Hmmm...

I might still have to test it at least... The position of the players aren't super crucial anyway since the AI will take over for the other players and make them "help" your player.

The location of each player only has to be very approximate (say a 300 pixel margin of error).

The proposed AI idea would involve the other players characters sticking close to you... only if they leave the immediate screen area, go into a building, or leave the screen and go to another screen all together would it really need to "adjust" the players location.

"Teleporting" the player in the event of them doing one of the above things wouldn't be such a bad thing... I'm just really keen to get something at least functional until such time that we get a full online object that can track things better.

Am I still barking up the wrong tree or is it worth testing out? I don't want to waste a bunch of time if it's just not going to work, but I also don't want to wait forever for an online plugin.

~Sol

You could do something based on nodes, or way points, like if another player gets to building A it sends an update. Then the individual clients could use some sort of interpolation using last know coords updated to the new one.

Yeah that's pretty much what I was thinking. Using nodes or target waypoints, and the players "AI" will move them to the next updated position. Sort of like a real-time game using turn based node updates.

So, if I shoot at a player, I'm not even sure I'm really hitting that player? And I also wouldn't be sure that a player is shooting at me? I think the HTML object can be used just for online card games, etc. Interpolation between the last known coords is way too inaccurate.

It's going to be a cooperative game, so shooting at other players will not be an issue.

Linkman: There was an alpha test of the online object that was being worked on in the chat room one night. It was never officially released but seems to work ok for a chat program.

You need a proper online object, it won't work via the HTML object. It's the wrong technology, it's based on clients making requests to servers, which then respond with a document. There's no way for the server to notify the client out of the blue. You might get a turn based game or something working but still - it's like using a SQL database for a todo list. Way too much technology involved, which is going to make it clumsy and slow.

So... you are saying it won't work at all? Or just not work very well?

Slow won't really matter... that's what the AI will be for. As long as the other players are kept relatively close to their intended position, and shoot when they are pressing the shoot key (specific number of rounds fired etc won't matter... it will be shooting until the clip is empty for all I care.) then it would be fine.

How slow are we talking here? Would updating the player location every 2-3 seconds be too much?

Cheers all!

~Sol

I had this theory that I could make a complex multiplayer game suing the HTML object, and I was wondering what everyone's opinions were on the feasibility of it working properly.

My plan is to make an overhead view game like the old GTA games. The player will run around shooting at things, picking things up etc. There will also be other players that can join the game and run around with the other players.

My overall idea is this:

Knowing that the HTML object is way too slow for accurate and real-time movement I was planning on "faking" it. The player will move around as normal on his/her screen, while the other players will use a pseudo AI to control the more complex actions. The HTML objects comes in to play by sending the players X and Y coordinates every few seconds, and data such as "is shooting" and "is searching" etc.

In between the pings of data, the AI takes over and moves the player towards the coordinates the last ping data was set to. There will also be some correction data so if the players coordinates don't update correctly for any reason, the character may appear to "teleport" nearer the actual location of the real player. This may look and feel like the old-school lag from dial-up days, but I was thinking it may just work.

Any thoughts or ideas to add to this?

I will also be using Rich's early alpha online plugin he has made for a chat system since it seems to work just fine for that as it is.

The only problem I can foresee at this point is how the client and server will communicate? Should it be made to work as a straight peer to peer data transfer? Is a server required at all? Will it even work?

Any input or considerations is much appreciated. Thanks in advance.

~Sol

Looks and sounds interesting. Will be checking out the demo/beta as soon as it's up!

~Sol

Wow! Very professional looking graphics, with polish to match

I will have to download and try this out for sure, it looks fantastic!

~Sol

Wow that's really quite impressive. Nice work mate

I'd love to see how you made the player work... a physics platformer has been on the back of my mind for a while now, but I had trouble getting anything to work convincingly.

~Sol

If you're just skimming through, you're likely not to notice anything out of place. They even copied one of Ashley's posts in one thread... funny how a newbie with only one post is suddenly a developer commenting on the future of Construct.

~Sol

OoOoOoOo Zombies... Awesome!

~Sol

Well I am open to suggestion if anyone can work out a way around the problem... :/

~Sol

Yeah this game is pretty damn awesome... I finished it on the hardest difficulty setting... I know what you mean. It's not that bad once you get the hang of the battle, you just have to prioritise the kills.

I had a lot of fun with this one... highly recommended to anyone looking for a good single player game.

~Sol