Hi everyone! I've been keeping track of PA and it's come a long way! I fired up the game today to see how far everything is progressing but with the news of the 1.0 on the blog, I'm starting to get a little worried about the availability of certain features. One of the things that I find most annoying are that pathfinding does not really seem to work from time to time, especially with assisting engineers and a commander that walks backward. They bounce into stuff and move around way way more then they need to. I thought this was going to have the same system as supcom 2, flow fields? Did that not make it in the game yet? The second thing I cant help but miss is replaceable build orders. Supcom did this fantastically, shift and replace buildings that are yet to be built. Other stuff like the ferry system (which I saw mentioned here before) would be great, just like the megabot and the unit cannon. I know there's been a lot of back and forth for the last two but I really hope someone can reassure me on the first two points. Cheers!
Yes, it doesn't work. Yes, it does use the same tech as SupCom 2. No, SupCom 2 never used flow fields. The devs just kept using a wrong term, what they actually used were vector fields for a goal based navigation system, expanded by a flocking algorithm for formations. PA extended that system from 2D into full 3D space, but obviously the system is still very much bugged. Apparently they tried to take a short cut by simplifying the generation of the flow field, and as a result smaller obstacles have not the effect they should have. To be specific: Causing units to avoid these obstacles in advance. It did work much better with the previous port of the 2D system, applied by cube mapping the field onto the planets surface, but that system had bugs of its own. Yes, there is still much stuff missing in the command system. ETAs, persistent orders, modifying existing order chains etc. Most of that stuff is still on the TODO list, but it's not top priority.
There were some blog posts from devs who worked on Supcom 2 about the tech they used and how "innovative" it was. Well, I must give them that you actually found only very few, complex research papers for the terms they used, which contained some graphics and ideas which looked remotely similar and which definitely were a source of inspiration for the devs. But couldn't change the fact that these techniques were actually well known by that time, under different names. Some of the techniques have been reused for PA, and the devs happened to drop a piece of information from time to time, mostly Sorian did. On top of that, it's any easy task to figure how they could have possibly combined the techniques they listed to achieve the visible behavior. (And also where the bugs originate from.)
Interesting. I guess I just thought that the term used was flowfield pathfinding and was used on both games PA and supcom2. But couldn't change the fact that these techniques were actually well known by that time, under different names. (What names are they?)
goal based vector field pathfinding : http://gamedevelopment.tutsplus.com...-based-vector-field-pathfinding--gamedev-9007 Neat flocking algorithm http://harry.me/blog/2011/02/17/neat-algorithms-flocking/
