AI & A few other questions?

[Draugr]

Hi,
I wanted to ask a few questions on EQemu's AI & pathing for MOBS. I'm not a programmer and haven't followed the development for a while so I just wanted to ask some questions and see if there is anyone with answers.

1. I understand there is no NPC AI and I remember reading that there was some problems with the concept. Does anyone know what the problems are or if AI is being worked on?

2. What about pathing? Is it possible to set an NPC and say go from point A to Point B then wait 30 seconds and Return to Point A or even GOTO Point C, etc....

3. With regards to the databases. Does the use of the databases make the server fully stocked (or close to it) like the real EQ? I mean is the databases what position and place all the objects, mobs, etc... So a quick installation and BINGO you have a fully or near fully populated server?


Well, I guess thats about all. I used to play EQ but got tired of the BS and quit. I still long for the days of wandering West KArana and bumping into those two goons, Froon and Choon or that damn Hill Giant! EQ is a great game, but the operators of the official servers just don't know how to make it work right. :( I can't wait for EQemu to get further along so I can setup a server and play locally with a few friends.

[killspree]

I think memory limitations are what's keeping pathing from being put in.

[Wiz]

Would really memory usage be that bad? If the waypoints were loaded into zone, and the NPC had fairly large "destination" jumps...

Ah well, I haven't tested it, so I guess I wouldn't really know.

[DeletedUser]

You need to look at the bigger picture, on EQLive not only is there pathing for walking, but theres pathing to follow people.

[Draugr]

Image,
I'm just curious about the memory usage and the pathing in EQ. I'm no programmer so I don't necessarily know the programmers side of the issue, but what if the pathing was similar to the pathing found in Neverwinter Nights? Now I know EQemu is a LARGE LARGE program much larger than NWNs, but what if a server was only running 3 or 4 zones per system? Would a pathing feature like setting waypoints A, B, C and telling a specific mob to spawn at point A goto point B wait 30 sec, then goto Point C. Then vice versa. If a MOB sees a player or NPC that is a negative faction to itself it would leave its pathing and chase that other entity. After catching and killing or losing the entity the mob was chasing it could just goto the clsoest waypoiunt for it and continue its pathing pattern.

Again, I'm no programmer and I'm certainly not trying to second guess or out guess you. I'm just trying to understand the whole picture.

Sincerely,
John

P.S: I understand that pathing is a problem right now, but is it an issue that can eventually be resolved or is this going to be a very difficult issue to make functional due to the memory limits you spoke of?

Thanks Again

[DeletedUser]

Thats the whole entire point, those 'waypoints' are within a map which is loaded when you play a game. There is no way that we are going to cut zones per computer down to 4 for some pathing. My server alone peaks at 30 users and running 23 zones off of 3 different computers are taken up, I cannot bring that down to 11-12. Map sizes could vary depending on the type of map we create, if its 2 dimensional, it will be very small, however, NPCs would not be aware of their Z coordinate. If it were 3 dimensional, file sizes would be rather large (note these are put into memory). But if we were to use 2D, the NPCs would continue bouncing like they do now, which no one wants.

Once we actually come up with a solution, not only do we have to make the ability to create a map, but a program to actually edit and create them. It is a lot of work for our team right now, we are not as big as we were from December-March (Team is down to roughly 2-3 active people, another 2 that come in once a week to work for about an hour).

I will start brainstorming some ideas, see what creative things I can come up with that shouldn't kill our computers :)

[Trumpcard]

Yea, resources become the issue, and it helps to have someone with experience in dealing with specific implementations.

My opinion is put the project up on freshmeat/slashdot with an open call to new developers. Not that the current team isnt doing great, but expanding the team will make it easier to delegate work out, and you might be able to pull some developers who have prior experience dealing with those issues/coding problems.

Yes, you flood the project with new people, but you also open it up to alot of new development talent.

Oh course, you'll probably have to have someone assume the mantle of full time 'project/systems manager' and start to
develop some design strategies to keep everyone working in the same direction.

I've noticed this team will do that, grow, then shrink down to a few folks, but not actively recruit new developers. Thats the beauty of opensource projects is how easy it is to pull in new development blood when you have a project thats interesting, and this one is one of the more fascinating ones I've seen. Hell, I'd love to be more actively involved in the development of eqemu, but it seems you need to be knighted by King Arthur first to get access to the developer channel..... LOL

I think if you want to spur the development of the project, you need to first have a more open way to recruit new developers on a constant basis, not just when you get down to 1 or 2 people.

There should be a developer application form right on the front page, or some method for pulling new blood in.

Also, sceheduling a weekly design/development meeting thats open to the public will go a long way (IRC would work), at least you make a scheduled attempt to hear what the community has to say and get new ideas.

Hows that sound? I think a weekly/biweekly design meeting on an IRC channel would be an awesome idea...

[DeletedUser]

Trumpcard, the new Events Coordinator :) j/k

The thing that insults me at the moment is the fact that we have 5k users signed up on these forums and our chat room peaks at 40 users :P

[jackass01]

Exactly what kind of server does it take to run a world like VI's. Just curious. I do not know jack about these things so dont flame me too hard.

[Draugr]

Alot, I don't know exactly how many, but when your dealing with thousands of players online at once and all the AI and zones, I'd probably think it would take 1 high powered server per 5 or 6 zones maybe, with the more popular zones possibly getting a server to themselves.

[DeletedUser]

Lets just say it would not be in your budget :)

[OverlordvI]

LOL.


I read a long time ago that Verant has approximately two zones per machine (decked out PIIs is what I recall, but that's probably wrong). One of those two zones is rarely used (only a few people on one zone - i.e. Ogguk), and the other zone is busy (i.e. Kael). One machine basically handles about 100 people, then you start to notice serverside lag.


I agree that pathing would be nice to have, but to me it is hardly something that would be *the* defining feature of EQEmu. I appreciate what these guys are doing, and how far they've come so far. To me, what would make EQEmu really amazing in the next few months is this:

1) Work on the inventory a bit more. You guys have done a great job with the random itemizer for specific mob ranges. There are databases at Allakhazam that have specific drops for the higher end mobs - when the Avatar of War starts to drop a blade of carnage, pants, AND a mask, that will be pretty neat. To me, this is something that can be easily delegated out because all it needs is a lot of grunt work itemizing specific mobs (approximately 100 - 150 noteworthy mobs - i.e., big bosses, and a few specific quest mobs).

2) Get working portals into the game. For example, in Ssra Temple and Vex Thal, none of the portals work. The zone out from NToV is bugged. You can't click on the entrance to the ToV from WW because it doesn't exist. You guys have done a stellar job getting zone lines coded into the game, and with the last patch, most of them seem to be working well.

3) Along with portals, get some doors up! I like the doors you've coded in certain zones. I'm not sure how you'll be able to do this except by having someone walk around in all the zones getting the loc of every stinkin door. Perhaps something to delegate out?

4) A few quests, just to get the coding into the game? If you start with a few epic quests (if you want the complete textual exchange that occurs between a PC and NPC, let me know and I'll e-mail it to you - I have it for *all* of the epic quests), that might be a pretty neat thing.

5) Related to (4) is NPCs that understand faction. From what I've seen of the database, factions are already in the Emu. However, there aren't very many NPCs that aggro me. What I think is funny is that they consider me indifferently before *and after* I aggro them :D .



Like I said earlier, you developers are absolutely incredible. You've done a stellar job bringing this to life and keeping it going. I use the Emu before every raid to scope out the zones, figure out new tactics, and more - and it is very valuable to me in that regard. It's also a fun thing to play around with - I like how looking at your reconstruction teaches me about how the game works, and I think that's pretty cool.


I wish I had more time to help you guys out. This is something I'd love to do when I take a vacation next year...



Cheers!

[cray]

Actually 22 computers as of the Luclin expansion make up a single server. Not quite sure now that PoP will be comming out.