Sadly, most of your changes seem to have a differing aim.
Yes, we know. But anyway, comments are welcome.
Sure, most part of projects has no interest for MGdev mod, but MG team, like DnC team, has a very good vision of the tremulous code, their weakness, and we are very interested to know how do you think about our stuff too !
don't complain about the balance of a weapon that has long since been changed; regarding dretches: why shouldn't luci own them just as much as tyrant owns naked humans?
How can you compare human weapons with alien evolution ? There are too differents ! Yes tyran can kill easly all humans, but tyran need a goon+ to destroy a base against good players, event if everyone has a rifle.
If everyone is dretch, luci need no one else to destroy an alien base. Tyran is more expensive than luci (luci: less than half max points, tyran: more than half max points). Luci has a long range, tyran a short range.
Yes, my comparaisons are stupid, because it's stupid to compare their two classes.
So, I compare luci with human weapons: all human weapons has streigth and weakness. All... but luci in 1.2
Why do you want to use a chaingun when you see the power of the lucifer canon ?
With current lucifer canon, good player can easly rush alien base alone, and kill all aliens, dretch or tyran...
He need no support, no teamplay. It's a bad thing : it *must* have weakness !
When a tyran fight against rifle, rifle has time to hit tyran before to die.
Currently, you can kill all dretchs with a lucifer canon without any damages, even with a tiny charge. The only cost is some power for your luci.
With a variadic radius, you can kill dretchs, but you must have a good aim, or wait to a full charge but maybe dretch will have time to hit you.
-all the dev's i've asked are strongly against adding any form of credit sharing to tremulous
For a "pro" balanced game on a standard slot server (max 16), I'm against share/donate too. But you know, tremulous is a pretty hard to learn. More over, strong player are very powerfull in games : they take all kills, and kill beginner very easly. So, beginner don't want to loose the few credit/evo they have, and they prefer to camp to stay alive.
With donate and autodonate, beginners has more credits/evos, and they know even if they die, good players will give nextly some money, and they take risk to attack. We tested it on high slot server, and it's work.
I know, tremulous is not made for high slots server. But many players like it, so I think we must accept opinion of their players and give them tools they like.
* No Private Message Traces - Don't give people the expectation that something said on a game server can't be monitored. Additionally, people cheat using private messages. Please don't try to cure stupid, it won't work. Better to say it to players the first private message they send that it will be logged, maybe then they'll think about it.
So it has a bad name. It's a bug to call "private message" something logged.
Cause of private messages, I saw sometime inner discussion, they though nobody saw them... Admins are not big brother, and it's idiot to talk about privacy if there is no privacy.
* Players Ranking of Master Server - I have yet to see ranking ideas that make a bit of sense. Kills don't really matter (Khalsa once said he tries to end the game with the fewest kills possible.) You can't measure builders. Building lots does not mean building well, not to mention that you can't determine the quality of building placement effectively. More than once I've fed s3 humans on atcs with s1 dretches causing damage and distracting them enough for the builder to build and slowly getting kills up to s2 trappers and ultimately s3. Sometimes feeding is really, really good and appropriate.
I agree with you, it will be probably very hard to make a real ranking. It's only a project, we don't talk about many details and it's maybe impossible to have a very balancing ranking. But it's not a reason to don't think about it

* Climbfall - Dretching will suck with this. Dretches constantly take damage, whats the point of wall walking if you constantly fall off?
It's very uncommon to fall with a dretch, and also with a basilisk and basilisk+. The only one alien which easly climbfall is granger. The idea is maybe a bad idea for dretchs and basilisk (we currently test it), but I think it's a good idea for granger.
Why does everyone want to "fix" camping and egg spamming? This isn't a deathmatch FPS and there actually can be a fair amount of strategy involved. Thats one of the many reasons donate and its ilk suck, it kills the strategy. Please don't kill the expectation of using strategy and turn tremulous into a pure twitch fps, there are plenty of those and only one tremulous Smiley
About camping, a game must not became boring to be strategic.
About eggspam, on big maps, humans can't win if alien team is eggspamming. I and my clan, in clan war and with VOIP, fight against an eggspamming on nexus. Even using outpost, we did not win, and we had a *big* advantage on aliens ! It's not only a strategy, it's a gameplay unbalance.
Done for now, remember, you said not to hesitate
I hoped you could find just only little positive things, but... you don't like me

Thanks a lot for your opinion

What's the problem with the admin system? By default, it is clean, efficient, and easy to understand. It is kept separate from /commands because they are different things with a different purposes and restrictions.
Several things:
First, we think ! and / commands are not different things. It's commands, commands affect the game. I'm ok to say moderation is not administration. But in original quake3, there was only way to comunicate with game: commands.
So moderation can be a layer over administration, it's clean and more stylish than current '!' commands !
One of our ideas is to make '!' commands will become wrapper to / commands. So, all commands are / commands, but you can define in admin.dat '!' commands with same names and give their all restrictions you want.
Another idea is to drop all '!' commands and make '/' commands instead, and give a way to user to define its own '/' commands in a config file.
Second, we think admin.dat has not a good format :
In a side, it's a config file to give rights to administrators and define custom admin commands
In other side, it's a data file to store admins data and ban data.
One idea is to keep the same admin.dat for admin rightsand commands (or use lua instead, but it's an other discussion

) , and to make a real data file for real data, with, for exemple, a sqlite database. So, two servers could use same database to manage admins and bans : some servers has managed by same admin team and want to share admin datas, it's currently impossible.
Third, we think admin flags are crappy : there is not enough flags to add many admin rights, and flags are very hard to read (to understand an admin-right flags line, need to have a documentation and time to loose to translate flags), so we need to think about a better idea to manage admin rights.
RSA identification is now only for convenience, since server-unique GUIDs are completely secure.
As I said upper, with server-unique GUIDs it's impossible to reconize a player trough servers, so it's impossible to share admin data.
With RSA, it will be possible

There are a lot of your changes, however, that would be appreciated on the official bugzilla and some on your features list that have been already fixed. The -ffast-math thing is especially significant, in fact I will submit that myself as soon as I have confirmed it.
edit: it seems some of your listed features are bugfixes for bugs which aren't in SVN (e.g. the left the humans thing). What codebase did you start from?
double edit:
http://bugzilla.icculus.org/show_bug.cgi?id=3533Yes, you have right.
We need to take time to submit them.
We currently work on svn992. It's possible than some bugs don't come from tremulous svn but from patch we added and stuff we made. That's why we must sort them, that why it will take a long time to make pretty bug reports on tremulous bug tracker
