The time has come to formulate exactly what goes into Demo 5! (Well, actually it's probably past time, but we won't go into that...) If we are going to get something out by February, we need to organize and start work now.

Here's what I would like from each of you

Model Leads:

Come up with a fairly well-constrained start on what your model will become. There is already some existing code there for military and diplomacy. For the other areas we can try and salvage some things from my old code, but the programmers are in essence starting from scratch. If you can have a crude proposal out by the end of this weekend that would help a lot. You need to talk with your respective coders and try and figure out what level of progress is reasonable to achieve with something like a mid-February target. (I'll be the Military coder for this round; Darren, on interfaces you get to talk to yourself ) My advice on the simple models for demo 5 is to just test out a few big things. The most important things to check out are probably the ones you are concerned might not work. Otherwise, I'd say emphasize the things that will be fun to get in as soon as possible. It's probably best at this early stage not to rely on too much connection with another model. Once you have your crude model, please go ahead and start a thread dedicated to your demo 5 version so it can be discussed a little bit, with a minimal intrusion on the main model development threads. If anyone has any other bright ideas, speak up You also need to think about a very crude interface to allow the player to interact with your model. That leads us to the next point.

Darren, how are the interfaces coming? Do you think you will have a few widgets that are basically ready for prime time in the next few weeks? It would be great to start using your interface stuff for the limited interfaces we will need to put in for the models. But I know you've got quite a full plate between the interface and diplomacy programming. Just let us know which way we should plan for.

Programmers:

You need to give what ever input you think is necessary to the model lead in your respective area ASAP. If it would be much more efficient to build up the classes for the code in a particular way, now is the time to get in your suggestions for what the first shot at the model should be. Once a model has been proposed for the demo 5 version, you need to make a realistic appraisal as to whether you can do all that stuff efficiently by February 15th or so. If it turns out that you think you just can't code a reasonable first shot at the model by the deadline, you just need to figure out with your model lead if that particular model should skip demo 5 and shoot for demo 6 instead. I think the overall progress of the project takes precedence over any particular model making it into this demo, although clearly I think we would all like to have all of those things I've outlined present in some fashion

I'm Sure I haven't covered it all. Let's use this thread as a meeting room for the model leads and coders to discuss the demo 5 issues. Speak up soon if there is a big hole in the plan, or there is handshaking you need to do with some other particular model. If Darren thinks he will be ready to go with the basic interfaces, then we will all need to give him feedback as soon as possible as to what sort of widgets will be required for our individual models.

Thanks Everybody!

How limited? As you know, interfaces at present are little more than blocks of glorified graphical arrays, but can hopefully be stuck together quickly to produce something useable.

I'll spend this weekend putting together a proposed Demo5 DiplomacyUI and post some kind of demo to my webpage so everyone can see its uses/limitations.

I'll also make the source available for coders to peruse, along with some kind of breakdown on the new classes.

from the interfaces thread I think we are basically squared away. If you need more input, just let me know. One thing I thought of, is that your slider bars need "locking" checkboxes next to them. We will need these in the economy model where the total of all slider settings adds to 100% or something like that. If you don't know what I'm talking about, let me know and I'll elaborate.

All:

I have to get something running at this level as it may give me more ideas about how to use GAs at the higher, strategic levels. I am still trying to get my head around the literature on broader applications for GA. I expect this to take the next week, at least.

I can see a couple of possible approaches at the tactical level, but I am still a bit bemused about employing it for strategic analysis.

For the tactical AI I plan to use a very basic implementation model based on rules and heuristics only. This will provide decisions on tactical movement.

At this level the AI model will need to tie in with the Military so I will need to seem some code for this. Is that you Mark?

Yes, military is me... (I'm praying another programmer will show up, since my dance card is full even Before considering military coding .

My recollection is that I already sent you some code a few weeks ago. If that's correct, then what you have on the military model is basically state-of-the-art. Garth went through several packages and added getters and setters, and better Javadoc comments. I'll gladly send you that update. Just let me know if you need the update only, or the whole thing. I don't have any new code, or even an outline yet for Demo 5. My guess is I'll need to wait a few days and see what exactly the Demo 5 version of the military model looks like.

I know we are all wrestling with issues in our models, and trying to make them the best we can.
But we as a group need to pick up the pace, or it will be mid-summer before demo 5 gets out!

Please don't let the perfect be the enemy of the good. Only through playtesting are we ever going to be able to tell if the models really work or not. Please redouble your efforts to just come up with something that is reasonable for a start that we could perhaps turn over to a coder. Any thoughts that people have on these issues, are of course welcome. Let me know if there's anything that you need out of me.

That said, the only major modification I propose for Demo 5 is that it include the Govt/Social stuff that F.Smith has in the beast. Merging the two code models together (demo 4 vs Beast) is going to be very interesting to say the least. We will need to discuss that sometime soon in another thread.

quote: Originally posted by Mark_Everson on 09-16-2000 09:09 PM Well, this is a somewhat embarrasing thread to bring back to the surface, what with end-of-febuary target dates and such

Don't sweat it, Mark. I've done the exact same thing more times than I care to admit

Ron