Archived Official Modmix Distributions

ReleeSquirrel

New Member
Reading some of the comments on modding, and how much should be in the 'base game' and how much should be added in mods made me think about some friends of mine. They don't like modding games. Some won't mod games at all, and some won't mod games untill they've 'finished' the base game, and often by that point they've stopped having interest in the game anyways.

If Terasology's core game is very feature-light, that's what a lot of people are going to go to straight off the bat, and that's what people will judge the whole thing on.

When Terasology is published for consumption, with various choices for mods, I think that we should publish several 'official' builds, all presented and delivered the same. That way people will have confidence in the different mod sets, instead of thinking of them as random things jammed into the otherwise 'pure' game. It'll also help these mod-wary folks become more comfortable with the idea of making their own mods.

One way to do this is by offering Terasology with different mods preloaded for download. Conversely, if we distribute all of our 'basic' content with the game, but the player can toggle it on or off at world generation, there should be several preset buttons in the world generator, so that players don't get frightened by the otherwise seeming 'advanced' options.

What do you think?
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
You're already right on target - we aim to provide several "game modes" in the primary download each of which will essentially be a combination of mods that work together to provide a certain form of gameplay :)

Modding terminology is sort of tricky as we're naming every single bit of content as a "mod" - but a wide array of said content is considered core content anyway. I suggested a goofy series of made-up terms (or just the concept, the exact terms don't matter) to better differentiate between the different "scopes" of mods, but I'm not sure that's the way to go ;)
 

Skaldarnar

Development Lead
Contributor
Art
World
SpecOps
Another important point is to hava only a few central »mod spots«, i.e. here on the official forum, and then bundled in the specific (country related) community forums.

I would highly appreciate a package system for mods (similar to linux package systems, I like them because its easy and uncomplicated to (un-)install things. This system could be integrated in the launcher (or just ingame, another submenu ...), providing easy access to to mods...

I'm not quite sure, but it should be possible to maintain a databes with mod names, an url and a command (i.e. a github repo of a mod and the git clone command to copy the files into the local game directory) in the simpliest form.

But that are just my imaginations, let me hear what you think about it. And of course the would/could be »basic configurations« of pre-installed or featured mods (they can vary from time to time to support modders and stuff)
 

ReleeSquirrel

New Member
A package manager can be good but there has to be a way to distribute stuff outside of it. Otherwise, we'll have to add people's mods all the time, and the system will get flooded with crap people don't want. Even with a search feature, then you need to know what you want ahead of time.

But then, on the other hand, what if you want to join a server and they need mods? There should be a way to automatically transfer mods. Ideally from a tertiary server, so that people's individual Terasology servers don't get bogged down updating clients.
 

Skaldarnar

Development Lead
Contributor
Art
World
SpecOps
I thought about coupling it (automatically) to the forums, reading the header of a post (like the incuator thread header) containing all the necessery information. This way, a mod just needs a forum post and nothing more, different forums mimic different repos/mirrors and mod dependencies could be handled via the database...

Enhancements of this feature could include categories, related mod hints, dependecy and conflict checks, ... and of course a mod manager. And with the package system the server could just say I need package "mod-xyz" and the client can query the mod location and install command form its mirror (i.e. the forum).

Of course users can still install mods from other sources, but than without any guarantees (i.e. checks mentioned above) or something.
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
I'm somewhat attracted to the idea of eventually hosting each mod in its own GitHub repo, akin to how the Jenkins project does it (might seem scary at first, 843 repos?). Forum is probably too fragile, and we're hosting it ourselves so if Terasology suddenly gets popular (more Gronkhage) it could get very fragile even on our new beefy server ;)

I would very much like to connect much more officially with a modding community than it has been done for MC. Well, imagine for a moment that Mojang did connect seriously with the modding community :p Rather than just rely on an assortment of vastly different sites (or even just a forum thread, yuck) prepared by individuals we could focus on a series of officially approved or even hosted mods under our Organization account on GitHub, built by our Jenkins with files re-uploaded to GitHub for downloads (saves us bandwidth and is more stable). In-game / launcher can link to and pull from there.

Anybody who'd like to keep their own GitHub and/or Jenkins can do that too, just as long as it follows the right conventions. You can then add another repo definition in the launcher/game where mod updates can be pulled from - just like a package manager. Finally you can of course also just download a jar file from some random site and add that to your local install manually, but then there are no official guarantees or support involved.

You shouldn't need to interact with Git at all to grab mods, even in code, I'm just talking about the Downloads section on a GitHub repo (which I believe can be updated via API). Unless perhaps you want to run the game and develop against source from an assortment of mods.

That's not a small undertaking, however, and we shouldn't enforce too much closeness with a modding community so we become a bottleneck instead (official approvals / creations of GitHub repos / Jenkins jobs, etc). The Jenkins project has a nice IRC bot setup where admins can simply tell the bot to create a new GitHub repo, make people admins there, etc, that seems like it would be very handy and light-weight. It could also be a way to not lose great mods if an author goes inactive as we would then have the ability to "bless" a new curator, even if temporary, with full source access. Each mod would have its own issue tracking system, wiki, etc, and we can automatically parse all that for a separate site (like a modding section of our own GitHub wiki)

I've always thought hosting MC mod info in a single thread in an unofficial forum was an awful setup (might be stating the obvious there). I would still love to put a modding forum up here (probably two - accepted & active vs everything else) with a thread per mod for discussion and some spiffy hooks to GitHub/Jenkins, but don't think that should be the primary source for downloads, bugs, integration, etc. It certainly has its place though, since trying to scroll through a single list of 800+ repos on GitHub is probably not really the best way to find a specific mod :)

I'm sure not all authors would be thrilled with going through official infrastructure, some probably like the idea of using adf.ly for their downloads, but I expect we can meet the majority somewhere in the middle, like using their adf.ly links in generated lists that just go to the GitHub downloads. People can skip that, but then you can skip adf.ly pretty easily too.
 

Skaldarnar

Development Lead
Contributor
Art
World
SpecOps
That sounds good, but we still need some kind of data base if we want to organize available mods somehow... for example for dealing with dependencies and/or conflicts betweeen different mods...
Would it be possible to run just the database on our sever, from which the client retrieves the information where it can actually get the mod? So, for example, one mod is hosted on github in its own repo, the client could just clone it and put in the right directory, another mod is accesible via adf.ly, so the clients opens a web-broweser-like windows, displaying the ad, and the downloading the jar or something...
If the database contains meta information for the mods one can search via the database... One attempt to save traffic would be to keep a local copy of the database, and just pull updates from the server and do the search locally...
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
We could probably survive that, yeah. Although I'd think we could maintain a file on GitHub that describes what is available, with dependencies, etc, and leave a copy with the game downloader. Then on startup the game could check vs. that file online for updates.

Then the infrastructure building mods just keeps that file updated. Same thing for external "package repos"

A full DB might be more useful if you were trying to review mods available outside the game through some sort of mod searching site?
 

Skaldarnar

Development Lead
Contributor
Art
World
SpecOps
Yes, for example... I think a website can give you a much better overview over complex projects, composed of many single mods... and you don't have to be ingame to search for new mods.
But all that are future plans... for the beginning a working launcher reading a single file on GitHub would be enough... we can switch to a DB sooner or later if need to...
 
Top