Plain SVN (ie, svnserve) will not work.  Other options I can look into if everyone on the group with checkin access feels there is a need to do so, though I make no guarantees of how fast that process will happen :>

Also everyone would need to update their sandboxes, but that's semi-trivial.

MGDev isn't the official 1.2 branch, it's a sandbox for Norfenstein's ideas. No one has ever said that MGDev would be integrated as is. Even Norf's ideas are subject to the core-dev veto. More likely Timbo will force us to split every feature off into its own individual patch and convince him to include each. Half the changes we've made will probably fall by the wayside as a result.

He doesn't really ask for much. We're not backlogged with his suggested changes, I actually have to prod the poor guy into making decisions on things (hives was a recent example). Most of the changes we make are bugs/missing features/enhancements we come up with on our own.

I should also point out that we always welcome new ideas for solving balance issues. Start hanging around our IRC and join in the discussions when something comes up.

Note: If the problems are having to do with infrastructure, dont worry your cute little heads about it. Arrangements can always be made for more infrastructure. The current setup is the way it is because it was basically what we had on hand .

I'm always one to have more infrastructure, I like trac, and there are plenty of options available to us.

I (not being one on the dev team) for one would love seeing the dnc crew come help out :-)


Khalsa

Eh, Risujin, we're about as official as it gets. At least what we're
doing stands a chance of inclusion. From what I hear, at least TJW likes
a lot of what we've done. That's a bonus.

We do need to get more organized, though. I've always been saying that.
I'd like to see more dev-tools. Even if we just register with
googlecode, that would be a huge leap forward. A wiki+integrated
bugs+integrated svn would actually be pretty perfect.

Shall we consider googlecode? I would guess it's a good option. I use it
for work now (my project is Versaplex if you want to go find it). Trak
is kind of "meh" to be honest.

As for what to actually do, I have a slight plan.

1) We gotta appoint somewhat of a secondary leader. Norf can still
approve/veto our ridiculous changes, but we need someone interacting
with the other developers on a day to day basis and organizing things.
Even without the power of Norf, this person could do something like this:

  - Organize MGDev play times when we agree the focus is building ideas.
Half of it will be chatting and half playing. The "Leader" will take
notes for what might be implemented.

  - Decide what should actually be done, in moderation

  - Assign tasks

The traits of this person would be: moderate conservatism, leadership, being around a lot, and decisiveness.

Anyone around that think that they could somewhat fill this description? The most important thing is having the time to spend on organizing people. The rest can be more or less figured out.

First, I'm really happy to see your reactions, thanks

About place to set stuff, some ideas:
- Source forge -> it's lagging, and too often down, so, I think, stupid idea.
- Google code -> I don't really know their services, but it's probably good, like all they do. But I don't like the idea to use google to everything: everyone know a monopoly is never a good things, and google has too many monopoly in many sectors.
- GNA -> free, many services, seems to be good, maybe a good idea ?
- Other free hoster -> many members of tremulous community have their own servers. Maybe less stable than a real hosting (if howner stop to pay the server, must migrate all stuff :/ ), but it's really more easy to set new services, can have more freedom if you know main admin
Some servers where I have access and probably ok to become hoster:
* cygg.fr, used for a tremulous server and all DnC hosting. Little config, but probably enough for a website and svn.
* thcteam.org, used for thc team, a really good server with unlimited bandpass and good processors. Two tremulous server + 1 private, but many unused ressources.
* noobulous.fr, used for noobulous servers, Three
I have, or know people with root on their servers. Of course, subject to debate, I know you can't give all your access and your security to guy you don't know

Mind you, it's not that we don't have the horsepower.

1) This server belongs to Khalsa (the one this site and svn is on, we have full access)
2) At least 3 other members own servers
3) We pay for a dedicated server that is outstanding. It is used to host Pure Tremulous, Pure Tremulous 2, Pure Tremulous 3, MG Development without a sweat.

If we wanted a private solution, we could just use Khalsa's server. It's no big deal, but it requires that _we_ set something up. I like google code because it's simple, fast, and free. It doesn't have frills, and it's easy. The only main requirement is that committers need google accounts.

We'll probably end up setting up something on this server, though, since we have access.

This server isn't khalsa's and isn't full access.  Its a shared web host.

Nobody said I didn't have other servers...

I suppose. But at least we have ssh into it, right? That more or less means we could set up whatever we want with a bit of effort.

Real example of one of the worse servers ever:
I have a Fedora Server running on VMware on Windows on my brother's desktop (Pentium 4), sharing a 1Mbps ADSL connection with 3 other computers, of which 70% is consumed by my brother's bittorrent. Also the fact my brother plays a lot of computer games, and since the computer only has 1 CPU, this will affect the server's performance.
Since the desktop is in my dad's office (he works at home) he can just turn it off (and he often does because of the fan noise).
Since I live in China the server will be very slow.
The only thing I use it for is ssh tunneling when I am at school.
My server is definitely no suitable to host any serious website, because I live in China, therefore the server will be very slow for everyone else living in Europe/US.

Now back to the topic (Sorry about the Off-Topic):
I think any system that allows us to collaborate will be fine, whether it is trac, google code or something else. The important thing would be to to have that system up so that we can start by seeing which features of DnC should be put into mgdev, and then we could continue normal development, but with a larger team, which as Garoth said, needs a lot of organization.