Y'know, I've got a whacky idea. How hard would it be, really, to remove all of the economic code from Spring and just make it all run through LUA instead? Doing that would open up a HUGE number of opportunities gameplay-wise...
I don't think it'd be terribly hard, honestly. The only hard part would be to have this script track units as they were created/destroyed and then get the values either internally from itself or from the unit's definition. It'd be really interesting to combine that with unit-morphing and couple of other things to create totally new economic models, freed at last from Spring's core constraints.
Economy via LUA!
Moderator: Moderators
Please provide specific examples of what you'd want to do
(so that I can let you know how hard it would be to do). Try
to take into account building / repairing / reclaiming type
operations, or new modes such as morphing, cloning, resource
transfers, etc... Would you also want the 'proximity benefit' type
of setup that I've been told SupCom has? (might not be that
hard). How many new resources would you want? Would they
require carriers as in StarCraft?
P.S. I'm guessing that replacing all economy operations will be
prohibitively expensive, but it might be worth trying it out. If your
requirements aren't too great, and with a few 'smart' tricks, it
could be workable.
(so that I can let you know how hard it would be to do). Try
to take into account building / repairing / reclaiming type
operations, or new modes such as morphing, cloning, resource
transfers, etc... Would you also want the 'proximity benefit' type
of setup that I've been told SupCom has? (might not be that
hard). How many new resources would you want? Would they
require carriers as in StarCraft?
P.S. I'm guessing that replacing all economy operations will be
prohibitively expensive, but it might be worth trying it out. If your
requirements aren't too great, and with a few 'smart' tricks, it
could be workable.
WC style food or pilots? Mmmm :)
That's one major features I've been looking for.
Hmm I assume it's possible to change the limit tag on the fly via LUA. I was thinking of a support type of building which sends something like an "I'm alive" signal to some LUA script. The script then remembers the unit by ingame ID (to allow more than one support building) and changes limits for a certain unit (or units) accordingly or resets them if the unit doesn't respond anymore, or one can send a signal that it's dead in the Killed() function, yeah that would be better I guess.
That's one major features I've been looking for.
Hmm I assume it's possible to change the limit tag on the fly via LUA. I was thinking of a support type of building which sends something like an "I'm alive" signal to some LUA script. The script then remembers the unit by ingame ID (to allow more than one support building) and changes limits for a certain unit (or units) accordingly or resets them if the unit doesn't respond anymore, or one can send a signal that it's dead in the Killed() function, yeah that would be better I guess.
-
Guessmyname
- Posts: 3301
- Joined: 28 Apr 2005, 21:07
-
1v0ry_k1ng
- Posts: 4656
- Joined: 10 Mar 2006, 10:24
how so? for gundam and EE, they already have a hard time thanks to the fact neither mod works off of TA styled econ. Currently RAI seems to play gundam the best because it doesn't con assist and it builds more random stuff. I have not tested EE but it may work well with ee.1v0ry_k1ng wrote:woudnt doing this shoot every skirmish AI in the head?
Once a project goes away from a ta-styled econ it already has issues.
If the resources just require a building then some ais may figure it out. Others will not at all. Some like NTAI with complex build configs can also work it out. In the end the projects are going to start becoming separate games in and of themselves that utilize spring as a platform to run content and will require custom ais.
-
1v0ry_k1ng
- Posts: 4656
- Joined: 10 Mar 2006, 10:24
gundam and EE could be played by NTAI perfectly, the config tool is very flexible.
KAI played EE like he-man before the hubs change.
people dont seem to realise but most non-star players do use the AI's to learn mods and practice. if you check the lobby at least 1/2 the games contain just one person; AI skirmish players. not counting players who use spring SP.
If this was done, if would have to be optional. its not fair on players and the AI devs who have put in a billion man hours to break their AI's for the sake of experimenting with new mods.
KAI played EE like he-man before the hubs change.
people dont seem to realise but most non-star players do use the AI's to learn mods and practice. if you check the lobby at least 1/2 the games contain just one person; AI skirmish players. not counting players who use spring SP.
If this was done, if would have to be optional. its not fair on players and the AI devs who have put in a billion man hours to break their AI's for the sake of experimenting with new mods.
1v0ry_k1ng wrote:gundam and EE could be played by NTAI perfectly, the config tool is very flexible.
KAI played EE like he-man before the hubs change.
If I made the config for gundam I can promise you that it would be outdated in a new release. I tend to change the build menus and resource system around frequently to meet new features in the spring engine as springs features allow me to move closer to what I originally planned. So I would have to stop development at each version and redo the config for NTAI at each release. Not exactly productive.
*edit please stop talking about ai issues, we are derailing the thread.*end edit*
Hence the addition of 20-30 generic universal build routines. In such a case as gundam you would build a small number of tasklists and assign generic keywords as much as possible. That way NTai economy algorithms do the work as best they can rather than you defining each and every unit needed for every revision.
However I doubt that'll pursuade you. In the mean time a gundam config that played rather well can easily be built in several minutes thanks to the nature of gundams economy in 1.11.
I'm sure that when the time comes trepans AI interface bindings can be leveraged to AI devs and modders a platform for standardizing lua features such as sockets, superweapons, or resource harvesting.
However I doubt that'll pursuade you. In the mean time a gundam config that played rather well can easily be built in several minutes thanks to the nature of gundams economy in 1.11.
I'm sure that when the time comes trepans AI interface bindings can be leveraged to AI devs and modders a platform for standardizing lua features such as sockets, superweapons, or resource harvesting.
-
1v0ry_k1ng
- Posts: 4656
- Joined: 10 Mar 2006, 10:24
I'd be concerned with speed. I think LUA would work marvelously for things like a food limit, or even *craft style ore and gas, because the frequency of changes are so much smaller than in TAesque games (only need to update when a supply depot is killed or built, or when a worker returns with ore / something is build/cancelled)
Now that I think on it, 90% of RTS's use discrete economic units that would be LUA-able without too much of a performance hit (Disclaimer: I'm speaking of LUA in general, here, as I'm unqualified to comment on Springs specific implementation. So some things I say are "possible" may be really hard.) Starcraft, C+C (even generals)... most anything. The only things is might be too slow for would be TA-style, and Z-style economics.
Now that I think on it, 90% of RTS's use discrete economic units that would be LUA-able without too much of a performance hit (Disclaimer: I'm speaking of LUA in general, here, as I'm unqualified to comment on Springs specific implementation. So some things I say are "possible" may be really hard.) Starcraft, C+C (even generals)... most anything. The only things is might be too slow for would be TA-style, and Z-style economics.