Version 1.0 is content-complete

Started by Tynan, October 01, 2018, 10:29:21 PM

Previous topic - Next topic

Kirby23590

Are going to keep the Alphas and the Betas, Tynan?

I hope you don't remove them from the steam betas, since some people are still playing in B19 or B18 & have built large colonies that they built so hard for hours and days in your game, they don't want to loose their hard work they have built in. :(

One "happy family" in the rims...
Custom font made by Marnador.



Tynan

Vanilla savegames should be compatible A18->B19->1.0 and possibly from A17 as well.

Let me explain why I'm requesting that modders not update in-place.

So I never felt it worth worrying about how modders updated before. Every update there was some confusion over mods, but it wasn't that big a deal.

However the B18->B19 update was very rough. There were a lot of permanently broken saves (without warning), a lot of anger, a lot of angry messages to me (justifiably!), and a month-long period of instability as players had to repeatedly re-check which mods are and aren't updated and update their mods list.

The size of the community has grown. Most players will never interact with any dev or modder in any way, but are still being affected. When it's 1,000 players, that's one thing. When it's 100,000 players, that's something else.

This is why I figured I'd ask modders to update using a method that, while slightly less convenient in some ways, significantly reduces the suffering in the world by guaranteeing that nobody's game gets trashed out from under them.

To me that's a solid expectation I have as a player: If I'm playing a game, it won't just get trashed one day with no recourse. So I don't want to violate that for RimWorld players.

I think the benefit of being able to play historical versions with historical mods is also significant.

It's worth noting that many modders already keep separate versions like this on the Workshop.

I do wish Steam had better support for all this. But they've not seen fit to address my requests on the subject.

Quote from: Fluffy (l2032) on October 02, 2018, 03:47:43 PM
I've always kept my mods in sync with the current 'main' release of RimWorld on Steam. I believe that for the vast majority of players, this is by far the best solution; their mods will continue working on new versions of RimWorld without any action on their part.

This works fine for people using one mod. But many, many players use many mods, and for them it sadly won't work this way. e.g. A user with 30 mods will have his game rendered unusable the moment Steam auto-updates to 1.0, unless every single mod author of all those 30 mods instantly updates (and solves all inter-mod compatibility issues).

The B18->B19 experience shows that this isn't going to happen. In fact it took several weeks for the ecosystem to stabilize, and I was receiving angry reports of peoples' games being trashed for a month after. This is an experience I want to not repeat.

Players would go back to the old beta branch, game broken. They try the new version, game broken. It's awful for them.

One mod may keep working, but there is no chance that a player with 10+ mods will be able to keep playing without any action on their part. They'll have to struggle, wait, search through forums - and possibly never be able to continue at all if any given mod doesn't update perfectly.

QuoteYes, this will break save games for some people. I'm sorry, but that's on you for a) changing the API (seriously, moving Translate?!), and b) using Steam Workshop to begin with (the lack of version control is laughable).

Regarding the API, we actually put a lot of work into making sure that 1.0 code required no API changes from B19. I considered this a requirement. Translate is an extension method and we always use it that way, so it seemed here that the API worked without changes. It just never occurred to us that people were using it another way. I'm sorry about that.

(Translate had to be significantly refactored to handle other languages noun gender and other such complexities - it's a much more powerful method now. This was a major project for 1.0. But we did intend the API not to change. In any case this change seems relatively minor - a single find/replace, I hope.)

Even if we didn't change the API, the binary interface would change and compatibility would still break. So keeping the API the same makes updating easier, but it won't make mods just work without recompilation.

I agree Steam Workshop is awful with version control, but you can't get angry at me for using it. Many, many players get great value out of it. It's so much more convenient than manual installs, I think mods mods would get much less use if it weren't for Workshop. So in that way it enriches the mod community.

QuoteI'm already hard-pressed to maintain interactions with the people that use my mods, with discussions spread out over the forums, a dozen or so steam pages and my GitHub repositories. I'm certainly not going to spread the discussion out even more. I'm not going to lose my workshop item stats and create extra work because some people wouldn't take the effort to fetch old version of my mods from alternative download locations.

I didn't think it was any significant extra work. Just copying the mod directory and changing some metadata. You're certainly not obligated to keep visiting discussions about old versions of your mods.

QuoteI'd also like to add that your tone in this has seriously pissed me off. I understand you're under pressure for a good 1.0 release, and I'm aware that mods are a pretty big chunk of RimWorlds' appeal. That should be all the more reason for you to treat modders with a bit more respect.

I certainly never meant any disrespect. I didn't interpret this as asking for anything that adds any noticeable difficulty for modders. The goal is simply to not irrevocably trash thousands of peoples' savegames and thus ruin their day and piss them off. It's about serving players better; I thought this would be desired information.

Never imagined it'd be interpreted as some kind of burdensome demand.

You'll understand why I'd make that recommendation, right? The calculation is: We did it the old way before and it demonstrably created huge problems. If we do the same thing, we'll get the same terrible result. But there's a different way that would benefit the 100,000+ players at a comparatively tiny cost to modders. It makes sense to request that modders use a new method when the old method was so awful. (And many modders were already uploading separate versions of their mods anyway).

QuoteI wish you a great 1.0 release

Thanks!

QuoteI'll keep releasing my mods like I always have.

Well, I hope you'll reconsider. Your mods are great, lots of people enjoy them, and irrevocably trashing all those games-in-progress would hurt all those players. It seems unnecessary when there's such a good alternative that many modders already use.

However what I wrote is simply a request. I'm not going to police this or start banning mods that don't update or something. In truth I never expected anyone would have an issue with it given the cost is so tiny for such a huge benefit. But, I've made mistakes before. I should've explained the reasoning better the first time.
Tynan Sylvester - @TynanSylvester - Tynan's Blog

Tynan

Quote from: Kirby23590 on October 02, 2018, 11:11:18 PM
Are going to keep the Alphas and the Betas, Tynan?

I hope you don't remove them from the steam betas, since some people are still playing in B19 or B18 & have built large colonies that they built so hard for hours and days in your game, they don't want to loose their hard work they have built in. :(

I'm not removing any historical versions. While vanilla play is compatible between recent versions, modded play won't ever be and I'd like people to be able to finish their 100-hour game without it getting trashed!

I also think there's a lot of value in being able to go back and play old versions out of historical interest.
Tynan Sylvester - @TynanSylvester - Tynan's Blog

Kirby23590

Quote from: Tynan on October 02, 2018, 11:24:07 PM
Quote from: Kirby23590 on October 02, 2018, 11:11:18 PM
Are going to keep the Alphas and the Betas, Tynan?

I hope you don't remove them from the steam betas, since some people are still playing in B19 or B18 & have built large colonies that they built so hard for hours and days in your game, they don't want to loose their hard work they have built in. :(

I'm not removing any historical versions. While vanilla play is compatible between recent versions, modded play won't ever be and I'd like people to be able to finish their 100-hour game without it getting trashed!

I also think there's a lot of value in being able to go back and play old versions out of historical interest.

Whew!

That's actually good to hear! Thank you, Great Tynan! ;D

One "happy family" in the rims...
Custom font made by Marnador.



RemingtonRyder

#49
If a Workshop item is really popular, it doesn't matter if you put a new version of it on Workshop. It will regain its following in a relatively short period of time.

As you may know, I have always released separate versions. The user experience of this (at least from the ones who comment, I'm not a mind reader after all) has been very positive. Occasionally someone lands on an old version and posts a comment asking when it will be updated. I can't be everywhere, but I do get the word out when I put a new version up so that people have a decent chance of hearing about it.

I'm inclined to go with what the users of my mods seem to favour.

Honestly, Tynan, Steam Workshop is really showing its age. We should be able to have like a mini-hub which connects related Workshop mods. Instead I have to copy and paste markdown code like a chump so that people can navigate between versions. You can quote me on that next time you have a chat with Valve. ;)

YokoZar

Tynan, couldn't you make some way to put backwards compatibility into the mods themselves?  Like a 0.19 subfolder inside the mod dir that, if detected, would be loaded instead?

That way a mod could be compatible with multiple versions and things would just work, if the mod author enabled it.

UncleIROH

I made a new steam page for my mods b18->b19 and I regret doing it for these reasons:

  • The discussions / comments do not transfer.
  • The stats start over from zero and that makes me sad.
  • The search function for steam doesn't filter only current version by default, so they show up twice now which is confusing. Some people have resorted to putting the game version number for their mods in the actual thumbnail picture.

I also think if everyone makes a new page as you suggested, it will make the 'popular items' and 'most recent' views less useful.

When rimworld updates I'm going to try the following instead, and see how it goes:

  • Update in place my mods so they keep their same steam ID#, stats, comments, discussions
  • Delete any versions that are too old ( like a b18 version for example )
  • Make a new steam page ONLY if it's requested, but the new steam page will be for the older version of the mod.



Dingo

I don't think I have ever had comments saying "this broke my save game". They're ALWAYS "when will you update???" because Steam users expect a mod to auto update in-place much like their game auto updates in-place. I keep previous versions on GitHub where anyone savvy enough to opt into a previous alpha is likely enough to find them and I link those in obvious places.

I'm sorry if you've had outlash from users regarding B19, but let's face it B19 was marketed as 1.0 for the longest time and got suddenly released as another beta. I feel you're underselling how annoying this was to modders and mod users alike.

Tynan

I think it's largely a matter of point of view.

Considered purely from a modder's point of view, doing destructive updates seems better because it concentrates discussion and stats and makes the update itself easier.

But from the player's point of view, destructive updates mean your savegame is destroyed with no way to recover, without warning. That's perma-ragequit material.

I try to balance the interests of players and modders. Considering the huge benefit to players and relatively small cost to modders, simply requesting modders do non-destructive updates seems like the right thing to do. The alternative - just staying silent and letting everything break - seems wrong to me.

As I said I'm not enforcing anything. At the end of the day it's your mods, your choice. But I'd be doing wrong if I didn't even ask.

But I really hope people won't do destructive updates just to keep old stats around. You can count your "score" as the usage of all your versions together, they don't need to be counted on any particular page. I love watching numbers go up but just consider the impact on people who have their 50-hour epic RimWorld story trashed and forever uncompletable.

I hoped this stable pre-release people would be helpful. Previously we've released updated without warning, which is easier for development. Hopefully this pre-release period means that modders won't have to check constantly for a game update like before.

I do say this a lot (I am Canadian after all) but I'll also be really clear that I'm extremely grateful to everyone who contributes anything to the RimWorld community - modders, comics makers, people who post screenshots or stories or bugs or feedback. I try to make a sense of gratitude the foundation for all my communications with the community at large.

Quote from: Dingo on October 03, 2018, 01:21:21 AM
I don't think I have ever had comments saying "this broke my save game".

I think it's a point-of-view thing. People blame the thing that changed last. When a game update happens and a save stops working, people see the core game update as being the cause. They wouldn't think "my game is broken because modders didn't update yet"; nobody expects modders to all update the moment a new game version is released. They're really pissed off, but you don't hear about it because they all complain to me instead.
Tynan Sylvester - @TynanSylvester - Tynan's Blog

Antaios

I think you're not seeing a lot of compassion for steam users, because, frankly, they suck. They're largely the ones who complain the most and without reason, the ones who don't ever bother to check anything. It's hard to care about people who don't put any effort in.

Most people modders will actually care about are people who accept that when they add mods to their game, they're taking on responsibility, and accepting that there may be issues. They're taking on the mantle of debugging their own game and at least putting an even small effort into sorting out problems before complaining, or rather ideally, asking for help - and when they do ask for help, they provide some relevant information. The vast majority of this forum is like this.

The average steam user, just doesn't do that very often - you can't please them, so why try?
This:
Quote from: Tynan on October 03, 2018, 02:27:02 AM
I think it's a point-of-view thing. People blame the thing that changed last. When a game update happens and a save stops working, people see the core game update as being the cause. They wouldn't think "my game is broken because modders didn't update yet"; nobody expects modders to all update the moment a new game version is released. They're really pissed off, but you don't hear about it because they all complain to me instead.
is indicative of that.

If someone can't be bothered, when they probably know an update is coming because it's in the news for the game which shows up on the library page, to be ready for the update of an early access game with mods - be that locally copying the mods, or finding previous versions on available hosts that are almost always linked in the description of the mods, or just plain finishing their save up before the update hits - then it's really hard to give a damn.

I'm not saying that's fair, but it's the kind of issue that the click-and-go style of modding that steam workshop has created.

RemingtonRyder

You know, what really irks me about Steam Workshop is that it's clearly possible for it to deliver a version controlled mod to the user - because right now, a modder can actually roll back a mod to a previous version from the item's change notes. And yet, when we could really really use that functionality, it isn't available to users (yet).

Thanks Valve. Really appreciate you listening (not).

Mehni

The tedium of having to go through my list of 200+ subscribed mods to hunt down each new version on the Steam workshop kills my enjoyment of the game quicker than a lost save. "Why are there 3 versions of this mod? Which version do I need? Is this for the B19 version of 1.0 or for 1.0 version 2?" Take note of that last question: There are mods tagged 1.0 that won't work for 1.0! The update from B18 -> B19 was rough alright.

Players can make a back-up of their Steam mods if they wish to continue on outdated game versions. This is a copy-paste operation which requires little effort: all it takes is some preparation. A lack of preparation on their part does not constitute a problem on my part. It would be grand if RimWorld did that on behalf of the player prior to an update, but that's another discussion.

I'll be updating my mods in place. I'm not going to clutter the workshop or make hundreds of thousands of people go on a hunt for mods. Sucks for those that lost an "epic 50-hour save", but from my point-of-view it's the better system: consider the huge time-saving benefit of an automatic update for all those players using more than a dozen mods. Saves will break whatever happens. Plenty of modders won't read these forums, or use the new release to finally implement that breaking change.

Alenerel

Quote from: Mehni on October 03, 2018, 04:28:07 AM
There are mods tagged 1.0 that won't work for 1.0! The update from B18 -> B19 was rough alright.

I think that Tynan should clean those mods. Either take them down or just untag them all. The old 1.0 (pre B19) is not the new 1.0. It should be easy, anything tagged 1.0 before a few days ago walks.

RemingtonRyder

So... I just tried out Fluffy's mod manager. It looks like a pretty good way for people to stick with a particular version of a mod. Made short work of backing up the mods I had downloaded from Workshop. Gave it a thumbs up.

If we could somehow get the entire playerbase subscribed to Mod Manager and get them to back up their mods, then saves wouldn't break.

Tynan

#59
Quote"Why are there 3 versions of this mod? Which version do I need?

Steam workshop has tags, if players just click the appropriate version tag all the non-matching mods disappear. Beyond that I recommend modders put a [B19] or [1.0] tag in the title, as many do already.

Quote from: Mehni on October 03, 2018, 04:28:07 AM
Is this for the B19 version of 1.0 or for 1.0 version 2?" Take note of that last question: There are mods tagged 1.0 that won't work for 1.0! The update from B18 -> B19 was rough alright.

This won't be an issue. I'm checking everything marked for 1.0 to get rid of the old ones that aren't updated. There are just 100 such mods - not too hard.

In my pre-release announcements I will be recommending people copy/paste their mod folders. But the great majority (>90%) of players don't read news like that. So most players will just get their games trashed without warning, if any mods are updated destructively.

Another thing with in-place update is that it's impossible to put the mod up during the current pre-release testing phase. It might be useful to test that way. Then, when 1.0 is marked final, the 1.0 version of the mod becomes final too.

Anyway, I'm surprised at the refusals. In-place updates don't add convenience for players because many modders won't do them promptly or ever. If half the mods are updated in-place and half get new versions, now the old save is broken AND players need to hunt down and figure out new mods and deactivate the old broken ones - while the mods are changing day to day. It seems much, much more straightforward to simply guarantee that a mod will always work for the version it was created for.

E.g. Finish your B19 game on B19 branch, then on starting a 1.0 game, select what mods you want that work and use those. The delay between 1.0 release and starting a new game adds some time for mods to get updated. As opposed to just trashing everyone's game and demanding they start anew in an environment where mods are changing daily.

From player's POV, both approaches have downsides, but the destructive approach's downsides seem universally worse.
Tynan Sylvester - @TynanSylvester - Tynan's Blog