GSOC 2009?

[Agon]

Hello,

could we participate this year in the google summer of code event?
It would be a good advertisement for Spring.

It's starts in March:
http://code.google.com/support/bin/answ ... opic=10729

Ff we take part in gsoc we should collect some project ideas before March.

[imbaczek]

ideas are plenty:
- implement md5 in spring
- add possibility to script unit behaviour in lua
- add p2p/roundrobin for hosting (so host exit doesn't stop the game)
- fix sm3 or make a completely new map format (JC says somebody's working on it, but nothing public...)
- make a proper gui for spring.exe, rip out libspringlobby from springlobby and integrate it in the game
- lua-fy more parts of the game (command AIs, move types)
- bring uberserver up to speed, add irc and/or jabber for the lulz

i don't quite believe google would want a rather niche game... and we're not exactly an organization.

[Auswaschbar]

i don't quite believe google would want a rather niche game... and we're not exactly an organization.

They had projects for BZFlag, Battle for Wesnoth and ScummVM in the past, so why shouldn't we try?

[imbaczek]

they did? must've missed them. in that case, i say go for it!

[Agon]

Yeah they had all sorts of projects:
http://code.google.com/soc/2008/

Ideas are not the only problem
We also need mentors...

Can 'students' take part in gsoc even if they are not real students?

[jcnossen]

Can 'students' take part in gsoc even if they are not real students?

You need to be registered as a student somewhere afaik.

Hm fixing sm3 suddenly sounds a lot more attractive when you get 4500$ for it :) I think I could work that into my schedule :)

For ideas: How about deferred shading? Seems to me that it is a pretty good way to get fast enough dynamic lighting into spring. Is it doable in 3 months?

[trepan]

easily, depending on your approach :)

[el_matarife]

ideas are plenty:
- implement md5 in spring
- add p2p/roundrobin for hosting (so host exit doesn't stop the game)

- implement SHA-256 or SHA-512 in Spring
There's no point in implementing MD5 or SHA-1 in Spring since they're no longer cryptographically secure. CRC32 should work just fine for error checking, so the only reason to need better hashing algorithms is security. That being said, someone could just edit Spring to return bogus file names / hashes so it would be pretty tough to really ban any content in a permanent way.

As far as P2P / roundrobin goes, I'd also like to see joining games in progress and save games in progress, and maybe even resync. Also, we'd probably want a screen to popup when someone drops or desyncs giving the options like "Save and quit", "Save and Continue", "Continue", etc.

Also I totally forgot. We've had a few abortive attempts to join the GSOC program before. I think they're chronicled in the old Berlios mailing list, and maybe some older threads in this forum.

[imbaczek]

ideas are plenty:
- implement md5 in spring
- add p2p/roundrobin for hosting (so host exit doesn't stop the game)

- implement SHA-256 or SHA-512 in Spring
[...]

uh, the other md5. md3 would also be nice.

[Kloot]

that'll be another $4500 then

[AF]

md5 was always suggested out of convenience not necessity. The file sites all tend to use md5 hashes.

[Tobi]

Gathering ideas is easy, the question is who wants to be mentor.

[el_matarife]

the other md5. md3 would also be nice.

Wow that's hilarious. Also, it is pretty confusing they gave a model format the exact same name as the most common hashing algorithm on the planet.

[Agon]

Yeah the 4500 US dollar for the student and 500$ for the open source project are very attractive.

If we take part in gsoc I would like to include another script platform for mods/maps in a gsoc project... but I don't know if it's possible in 3 months with a good result...

Tobi is right. We need at least one administrator and co-administrator for the spring. Plus a few mentors...
Or can one person be a mentor and a student at the same time ?

Another idea: MacOS port

I suggest that developer who don't have time to take part in gsoc at least join as mentors.

[KDR_11k]

the other md5. md3 would also be nice.

Wow that's hilarious. Also, it is pretty confusing they gave a model format the exact same name as the most common hashing algorithm on the planet.

id Software has been using md2 and md3 for previous games so reaching md5 was kinda expected.

I hope they don't reject Spring for being reliant on proprietary IP

[el_matarife]

Another idea: MacOS port

This seems like a really great idea.

I'd also like to suggest a project to get the Spring software stack fully 100% compatible with Vista / Windows Seven with best practices top to bottom (Engine, lobbies, settings program and anything else needed to run Spring). I'd suggest targeting the requirements of the Games for Windows program as a baseline for best practices, or whatever else is recommended on MSDN.

Totally wild project suggestion that's probably useless: Fix endian issues so we can get Spring running on non-X86 CPUs? Realistically, most of these have really really low market share and don't have enough power to run Spring right now, but I could see it being useful in the future. Maybe people could use a stripped down version of Spring to make a tower defense mod or something like that. There's a decent amount of ARM and other systems that can run open source software and have at least basic OpenGL acceleration coming soon or out now.

I hope they don't reject Spring for being reliant on proprietary IP

This argument was brought up last time. Now that we've got a fully GPL version of Spring with content available I think we're in the clear.

[Licho]

el_matarife - CRC are not enough..
We need proper hash for resource identification and not just integrity "check" ..
Atm there are tens of collisions. The fact that we produce thousands of new CRC with each spring release does not help it..

Its gone so far i plan to identify resources (for downloader) just by name instead of using unreliable and ever changing CRC.

[imbaczek]

one more idea - make spring able to run multiple games during one execution (for campaigns and stuff, with Lua-scripted intermissions, etc.)

[jcnossen]

But that still leaves the original problem... who's going to be a mentor?
Doesn't any of the current dev's want to be mentor? Making a multiple of 500$ just to track progress/evaluate doesn't seem too bad right. tobi/kloot/imbaczek?

[Agon]

But that still leaves the original problem... who's going to be a mentor?
Doesn't any of the current dev's want to be mentor? Making a multiple of 500$ just to track progress/evaluate doesn't seem too bad right. tobi/kloot/imbaczek?

It seems like this is the biggest problem to get into gsoc with Spring.
You forgot hoijui, thought . And what about you jcnossen?

What about SpringLobby and gsoc and the other sub-projects?