Working out a new Economy GUI for Demo 7+

[Mark_Everson]

Originally posted by Mark_Everson
Lee, I like your mockup. Seems very practical. Add a number slider option for military unit builds and some other types and I think you'll be close to a workable solution.

I meant an number spinner for buying X units per turn.

[Lord God Jinnai]

Ok, i like the tree idea (exept leave it in 1 window, just expand it), but have a few ideas to enhance them.

1. The 1st level should be the only one the player has to set. Everything else should be possible for the AI to handle by the following:

1a. Have a button to highlight to say autocalulate. This means that you say 70% on infrastructure, i don't care what.

1b. On following levels, there should be another button to 'lock' the item in, preventing you or the AI from changing it while its locked. So say you want of that 70% for infrastruture, you want 10% (no more, no less) to go to roads, you then push that and lock it in place. The other 90% can be moved around and balanced however then and subsequent turns by the AI and yourself

2. The sliders should always for each level all equal 100%, so lowering 1 thing should raise others.

3. It would be nice to also see a chart of previous spending and previous and predicted future growth eventually.

4. Just though of this, again eventually it would be a good idea to have buttons on the top for people who really don't want to do any micromanagment at all to click things like Wartime Driven Managment, Research Driven Managment, Trade Driven Managment, Exploration Driven Managment, Industrial Driven Managment (industrial as in the more general use, not the more contempary version for factories and whatnot), Maratine Driven Managment, Colonial Driven Managment, Balanced Managment, etc.

[LFS]

Ok, i like the tree idea (exept leave it in 1 window, just expand it), but have a few ideas to enhance them.

I tried a tree, It was my first attempt. Believe me you don't want it as a tree. there is to much on one line that opening it out results in a mess.

1. The 1st level should be the only one the player has to set. Everything else should be possible for the AI to handle

Thats the idea - though we need some AI fo that.

On following levels, there should be another button to 'lock' the item in, preventing you or the AI from changing it while its locked

That is the idea, I just havn't put it in yet.

2. The sliders should always for each level all equal 100%, so lowering 1 thing should raise others.

yup, already in - it just doesn't demo well on a static image

It would be nice to also see a chart of previous spending and previous and predicted future growth eventually.

Just though of this, again eventually it would be a good idea to have buttons on the top for people who really don't want to do any micromanagment at all to click things like Wartime Driven Managment, Research Driven Managment, Trade Driven Managment, Exploration Driven Managment, Industrial Driven Managment (industrial as in the more general use, not the more contempary version for factories and whatnot), Maratine Driven Managment, Colonial Driven Managment, Balanced Managment, etc.

Thats a good idea. multiple preset settings for the minister AI

[Lord God Jinnai]

It would be nice to also see a chart of previous spending and previous and predicted future growth eventually.

Well to see if by spending more on1 more military unit will serverly hamper my long term economy, assuming all things stay the same.

2. The sliders should always for each level all equal 100%, so lowering 1 thing should raise others.

yup, already in - it just doesn't demo well on a static image

Well the 1st level could go over or under to represent ongoing loans and rainy-day funds, but everything else should definatly equal 100%

[Richard Bruns]

Originally posted by Mark_Everson
Richard, that's what the econ info screen is for! ("Econ Stats" button on the orders window) You can see the production level to gauge productivity or the growth rate to gauge effects of spending.

The problem is that I have no easy way of knowing how much of that production is due to infrastructure improvement, how much is due to technology improvement, and how much is simply due to having a lot of population in a square.

Production per unit labor maybe?

That would be nice.

There isn't any site number improvement so far, but we will eventually need it.

So why is it that the squares in delenda have so many more farm sites than the squares in Jerico? I can´t find anything in the terrain or scenario xml files that would cause this.

[Mark_Everson]

Originally posted by Richard Bruns
The problem is that I have no easy way of knowing how much of that production is due to infrastructure improvement, how much is due to technology improvement, and how much is simply due to having a lot of population in a square.

Production is the result of a function of infrastructure improvement, workers, number of sites and technology. Its called a "production function" if you want to search on it to see the functional form. The technology part of productivity is easy to show separately. The others aren't easy to disentangle. Production per unit labor is one way to get info to the player, I guess the analagous per site and per improvement values might give you what you want.

So why is it that the squares in delenda have so many more farm sites than the squares in Jerico? I can´t find anything in the terrain or scenario xml files that would cause this.

That is a hard-coded kludge. All scenarios arbitrarily increase farm sites to the point where the local economy can support the scenario-defined square population. Have to get rid of it at some point. Sorry it ended up confusing you.

[LDiCesare]

As stated elsewhere, there are problems with obsolescence of units. Most of it is a UI problem:

I corrected one miscalculation which didn't account for the number of men in the formula, which can explain some of the problems seen with obsolescence. The bug remains, however.
I do not understand how the UI works. When playing Carthago scenario, every turn, the UnitArchetype isBuildable method seems to be called (through Buildable.buildSquare) on every unit being built. Additionnally, it is called sometimes when the econ windows is opened. It might also be called by the performTests() econ method in the case of Phalanx, though I am not sure. - btw I'd like the performTests() method to be called only if a seed is fed to the game. I wonder if it cause some slowdowns. -

I observed the following:
The first time I open the econ window, it seems that calls are made, but I am not sure on what they are made. Some calls are made only once, some several times (apparently those orders being built are called many times). When opening an econ window for me (Romans), I'd get lots of calls for Phalanx and Elephant (what the Carthaginians are building). I wonder why. I didn't see any call to the things I was building. If I set for example Bowmen to a value, I'd see one call to Bowmen.isBuildable. If I set Legion to something, however, nothing is called, probably because Legion was already set to a non null value.

Calls to isBuildable should be made before showing the econ panel every time, but that is not currently the case. I don't know how that code works, whether the current code can be fixed easily, but the next UI should manage it correctly (by checking every time the window is opened or refreshed if kept open for more than one turn).

[Mark_Everson]

[QUOTE] Originally posted by LDiCesare
I do not understand how the UI works. When playing Carthago scenario, every turn, the UnitArchetype isBuildable method seems to be called (through Buildable.buildSquare) on every unit being built. Additionnally, it is called sometimes when the econ windows is opened. It might also be called by the performTests() econ method in the case of Phalanx, though I am not sure. - btw I'd like the performTests() method to be called only if a seed is fed to the game. I wonder if it cause some slowdowns. - [QUOTE]

Hi Laurent:

isBuildable is called whenever an order exists and is either attempting to be executed or test-executed to figure out the cost of an order. That is why the econ GUI initiates calls to isBuildable. It is probably also called from performTests when the exectution of orders is tested. We can discuss when performTests should and shouldn't be run. Why do you want to connect it with the seed switch?

I observed the following:
The first time I open the econ window, it seems that calls are made, but I am not sure on what they are made. Some calls are made only once, some several times (apparently those orders being built are called many times).

I don't know why this happens. Can you send me the code that exposes all the calls, and I'll try and figure out what is going on.

When opening an econ window for me (Romans), I'd get lots of calls for Phalanx and Elephant (what the Carthaginians are building). I wonder why. I didn't see any call to the things I was building. If I set for example Bowmen to a value, I'd see one call to Bowmen.isBuildable. If I set Legion to something, however, nothing is called, probably because Legion was already set to a non null value.

No idea on this one.

Calls to isBuildable should be made before showing the econ panel every time, but that is not currently the case. I don't know how that code works, whether the current code can be fixed easily, but the next UI should manage it correctly (by checking every time the window is opened or refreshed if kept open for more than one turn).

If everything on the tech end tests out, I'll look into the econ side and see if I can find the bug.

[ogj20]

Treeview

I'm not sure why you say that a treeview would look horrible, I'd think it would look quite nice and provide the functionality we need, I don't see why it should look vastly worse than the mock up attahced (only a straight graphics cut and past job, so I'm not aware of any kind of coding considerations here...)

(There should also be a scroll bar on this screen, for when the list is expanded)

I think there should be an overall spending category, where you can decide how much of your tax income to spend (this slider should be able to go over 100% to allow you to (try to) overspend if you have money in the treasury.

I think there should be an option to let the AI handly distribution of funds within each category, so each expandable node will need a checkbox to turn on the AI, it should still be expandable to see what the AI is up to though...

Of course some categories (i.e. military) will require both a percentage allocation scroll bar and a total production target. I agree that it would be good to be able to set a desired force composition and let the economy build up to that, at the rate set, with the priorities you set, to do this I suggest the target be set in terms of total numbers, and the existing numbers be shown.

Others (roads, walls) will require a method of aiming either to spend x% or build them in x turns. I'd also suggest these be split into a build new category, and a maintain existing category, since this should be easier to work.

Others will require a ROI setting, though I think this should probably be described as "Turns to pay off", since this is easier for non-economists to grasp (or at least it's how I interpret it when deciding what level to set). These are the hardest, since they will affect the total amount spent in a way that is fairly unpredictable. I suggest that the economy turn processing needs to be set up to spend money in priority order, then if there is any left over (i.e. some investments weren't worthwhile) to go back and spend the leftover money in the proportions chosen within the investment category, then if it finds it runs out of worthwhile investments, to go back and try and spend the money on all the other categories in the chosen proportions.

Actually you could have a toggleable setting which decides whether to send left over money to the treasury or tries to spend it, this could either be by category or global.

[ogj20]

p.s. how do you include images that show in the message, rather than as an attachment

[LDiCesare]

Can you send me the code that exposes all the calls, and I'll try and figure out what is going on.

Just put a trace in UnitClass.isBuildable method.

About the sliders: I suggest they should be a bit smaller than what we saw in the screenshots, and the figure boxes if we use a tree should be aligned to the right of the window. Otherwise, it looks good (certainly better than what we have).

[Mark_Everson]

Hi Owen, I think your observations in your longish post above are pretty good. Some of them were planned already. I do like your suggestion for reviving the treeview. What do you think Lee? I also think the sliders should be shrunk, probably fairly significantly, which would certainly help with the treeview part.

Originally posted by ogj20
p.s. how do you include images that show in the message, rather than as an attachment

Believe it or not, the info is in the faq, you should check it out sometime. Here are the relavent formatting codes.

[Mark_Everson]

What do we do when an old order becomes obsolete?

I will be trying to get the order obsolescence stuff going today. I'm sure it'll be pretty easy to do. But that leads to a question. Suppose I'm building Warriors at 60% when Warriors become obsolete, and the order is no longer available. What do I do with the order?

My intention at the moment is to find the new order type, Legion or whatever, that made the previous order obsolete, and make an equivalent order for the new category. I would also as a first shot take any invested point in Warriors and transfer them to Legions. Does this sound ok? If not, speak up soon!

[Martin the Dane]

Well I must say I preffer a treeview, to a number of windows popping up. Hmm. I guess I've said that before but I still think that the fewer windows the better. I hate having to close several windows to get back to where I was.

I could live with Owens sugestion, though I'd preffer a slimmer look.

[Richard Bruns]

I think that I prefer the popups to the treeview, but I don´t feel strongly about it either way.

Right now it seems that a unit can only be made obsolete by another unit, and can only have one unit that makes it obsolete. I would like this to be more flexible in the future, for example by giving the possibilty of an event to make a unit obsolete even though there is no replacement. Is there anything in the code that requires this current inflexible structure?

In addition to unit obsolescence, I would like to see, fairly soon, the possibilty of Activities becoming invisible, as described in the tech thread. This should be slightly easier to handle, since everything invested there is spent every turn. The order would simply vanish from the screen, with no other changes.