Discussion: Uber's plans for Modding, 1.0 and Beyond

I was just saying that to draw the connection between what SC2 does well and what could be in PA. They have a big flashy button for modded games right there on the multiplayer menu (called arcade), of course all of their UI is pretty flash, so...
I think 'modded game' should be right next to 'vanilla multiplayer' (I want to avoid calling it multiplayer, because modded games are multiplayer too!) And then of course skirmish below those two. Once 1v1 and 2v2 and all that are installed, just add everything under 'multiplayer' with different flashy UI buttons from the home multiplayer page. That's kinda what SC2 does, and I think it attracts people well.

Hi everyone,

Even though I'm still deep in the trenches right now, I figure this was as good a place and time to give a general sense of what I'm working on right now modding-wise.

First, I'm not going to be talking about time frames or specific version numbers in here (1.0 or otherwise). Neutrino has been very vocal about not making those kinds of commitments, and I agree with that position. Making games is hard, and every game constantly makes trade-offs and reassessments behind the scenes during its development that most people don't get to see. What we're doing here is not that much different from most game projects I've worked on in my nearly two-decade career from that perspective, we've just simply been more willing to talk about it here. With that said, understand that everything I'm about to say is speculative, based around our current vision and plans; we'll only know how real those speculations were when they're in the rear-view mirror.

The modding platform/ecosystem has several pillars that we're thinking about, but two of our current main concerns are Publishing and Scripting. These are reasonably independent of each other, but both are important. There are other cross-cutting pillars too, like Security and Debugging/Diagnostics, but given that they're both open-ended and also dependent on how much the first ones are embraced, they take a natural backseat to Publishing and Scripting.

The "Publishing" pillar is concerned with sharing of mods, mod discoverability and ease of download, persistence of frequently-used mods, etc. The "Scripting" pillar is concerned with adding mod scripting APIs to make new game types possible, deep script-based inspection of game state without excessively hurting performance, and so on.

Scripting is more open-ended than Publishing; we can keep adding mod APIs until the cows come home (read: we run out of money or there aren't enough people buying the game anymore to justify additional engineering effort). Publishing is somewhat open-ended (especially on the UI side) but less so; there is definitely a set of core functionality that we're thinking about there.

Originally it was just me staffed on this whole modding thing, and since Publishing was the least open-ended of all these various issues, it made the natural place to start. I originally figured I would try to get the publishing situation to a reasonable state, and then move onto more fun gameplay-related scripting stuff. Originally I thought I had a pretty good idea of what we'd need for the publishing infrastructure, and Ubergaf was also helping out on the PlayFab side, so I figured it'd be a couple weeks of work on my end for the game client and server work.

My estimate was not even remotely in the ballpark; what I originally thought was going to be a couple weeks is now turning into a couple months. Like I said, I've been working in games for almost two decades now (18 years specifically), I feel I'm a pretty competent engineer at this point, and my estimation skills have gotten pretty decent over the course of time... yet even then, sometimes I still get broadsided, and this definitely broadsided me. Before I knew it, I had to start re-assessing some core architectural decisions I made when I first built the transient server mod architecture that people are currently using... decisions that I felt were correct at the time, and which I still think were the right decision to start with even now, because it got some basic server modding capability into the community's hands sooner (which helped us answer some other questions that we really didn't want to wait on). I might have made different architectural decisions earlier had I spent more time thinking about the publishing questions, but then when would you folks have gotten any kind of server mods at all? I have no idea; that's an alternate timeline that I don't exist in.

So I have no regrets, however the fact is sometimes things are just more complicated than you think they're going to be. Game development is filled with stuff like this, and it often results in features being delayed or cut to get a product out the door. Since modding is an important thing for PA in the long term, it's not getting cut, so now we're just trying to work our way through the delays.

One way we're working through it is that another engineer, Jorgen, is going to be looking the Scripting / gameplay modding API issues first, instead of me. I was looking forward to working on that stuff, but the fact is I made a bad estimate on the publishing work, so it's only fair that if I'm going to hand off work to somebody else, that it be the part that's more fun. So he'll have more to say about that pillar after he's wrapped his head around it.

For my part, right now my head is filled with publishing-related issues like these:

* Allowing you to publish completed mods to PlayFab, so we can host them on our Amazon-backed storage instead of forcing the host client to upload them every single time they host a game.
* If a host client has several mods mounted, we need to make sure that transient (host-client-provided) mods and PlayFab-provided published mods can be intermixed and arrive at independent times (this problem turned out to be a Pain In The Rear).
* Allowing the server and all clients to download their mods from independent sources (e.g. PlayFab links to Amazon Cloudfront distributions) and still synchronize their mounting capability.
* Making it possible for the host client to change mod configuration while already connected to the server and other connecting clients (since we'll need this down the line for mod-related game options), triggering additional mod downloads as necessary.
* Caching of published mods on the local file system in a way that doesn't compromise synchronization with other clients.
* Getting the game to understand the concept of mod versions and version ranges (for expressing dependencies etc.)
* Yadda yadda yadda

This is all just for the publishing piece. The deeper you go into certain problems in game development (or software development in general), you more you realize that some of them are deceptive from the outside, until you get into them. This stuff is Not Easy. I haven't been spending much time on the forums lately because my brain is having a tough time handling the context switch; if I got too immersed in forum discussion while trying to do all of the stuff listed above, it'd never get done ever. This is tough enough to keep in my head as it is.

So when Neutrino talks about it being a miracle that the game is in the state that it's in right now, he's not kidding. We're all fighting the good fight, and we'll get some good stuff to you. But you gotta be patient, because this is difficult, and we're human.

I love these updates from the trenches that we get from some of you devs. Good luck to you and Jorgen!

Are you and Jorgen also working on things like the UI for the mod stuff? How a player can learn about and start playing mods? I get that how a mod is hosted and shared through the server is important to get smooth, but for me that's kinda the "last" step of the process and I'm honestly a bit more concerned about the the earlier steps which is part of what got this thread going.

I can surmise that all the stuff(and more I'm sure) are rather "Foundational" when it comes to setting up a system like this and that Modding is a "Pillar" for PA's long term success but like everything else, you still need a good "start" because it's a heck of a lot harder to make up for lost opportunities than it is to get them from the beginning IMO.

Mike

How would people feel about a reputation systems?
http://sloanreview.mit.edu/article/...ms-how-to-design-one-that-does-what-you-need/

Does this rating system include a down vote because that is easily abusable. Having user interaction causes some concerns such as how trustworthy are they? How could you know if the user is acting in a malicious manner? Is there a way to prevent or lessen the damage that would be caused from it? Potentially user interaction could damage integrity and credibility of the author or the mod in question. It is make or break on reputation.

@forrestthewoods

• https://forums.uberent.com/threads/pa-client-api-requests.57785/
  
• https://forums.uberent.com/threads/pa-server-api-requests.61182/

Also this one: https://forums.uberent.com/threads/mod-ideas.56072/

Having a split windows top "Normal" Bottom "Advanced" read approved not to crash your system mod games.
Then if there is a full set of filters on each window with buttons to minimize ether one making the other full size but with the split as default it will show both mod and normal equally.

Speaking purely form a client side perspective, what I would like:

• A mod marketplace in-game (like PAMM but internal).
• The ability to install, re-install, un-install and update mods whilst in-game. (I NEVER want to restart PA.)
• Descriptions for mods (in detail).
• Direct link from the main menu. (and advertise it to players who don't bother. 'Try our mods to improve your experience' blah.)
• Separation in marketplace between server and ui mods.
• List of installed mods for yourself, maybe for others too.
• User ratings for mods. (perhaps even reviews)
• List how many users have downloaded the mod.
• Show games with installed server mods by default. (clearly labelled)
• Automatically download and install mods when trying to join a game with mods.
• Filter by specific server mod. (Only want to play matches using navalreduxwhatever??)
• NO FINANCIAL TRANSACTIONS

Financial transactions done properly, is just more choice for modders.

It already automatically downloads servermods, doesn't it?
And I could be wrong, but I am pretty sure that you can already filter by specific server mod. I have seen servermod names in the filter dropdown menu for mods.

Someone is going to make their mod cost money, this has been the plan for a while.

If he makes his mod cost money then less players will use it, but people will pay. I can see total conversions cost money, but the point of "what is enough? " is very subjective.

Technically no, but the server does! The server downloads the mod, and then hosts the game the same as it would any other, only with the mod's info, this means that you don't get the files after a game ends.

It's a free country (well at least some are ), people can do what they want. But I don't want to see their mod advertised in-game.

I don't know, wasn't a list of changes but a feature wishlist.

Hm? Why would it be advertised? I'd suggest only popular mods be advertised, mods with the most downloads and such.

Being in a list, inside the game, where people can look at it, and get redirected and have money stolen off them. Is advertising.

No mod handled by the marketplace should have fee's imposed, no click to buy crap.

Your view of mods is interesting.

I don't have a problem with people using their modding talents as a business selling a product, but not off the back of what should be a free service. It's sets a bad precedent.

How is money being "stolen" if you pay for something, you pay for it.

Something that takes them months, and months of work. Something that completely changes the game. I think that's worth money.

Total conversion mods, mods that took a lot of effort to create- some deserve money.

What I'd imagine is that there'd be three types of mods, completely free mods, a mod asking for an optional donation, and paid for mods.

To sell a mod, you need ownership over it, which you don't have. Uber own's it, so you have to cut them in, and they have to police the market place to ensure no one rips you off by taking it and offering it at a lower price. The entire marketplace soon becomes full of chargeable mods, honestly who would make it for free when all they have to do is give Uber x% and boom they get to go to college tuition paid. (or whatever, buy a house).

One mod goes onto sale backed by Uber and their new Mod Police and then begins a vicious cycle that leads first to the death of free mods and then to the death of good mods, and finally the death of the game itself as interest falls away.

Its a stupid idea and I wish to **** off and die.

If you want to sell your mod, then ask uber for permission to copyright it and advertise, sell it and protect it yourselves externally.

PA was supposed to be a platform for other developers if I remember correctly.

I can't wait for the populous game, that will be wikid.

Why should it be free? I don't like this kind of thinking. You basically say:
"I don't want to have payed mods because I am afraid that works and will result in me having to pay for more stuff in the future."
Personally I think that if it financially works out it is a valid business mode and that's fine.
I also don't think modding will die because of it. Quite the contraery: I think it will get more people to mod stuff.
Creating mods is much harder than using them, so you need to focus on making the act of creating them more attractive.

I don't like the idea of paid mods because to me mods were always labours of love. Done because someone wanted to make a difference to something they enjoy and share it with others.

Inserting money into that system just seems.... wrong, kinda.