Cut all maps but the most popular one

[captain sawrge]

Bottom post of the previous page:

+1 to op, meta isn't my favourite and it'd pain me to see box go but absolutely I agree with the sentiment

[Denton]

- Promote SEVERAL map maintainers. Denton is good. I'm good. MMMircacles is good. Most people are good, actually. Promote anyone who shows a tiny bit of mapping competence.

- Allow map maintainers to instantly merge their own commits on maps that aren't Meta. The PR system is an unnecessary hoop that makes map changes more painful than need be. Additionally, the PR system is the reason the map merger is required. If commits could be merged instantly there wouldn't be a need for the map merger, so kill the map merger on everything but meta.

- People adding objects should only required to update Meta, but for each map they don't update they have to tag map maintainers to let them know. People doing path changes or removing items are required to change all maps, however.

I agree - having more map maintainers would help tremendously.

Mapping doesn't take terribly long once you're used to it - IMO the whole thing where PR A blocks PRs B, C and D until it's merged slows down mapping progress by a lot.

[Qbopper]

I'm not a fan of removing map rotation for a couple of reasons but I'm at work so elaborating on phone is hard

I think pubby has the right idea in promoting those who understand how to map and there has to be some kind of solution to fix the problem that isn't "just remove map rotation"

Sure, we don't need 7 maps or whatever, but do what pubby said please, just removing all maps instead of trying to fix the bottleneck (github) seems like the wrong approach

[Alexch2]

rustle is lazy so im gonna post qbob's point thingies for him

shit we need:

-remove maps that aren't being maintained
-assign a "map maintainer"(?) to their map they work on
-allow map maintainers to be tagged when someone makes a change that requires editing maps (if the pr requires legitimate map editing and not "put this object in a locker on every map")
-allow map maintainers to merge map prs so the changes can be made quickly

I have no idea how feasible this is but I think it would solve the problem




My (Alexch2) own thoughts are that 3 maps is a good amount.

[Qbopper]

also would like to follow up on what I said with a disclaimer that I've been chipping away at budgetstation on and off for a few months now and it would be extremely disheartening to have all of my work just dusted because of this

I think at the very least we should try something along the lines of what I said before going nuclear

EDIT: also for the record the copypasted post above was from discord so that's why it's essentially what pubby said

[Wyzack]

I mean the obvious solution that should appease both parties is to do both at once, yeah? If we cut all but the top maps unless someone steps up to maintain them, we will see pretty damn quick if anyone is actually willing to do so

[Qbopper]

I mean the obvious solution that should appease both parties is to do both at once, yeah? If we cut all but the top maps unless someone steps up to maintain them, we will see pretty damn quick if anyone is actually willing to do so

yeah I mentioned that

I don't think anyone would object to cutting maps that aren't actively maintained

[MMMiracles]

Removing map rotation would be fucking horrible and bring back the staleness that comes with only playing one map, but having 7 maps is probably too much. Leaving the combination to something like Meta/Delta/Box would probably be the best combo for medium/high/low population situations while having some variety.

I'm not sure I share pubby's sentiment on the difficulty of the map merger since the most 'difficult' part is just setting up python properly and remembering to prepare/merge the map files before/after respectively, but having people required to only update the main map while having maintainers to push fixes through for everything else would probably help speed up the process quite a bit though when it comes to actually keeping maps maintained.

[Bombadil]

I vote we axe Box keep Meta and Delta.


So we have one lowish-medium pop map and one large map for huge playerbase

[captain sawrge]

I don't see the staleness, I played on the same map for years before meta was even a thing and while there's a novelty in a new map it's never been about the map, it's what happens on it.

All these workarounds or compromises still seem short sighted and present the same problems outlined in the OP, albeit on generally smaller scales.

The problem isn't that x map is not maintained or whatever it's that any new feature that changes the map must find a way to incorporate 3-7 map changes in a manner everyone is happy with. You can have "map maintainers" but then any major feature is stalled until they can react to it which still impedes development.

I'd rather play a changing game on a stale map than play a stale game on a (limited) variety of maps.

[Nabski]

Qbopper, I think you budgetstation and you falls into oranges "if it has a dedicated maintainer, that keeps it up to date it stays in rotation".

The key thing here is that you would be responsible for adding features such as a circuits lab to the map, rather than whomever makes the feature.

[CPTANT]

our feature agility is seriously harmed when you have to do the same boring ass shitty reptitive mapping actions on 7 maps, most of which you'll rarely ever play on.

I am seriously concerned that this is impacting our ability to introduce new features and remap existing departments.

Solution, run a simple STV poll with all the map options on it.

Cut everything but the first place winner (probably meta)

and commit to having a single map again.

What 7 maps are in rotation? The wiki only lists 5

[MMMiracles]

our feature agility is seriously harmed when you have to do the same boring ass shitty reptitive mapping actions on 7 maps, most of which you'll rarely ever play on.

I am seriously concerned that this is impacting our ability to introduce new features and remap existing departments.

Solution, run a simple STV poll with all the map options on it.

Cut everything but the first place winner (probably meta)

and commit to having a single map again.

What 7 maps are in rotation? The wiki only lists 5

I assume he counts mining and the centcom z-level as maps that are kept maintained, although it seems like they'd be affected a lot less by feature updates and only need to worry about path changes.

[Qbopper]

Qbopper, I think you budgetstation and you falls into oranges "if it has a dedicated maintainer, that keeps it up to date it stays in rotation".

The key thing here is that you would be responsible for adding features such as a circuits lab to the map, rather than whomever makes the feature.

Yeah, I think the idea with what me and pubby said was giving mappers the ability to be 1. quickly notified they need to make changes and 2. the ability to make the changes quickly

by having that map maintainer role, and having people making the changes edit the "main" map, it shows the mappers how to implement the feature, and lets them update the map without waiting for a maintainer

[Dax Dupont]

Where did you get 7 maps from?

Are you counting centcom and runtime?

[onleavedontatme]

We had 7 (and possibly more) maps recently before we cut several of them, was just a ballpark number I was griping about when talking about having to move the vault several times over

[onleavedontatme]

We also have multiple people working on maps so we'll be back up to 7 again soon if there isn't some sort of limit

[Dr_bee]

We also have multiple people working on maps so we'll be back up to 7 again soon if there isn't some sort of limit

Would there be any way to design around this problem? For example using spawners instead of manually placing items?

If code changes can all be made in one place and then automatically be divvied out to maps, it would make maintaining them easier.

[TribeOfBeavers]

Items aren't a problem. It's features that require remapping areas or adding new rooms (the citcuit labs for example).

[Qbopper]

We also have multiple people working on maps so we'll be back up to 7 again soon if there isn't some sort of limit

Would there be any way to design around this problem? For example using spawners instead of manually placing items?

If code changes can all be made in one place and then automatically be divvied out to maps, it would make maintaining them easier.

The problem is that someone needs to place those spawners on every map

Also, unrelated, but if budgetstation actually gets finished and added, by nature of the map some things wouldn't even need to be added

[somerandomguy]

Items aren't a problem. It's features that require remapping areas or adding new rooms (the citcuit labs for example).

If we're using circuits as an example, it didn't actually need a dedicated room. Sure it does make it more convenient, but they worked perfectly well as a research item too.

[TribeOfBeavers]

Well I was just using that as as a recent example.

Adding any feature that requires a new room or a lot of space means you have to shuffle a fairly large area of the map to make everything fit (especially if you're trying to add something near the center of the station), you then have to make sure disposals, atmos, and the power grid all fit correctly.

This is a pain for one person to do on a single map, let alone five of them.

Even something like adding a basic 1x1 machine means you need to either remove something else from the room, expand the room, or have the new thing akwardly block other features of the room.

[Dr_bee]

Well I was just using that as as a recent example.

Adding any feature that requires a new room or a lot of space means you have to shuffle a fairly large area of the map to make everything fit (especially if you're trying to add something near the center of the station), you then have to make sure disposals, atmos, and the power grid all fit correctly.

This is a pain for one person to do on a single map, let alone five of them.

Even something like adding a basic 1x1 machine means you need to either remove something else from the room, expand the room, or have the new thing akwardly block other features of the room.

Does the machine need to be added? I mean you could just as easily give the crew the ability to build the machine via spawners, techweb lathes, and cargo. Features should be added to those three things and if the mapper wants to add them to the map, let them.

absolute separation between the code and maps! But not really. It is just no one is making you make new rooms or add the machine proper. There are ways to put new stuff in game without mapping it in.

[ohnopigeons]

Lots of potential mappers have been turned away because of difficulties with the map merger and git. they're too complicated for people who aren't coders.

Even I've been turned away with all the fucking steps involved with making changes to a map. Making the changes in DM is the easy part. What pains me is having to manage the mapmerger, then baby the changes through github, which is NOT a good site for handling maps. Take my recent map PR for example. It's been 10 days, three commits, and still no end in sight. Right now it's holding up 2 other PRs, which is awful.

Here's what it should be, regardless of how many maps get removed:

- Promote SEVERAL map maintainers. Denton is good. I'm good. MMMircacles is good. Most people are good, actually. Promote anyone who shows a tiny bit of mapping competence.

- Allow map maintainers to instantly merge their own commits on maps that aren't Meta. The PR system is an unnecessary hoop that makes map changes more painful than need be. Additionally, the PR system is the reason the map merger is required. If commits could be merged instantly there wouldn't be a need for the map merger, so kill the map merger on everything but meta.

- People adding objects should only required to update Meta, but for each map they don't update they have to tag map maintainers to let them know. People doing path changes or removing items are required to change all maps, however.

I really don't see the value of "map maintainers" as described here. I understand your concerns with PR requests lying open too long but this is fundamentally an administrative/bureaucratic problem and isn't something you solve by throwing more bureaucrats at it. You cite it as a way to avoid the map merger and PR system but you're throwing the entire concepts of version control and incremental development out the window in doing so.

I would rather see the github bot automatically merge map edit PRs after 24 hours since the last activity (assuming it compiles and map files are the only files affected of course).

- People adding objects should only required to update [main map], but for each map they don't update they have to tag map [owners] to let them know. People doing path changes or removing items are required to change all maps, however.

This is good though.

[oranges]

- Promote SEVERAL map maintainers. Denton is good. I'm good. MMMircacles is good. Most people are good, actually. Promote anyone who shows a tiny bit of mapping competence.

How is this any different to having someone that is owning a map

- Allow map maintainers to instantly merge their own commits on maps that aren't Meta. The PR system is an unnecessary hoop that makes map changes more painful than need be. Additionally, the PR system is the reason the map merger is required. If commits could be merged instantly there wouldn't be a need for the map merger, so kill the map merger on everything but meta.

This doesn't remove the need for the map merger, people are still going to have updates at around the same time, coders are still going to need to do path refactors when they change systems, they're still going to conflict, we still need a git friendly map format (tgm) and people deserve a right to provide some feedback to changes to the maps they play on

- People adding objects should only required to update Meta, but for each map they don't update they have to tag map maintainers to let them know. People doing path changes or removing items are required to change all maps, however.

Whichever map is the most popular is fine


Most of your comment is insightful but your comments on git/the process are unworkable.

[pubby]

How is this any different to having someone that is owning a map

Single owners aren't reliable. I'm the only person who still "owns" a map and I go away for months at a time. Besides, mapping isn't hard. It's not like coding where some people are significantly better at it than others. We should involve everyone qualified rather than pick favorites.

Mostly I want "unpopular" maps to be like a wiki, where anyone can edit them. Pigeon's idea of auto-merging could be a worthwhile compromise.

we still need a git friendly map format (tgm)

Yeah, you're right. TGM is unavoidable. I shall hate it still.

[Qbopper]

Besides, mapping isn't hard. It's not like coding where some people are significantly better at it than others.

we don't have enough mappers right now for this to really be the case but trust me this is still something that a lot of people would not be good at

[Not-Dorsidarf]

Besides, mapping isn't hard. It's not like coding where some people are significantly better at it than others.

we don't have enough mappers right now for this to really be the case but trust me this is still something that a lot of people would not be good at

Mapping is hard, just in a different way than coding.

[Saegrimr]

All these Box haters are fucking weird, people used to bitch and moan whenever Meta came up randomly.

[Qbopper]

All these Box haters are fucking weird, people used to bitch and moan whenever Meta came up randomly.

this is off topic but I think meta has an "easier" layout - a lot more is provided for you than box, so people like it because they don't need to, say, scour the map for a toolbelt or whatever

people originally hated meta because meta was different and not box so it was bad

[PKPenguin321]

people originally hated meta because meta was different and not box so it was bad

no it was because it was fuckhueg (and it still is, but i think we're more tolerant of it now ever since we had to deal with massivehallwaystation)

[oranges]

this is irrelevant to the question at hand

[BeeSting12]

We should keep the three high pop maps and can the rest of them IMO. Use box as a lowpop map if needed.

[EagleWiz]

Have you considered cutting back on the number of coding changes that would require changes to the maps? Lets face it, those are usually bad coding changes

[oranges]

Yes, and the belief is that will lead to stagnation and will kill the server.

It's not sustainable to do nothing.

[Naksu]

Ideally I'd like to see us ape the model some other open-source projects have adopted where we have our one true map (meta?) and the rest are "contributed", where in our case it might mean that contrib maps are not guaranteed to have the latest and greatest features. As a player, I'd very much like to see us having all the maps we currently have in rotation, maybe even a new one or two. As for repaths/var removals etc, as the author of several of those PRs the maps were generally less tedious than the code, though for actual non-trivial feature PRs i expect the opposite to be true to an extent.

[Qbopper]

it might mean that contrib maps are not guaranteed to have the latest and greatest features

this gives me the gut feeling that players would begin to complain any time the map rolled off the "official" map because the feature they like isn't in the other map, plus the already existing issues of "i don't like it because it isn't the map i like" and other (more valid) complaints

I think the map maintainer idea is the best of both worlds, coders have to implement their features once as an example for others and then mappers can quickly push updates to their own maps, or in theory even get permission from other mappers to implement the new stuff temporarily (or not) to get the update out there ASAP if someone is busy

[Cobby]

Tell people to stop making maps and improve the ones we currently have >:(


SEE BOX!!! I literally hate the glass tunnel maints, feel free to flesh them out.

[MrStonedOne]

I think the maps contribute a worthwhile level of variety to keep around and we could solve alot of maintainability issues with them if we actually tried instead of giving up. Placeholder landmarks could be used to enhance expandability, type re-parenting as a form of type aliasing could be used to auto-convert older type paths to newer type paths, (you could make a bot that automatically deletes type aliases after say, 3 months and use datum flags to mark alias types so the map loader can warn when using them).

And since maps aren't compiled in anymore, we don't even need to have all the maps in the repo like we do now. A few tweaks could allow for server side maintained map files with automatic fallback on load failure on top of the referenced ones. (i'm thinking auto search maps/* for the file, but if it fails to find it there, searching config/maps/* for the map file, and load the last map if the new selected map fails to load as well as notify admins, maybe add on a way to do auto-updating maps from an external source thats configured server side.)

The issues and growing pains we are having with maps aren't new, many games have had them, and we can look to their solutions to help out here.


I'm thinking limiting it to high/medium/low pop reference maps in the repo, and expanding the mapping system to support server-side map files, working on adding systems to mark expansion tiles to give limited future proofing, and using time-limited type compatibility aliases (with a system that warns when a map is loaded that uses them) to make maintaining these maps easier.

This pushes from both directions, makes it easier on the repo by tiering the support levels of maps so all coders don't have to maintain all maps, just the official ones, while making it easier for people who maintain maps to maintain their maps.

[Bluespace]

Can we at least remove the unpopular maps though?
Everytime pubby comes up, the population drops and admins force change the map.

[Mickyan]

What's wrong with pubby

I suspect a lot of people don't like it because they're not used to the layout and forcing map change the rare times it gets selected certainly doesn't help

[Rustledjimm]

When metastation was introduced everyone hated it, population dropped and people suicided.

People are stupid.

[Lazengann]

Pubby's missing a lot of jobs for one

[CPTANT]

Pubby is rather mediocre. It just doesn't have very much distinctive characteristics but is different enough to get lost in.

[Dr_bee]

Pubby is rather mediocre. It just doesn't have very much distinctive characteristics but is different enough to get lost in.

One thing I like about pubby is that the bar is actually in the center of the station so it becomes a natural hangout, something other maps really should emulate.

[Alex Crimson]

this is irrelevant to the question at hand

Too much of anything can be bad. Is a slower coderbase really so bad? Do we really need a constant flow of features, most of which end up being shallow crap that eventually turns into an unsupported, dead feature?

On topic though, if having so many maps is such a big issue, then they should be removed. Especially if nobody is maintaining them. Hell, id say we should keep it at 2-3 maps maximum simply because its fucking annoying memorizing new layouts.

[Cobby]

BMD is the only acceptable rotation and you need a break from the game entirely if you get that aggravated about the map layout but aren't willing to flesh out those layouts.

[Atlanta-Ned]

well this'll make renderbus a lot less complicated

[PKPenguin321]

When metastation was introduced everyone hated it, population dropped and people suicided.

People are stupid.

Not really, it was ran 24/7 on the alt server so when map rotation came in box and meta were still both the most popular

[EvilJackCarver]

I actually miss Dream Station.

[oranges]

When metastation was introduced everyone hated it, population dropped and people suicided.

People are stupid.

Not really, it was ran 24/7 on the alt server so when map rotation came in box and meta were still both the most popular

headadmins have a loose grasp on server history