RTSUnits and Weapons

#1 I think you missunderstand the RTS approach. To have so many units on the screen we have to limit the amount of objects passed between server/client and so wemust live without such thinngs as mountable weapons. That would just be another object the server would be monitoring.

Instead make differant units that have a seperate dts file , and datablock.

#2 I hear you, I'm just not wanting to dance that specific dance.

Customizable units in a RTS is cool. We're trying to do it with Torque.

So, can I move RTSUnit towards my needs, or if I need to use regular units with the RTS Starter Kit instead?

(to defer the recommendation of implementing permutations and combinations, our concept involves multiple mount points with many options for each. It's certainly *doable*, but I'm not sure its preferred, or really necessary)

#3 We are working alonge these lines, except we've no plans to have seperate weapon objects. We'll end up implementing the different weapons based on enviroment, distance to target, etc, where each machine will calc what type of projectile to fire from where.

As for your wanting to mount them a la normal Torque, take a look at the rtsUnits pack and unPack functions, then take a look at the packUpdate methos for shapeBase. you will notice that the rtsUnit version, while it masks for images that don't need any updating, it doesn't send any info. It will need to transmit the image data like shapeBase does (don't forget the unpack methods too).

While rtsUnit is derived from player and shapebase, it bypasses their pack/unPack methods, so the only information sent to the clients is contained in it's own methods.

But do consider that going that route will add 8 or 9 more pieces of data to be sent per weapon, per unit. WHich could get to be a pretty heavy extra burden if you are planning on allowing hundred(s) of units on screen at any one time.

#4 Part of our desire to have the flexibility in weapons was a concession in terms of unit numbers. We'll be on the low side, for sure.

Thank you for your advice.

#5 If your plans are not have very large numbers of units, then you ought to do fine with the extra netload, especially considering that the RTSUnit is efficient to begine with. For us, I think what will eventually dictate the unit limit is how our implementation of AI turns out.

#6 Thanks for the detailed reply Martin, nice job.

Ed, as Martin points out, this is certainly possible in the RTS SK, and fairly easy to do. Sounds like you've got an interesting design.. don't forget to share previews. :)

#7 We're in the process of testing the idea of mounting weapons on RTS units vs "organic weapons"--where the weapon is part of the base model--ourselves. It is looking as if the better option is to have the weapons integrated with the unit model instead of mounted, for the reasons Martin pointed out--they trimmed down RTSUnit for a reason (performance, especially network), and it can be very easy to forget/ignore those design decisions and make ones that may bite you when you get to a high number of units in a game.