Oolite Bulletins

Mmmm.. impressive stats..

Now all we need is someone willing to re-write Oolite from scratch!

Perhaps not a complete re-write, maybe it is possible to just write a LAN-play OXP then just use some VPN to simulate LAN worldwide... just an idea

Staer9 wrote:

Perhaps not a complete re-write, maybe it is possible to just write a LAN-play OXP then just use some VPN to simulate LAN w

As a member of group of people who were actually daft enough to try this and failed miserably I would definitely put my vote in for a complete rewrite being required. There is no simple "just-add-water" solution to this, the whole thing is just wired up wrongly for a client/server system.

I'm still not endorsing Oolite MP but why client/server? Probably better peer-to-peer where everybody has an agreed common view of the Ooniverse - ships, weapons, system(s) pre-downloaded. Then positions, orientations, velocities, weapons firing are multicast to the participating players to deal with on their local machine.

However, either way, Oolite couldn't do it, the rewrite, I suspect given the way Oolite is designed, would be a complete new code that had nothing to do with Oolite in its current form but could perhaps load Oolite OXPs using the same data structures and pretty much nothing else (gfx engine may be?)

DaddyHoggy wrote:

I'm still not endorsing Oolite MP but why client/server? Probably better peer-to-peer where everybody has an agreed common view of the Ooniverse - ships, weapons, system(s) pre-downloaded. Then positions, orientations, velocities, weapons firing are multicast to the participating players to deal with on their local machine.

However, either way, Oolite couldn't do it, the rewrite, I suspect given the way Oolite is designed, would be a complete new code that had nothing to do with Oolite in its current form but could perhaps load Oolite OXPs using the same data structures and pretty much nothing else (gfx engine may be?)

We were looking at a persistent server solution with a single instanced universe. I can't comment either way on how difficult a peer-to-peer solution would be to implement, I've got no real world experience in deploying those mechanics.

Spooky wrote:

We were looking at a persistent server solution with a single instanced universe. I can't comment either way on how difficult a peer-to-peer solution would be to implement, I've got no real world experience in deploying those mechanics.

I must chime in and say that making a persistent server solution creates a can-of-worms in terms of needing MORE hardware to run the game instead of people joining someone else's already-running game. And a persistent server which can have random joins from outside, you will have unknown people there just to make others miserable.
Also, with a persistent universe...a late-joiner could be faced with far-reaching disasters to cope with...such as multiple systems destroyed by supernovas.

Player in-game clocks would have to be asynchronous relative each other, since a single hyperspace jump and docking+launching and repairs and buying ship equipment and even buying fuel increases their time. Making people wait to join "the big fight" would ruin the fun.

Difficulties in setting up/hosting/joining a game would no doubt be a massive problem. And 1 player with "bad pings" (high latency and/or low bandwidth) could ruin the experience for others depending on how the networking code handles that. At the least, "bad pings" could make that player hard for other players to fight, assuming they're not playing cooperatively. The whole universe may require tinkering to keep NPC ship/object counts low for multiplayer.

Why is it that everywhere there are people saying

of course, you could give it a try

has anyone actually tried? all you need is two planets to fly between and 3 or 4 different ships to fly and it would be great fun, one planet tech14 rich industrial and another poor agricultural tech4, that means that all equipment can be purchased and also gives the basis for trading (so therefore allso piracy ) for just two planets all you need are 30 people online at the same time, which is easy, so therefore no huge servers (and huge costs) are neccessary.

If this works and intrest increases another planet can be added and so on, donations however would be neccessary unfortunaly, because I am pretty sure that the open source nature of oolite doesn't allow advertising

the in game clock would have to be removed (making hyperspace jumps immediate)

what do people think?

Staer9 wrote:

for just two planets all you need are 30 people online at the same time, which is easy, so therefore no huge servers (and huge costs)

Actually, latency issues can occur with just 2 players. With 30, you're certain to have a few with really bad pingtimes due to distance, bad ISPs, and/or bad networking equipment.

I'd hate for the game to resort to this:
http://tvtropes.org/pmwiki/pmwiki.php/Main/HitScan

Staer9 wrote:

what do people think?

Definitely, do give it a try!



Or, to expand on that (ever so slightly): Oolite's codebase is fairly clean (very clean compared to a lot of commercial code I've had the 'pleasure' of maintaining in the past), but it's just complicated enough that most 'simple changes' are almost never simple.

Modifying Oolite the way you suggest involves a potentially unknown amount of changes, none of the dev team has got tons and tons of free quality programming time (me, I've had exactly 0 time the last couple of months), and you - the person who feels most passionately about it - is the one who stand the best chance of staying up at all hours wrestling the code to the ground until it's more-or-less working...

hmmm I find it tricky enough to get the plists for my ship oxps to work... coding an entire game , I would help if someone pointed me in the right direction, but I couldn't lead that project.

Everything is easy when you’re imagining someone else doing it.

Ahruman wrote:

Everything is easy when you’re imagining someone else doing it.

Very Zen.

I do realise how tricky programming is (I have tried and failed to make a very simple racing game), but I realise the potential that oolite oonline would have, as I played vendetta online for a short time, and to be honest it was rubbish compared to oolite that isn`t online, seeing as oolite 2 is going to be re-writen from scratch it might be worth implementing the possibilities of oolite oonline into the standard game and just waiting until it is possible to rig up a server system. Although the idea of creating a player AI is a good one it would never match fighting/trading against real players.

as I said, if anyone takes up the challenge, here is one person who is happy to help.

Staer9 wrote:

seeing as oolite 2 is going to be re-writen from scratch

Whatever gave you that idea?

As any programming project gets larger, complete rewrites (sometimes called "refactoring") trend towards 0. Often it is only 1 troublesome part that gets heavily reworked with each minor revision.