Dealing with a community server

#1

There has been a ton of talk on doing a community server lately (way more than usual) and, despite my own cynicism, I still love the idea. Rather than going on with members who don’t care about anything but the features, I want to query MoparScape’s elite on what they think should be done in the realm of a community server.

As far as the discussion is concerned, it’s worth noting that most people want to start a new framework entirely. This is both impractical and technically beyond the scope of a community server. A community server focuses on the community collaborating in content development (at least, that’s specifically what I’m asking). We already have a server base to use–Osiris. If you feel that a new framework is warranted and perhaps an even better project than just doing content, feel free to state your case. I like the idea of promoting Osiris more, though.

What does everyone think about getting some stuff stirred up for a community server? How would you manage it?

#2

most likely the best bet.

#3

I was referred to that a while back, but I have heard that there are significant stability issues, especially compared to a server like Osiris. Have you looked at that at all yet, t4?

#4

I would say have the elite cook up a good starting base. Get down the networking, protocol, all that stuff… Implement a good scripting/plugin system, and allow the community to work from there.

I know that’s bare minimum, but I’ll try to think of some stuff to expand on while I’m at work tonight…

#5

[quote=“m.dog311, post:4, topic:443579”]I would say have the elite cook up a good starting base. Get down the networking, protocol, all that stuff… Implement a good scripting/plugin system, and allow the community to work from there.

I know that’s bare minimum, but I’ll try to think of some stuff to expand on while I’m at work tonight…[/quote]That is exactly what I was thinking. It’d be a really bad idea to let the entire community mess with the core, especially with how difficult it can be to write a very stable and fast server that can easily be destroyed with content. There should be ranks for working on the server. For example, some people would have full access, others can only work with scripts, and there are some who can work with some core files but not all (I’d imagine that be the case for things like minigames. I’ve never worked with a RSPS scripting system).

#6

Which is why it comes down to needing a core to be written by people that are capable of doing it. Since Osiris is already done, it saves a considerable amount of time (especially since the success rate for server cores is very low). I will look into osiris in greater detail later on and post what I think about it. I can vouch insomuch as I worked on it in theory with blake and boomer, but I’m not thoroughly familiar with its progammability. To avoid making any assumptions, I’ll just end my post here. :stuck_out_tongue:

#7

Meh, I saw the development team for runerebels and wasn’t impressed. I know ry has the capacity to create a good base, but the content team would probably bork it up (no offense ry if you read this) stability wise. In fact I know one of the developers semi-personally as we were somewhat competitors in the 508 scene, and his programming was/is worse than mine was two years ago. And that’s saying something.

#8

I haven’t really looked into Osiris yet, but the other alternative is to just use github/ a private repo hosted by one of the team leaders, then build the base from scratch. Technically, it’s a very simple task. A server that uses a (mostly) binary protocol, with plugin/scripting abilities. Implementing it however would fare much more difficult, due to the different programming styles of the people that would contribute.

#9

Fundamentally, different programming styles shouldn’t matter when it comes to scripting. They wouldn’t be programming massive content systems but, rather, just small bits of content at a time. That plus absolute conventions and, despite slight differences in styles, everyone should be able to understand everyone else’s code.

#10

I’ve tried doing this before (I’m pretty sure Lin might have been part of that project at one point) and the biggest issue I ran into was how the plugins would interact with each other, build off of each other, etc…

#11

I would be interested in lending a hand with the core, content I really couldn’t care I’ve looked at osirus a few minor things I would like to see changed. 1 being the handling of user interfaces. I had a pretty cool system with one of the hybridscapes we worked on that cached what was changed for the player so the server wasn’t constantly resending useless changes. I will see if I can dig it up.

I think the best idea would be to fork Osiris and allow CM’s to submit changes once we have a content system in place open it to the entire community.

Also it would be good to set it up like the RSC community server is doing and have it auto update with Jenkins or whatever.

#12

Osiris is a good enough base we can already use. Another option is Hybridscape but I don’t know if there has been progress on it since 2010. Let’s not get people to write a new base, it will never get completed.

#13

anyone have a 508 client thats not butchered but compatible with Osiris?

#14

fork Osiris, allow anyone to submit changes but only merge changes that have been approved by CM’s. this opens up to contributors who may not be CM’s but still have lots to offer (and there’s a fair few of those). precedence of review will be given to CM’s though, as they’re more likely to submit quality contributions.

using github would make doing this all a breeze.

[quote=“m.dog311, post:10, topic:443579”]I’ve tried doing this before (I’m pretty sure Lin might have been part of that project at one point) and the biggest issue I ran into was how the plugins would interact with each other, build off of each other, etc…[/quote]i don’t see why plugins should ever have to specifically communicate with each other, and despite this, plugin intercommunication is not really a difficult thing to solve anyway (actor-like plugins anyone?)

#15

[quote=“mopman_sucks, post:13, topic:443579”]anyone have a 508 client thats not butchered but compatible with Osiris?[/quote]explain. I have a plethora of 508 clients, some raw and some more modified, what do you need for Osiris?

#16

[quote="_s, post:14, topic:443579"]fork Osiris, allow anyone to submit changes but only merge changes that have been approved by CM’s. this opens up to contributors who may not be CM’s but still have lots to offer (and there’s a fair few of those). precedence of review will be given to CM’s though, as they’re more likely to submit quality contributions.

using github would make doing this all a breeze.

[quote author=m.dog311 link=topic=546258.msg4005091#msg4005091 date=1338522420]
I’ve tried doing this before (I’m pretty sure Lin might have been part of that project at one point) and the biggest issue I ran into was how the plugins would interact with each other, build off of each other, etc…
[/quote]i don’t see why plugins should ever have to specifically communicate with each other, and despite this, plugin intercommunication is not really a difficult thing to solve anyway (actor-like plugins anyone?)[/quote]

One way would be module dependencies. That could get messy very quickly (depending on the complexity of the dependency tree)

#17

[quote=“Davidi2, post:15, topic:443579”][quote author=mopman_sucks link=topic=546258.msg4005158#msg4005158 date=1338538203]
anyone have a 508 client thats not butchered but compatible with Osiris?
[/quote]explain. I have a plethora of 508 clients, some raw and some more modified, what do you need for Osiris?[/quote]
have not looked that far into it I forked Osiris then left my computer, what I want is an unmodified (refactored would be great but not necessary) client and the cache

#18

[quote=“zyle1992, post:16, topic:443579”][quote author=_s link=topic=546258.msg4005335#msg4005335 date=1338575664]
fork Osiris, allow anyone to submit changes but only merge changes that have been approved by CM’s. this opens up to contributors who may not be CM’s but still have lots to offer (and there’s a fair few of those). precedence of review will be given to CM’s though, as they’re more likely to submit quality contributions.

using github would make doing this all a breeze.

One way would be module dependencies. That could get messy very quickly (depending on the complexity of the dependency tree)[/quote]perhaps discourage developers from overdoing on the dependencies

#19

What happened to other based community games? For a while we did intend to have pokemon as a game.

#20

Moparmon, but not enough people wanted to contribute.

As for the community server I think it would be great. Id love to help design content and do any graphic work required.