ on the website has this: Split a percentage of sales to top played games/addons/maps every month - might work if done right. If that is in fact a potential plan, maybe it makes sense that every month Facepunch selects a handful of games that are popular and don't have issues and both display them on the top of the front page as part of the curated list of games and also provide them with financial compensation. This gives developers an incentive to put effort into their games.

I also can't imagine disconnecting the end-user from a game just because there was a single error. Who here has never pushed a bug into production? Sometimes things make it through the cracks no matter how much testing you do. Also, what if the game is still in development and is listed because they want to do playtesting with a large number of people? There should be a way to easily support this scenario.

The user shouldn't pay for the mistakes of the developer. And some users might enjoy going around and playing poorly made hobby-level experiences. TLDR: Personally, I think that limiting what shows up on the front page of the game to better quality games, but still allowing users to browse through and search for any game (maybe with some indicator that the game isn't curated or high quality or currently has issues) on a separate tab is a better approach.

[michaelsshaw]

I am not sure how any of these, with the exception of the last, help anybody with anything. The game is supposed to be a very easy way to develop Source 2 mods, and all of these have a very anti-developer connotation, which sort of goes against everything I thought the game was about. Every single one of those changes makes it quite a bit more difficult to actually develop things and does absolutely nothing for the non-dev user.