Object Builder: Bug Reports

[axi]

Hey, F_Smith!

Were you changing stuff in the beast, just now? It got stuck on me.

As for the "add new EG" button, err, I think you have to fix it. I only get a horizontal line where it is supposed to be. I click on that line several times and I get two of the "select base EG" popups. Then the beast gets stuck.

No, it's not you after all. It's that bloody MSIE5 that crashes half of the time! Please, by the end of the day, pack up all the classes and a batch file in a zip, because this thing is getting on my nerves.

Btw, nice change in the civ window. And special actions too! The window headers, are the supposed to be titled "Foul Play" all of them? Are all these done by the book (the govt model) or by improvisation?

No it's you after all! I was doing nothing, something changed on the map, and crash!!!

As for the "Macchiavelian Ruler" problem, I think that we should leave it alone until the riots model gets implemented. I have a hunch that Rodrigo won't like your approach very much. As you can see from my quotations, he seems to insist in the importance of negotiations over policies (and of negotiated results), as opposed to the mere "passing" of these policies through a voting body. Your approach looks pretty much like what I had suggested and pretty much like what happens in a Parliamentary Republic, while Rodrigo's is more alike the negotiations between the companies, the state and the trade unions, for the annual Collective Labor Contracts (I hope you have them in the US too).
[This message has been edited by axi (edited August 20, 2000).]

[F_Smith]

I'm afraid 'negotiated' values is just not fun, and would cause a lot of micromanagement as players adjust their preference up or down by trial and error to get exactly that % pref they want in the govt.

So negotiating contract values was the model's inspiration? That is not an example of political decision-making. We need a political system here.

It also allows manipulation of the system that is not intended.

Man, I need some feedback and direction here.

P.S. -- that's improve, for the special actions. They don't actually do anything yet. The text was just to be a placeholder/example. Altho they sound fun, to me. Especially when you consider possible consequences . . .

[axi]

BUGS is my issue this time. New features attrack attention away from old features, so old bugs get overlooked. I accidentally stepped on a couple of bugs of the "Build/Edit Base EG" window:

1) After the addition of a fourth EG, the list acts funny; sometimes it can't get down to the 4th EG, sometimes it reverts to the 1st or 2nd EG, sometimes it just shows only 2 EGs. I remember that I have pointed out before that there is something wrong with that list and so had LGJ. It was then partially (?) fixed, so that the 4th EG was selectable, but nevertheless, the list is still buggy. Oh and one request: make it bigger, as you have done with quite a few others.

2) After building a new (not pre-made) Base EG, I tried to add some of it to a square. Although the Base EG had a Religion and a Culture selected, the local EG had none of the two; it had inherited only the name of the Base EG. Now this doesn't seem to happen with the pre-made Base EGs. Any ideas why?

3) That problem I mentioned with the "Add new EG" button for the mapsquares, when will you eventually fix it. It looks like the button was given a height of 0, so that the button was compressed into a line (which btw works fine). Can't you see it?

4) Another odd bug. After inserting the population of a new EG on a mapsquare, one click at the white background of the Editor (in any place, even up, above the map) would bring up again the population popup, which of course would not go unless one reinserted the population of the EG (you will have to put "cancel" buttons everywhere eventually). One click at another square of the map will fix this problem.

What I think you ought to do is to write-up and preserve a "known bugs list", containing observable and non-observable bugs. I've made the start for you. The benefits of such a list are:
- You will not forget to fix bugs, after they are reported and verified.
- We will not get in the trouble of always reporting the same known bugs.
- If you are unable or unwilling to fix a bug, someone with java knowledge will be able to check-up the source and provide you with the fix.
- The beast will cause us less frustration, once we know what to expect.

On a totally different issue:

quote: Man, I need some feedback and direction here.

I am doing what I can, but if you haven't noticed, you and I are alone here, for the time being at least.

About the political system: I will make another large quotation concerning this subject and hopefully this will make things clearer, especially in what concerns Rodrigo's point of view. Since he is not present, let his own writings speak for him. I only hope that you will be patient enough to read through all this.

Edit: The quotation was moved to the govt model thread.
[This message has been edited by axi (edited August 21, 2000).]

[F_Smith]

Axi:

1) It's still on my list of things to look into, but I'm not really debugging the GUI components at this juncture. This week, I'm focused on the turn logic.

After the functionality is complete, then I'll debug the GUI stuff completely, possibly next weekend.

It is worth it to mention the bugs, tho, because when it's time to do that part of the debugging I will go back thru all these posts and check to see if all these issues are taken care of.

And I do appreciate your feedback here. It is very helpful. I had just hoped for more input from Mark and Rodrigo, at a minimum. I kinda thought I had committments to be available.

Ah, well.

I do think I understand what Rodrigo wanted, but I don't believe it's workable, from a 'fun' standpoint.

The results while playing are not good. His system's focus is on arriving at a final number based upon inputs. We need a system focused on how things change, since that is what happens during gameplay -- you only deal with making changes, never on creating a value from nothing.

The results aren't realistic, also. When you want to lower ethnic discrimination by 5%, say, you have to mentally calculate how much to lower your preferences to 'manipulate' the system. Then, even if all other forces are against the change, the value goes down.

If I don't get some guidance on this by tonight, I think I'll leave one or two of the variables as 'negotiated' policies, and the rest I'll code up as 'approved/denied' policies. So a side-by-side test can be done, they can change each value and see which works best.

[roquijad]

I couldn't check the beast on saturday. Sorry. I haven't been there yet. I stopped by this forum thread before checking the beast so I'm gonna refer now only to the dicussions above.

F_Smith: will you code the govt model or whatever you want to code? It seems to me you're doing whatever you feel is fine with only a vague reference to the govt model Axi and I developed.

I can't participate in a process where the model is being built by improvision at every step of the way. I simpy can't. Sorry.

If you want to code the model Axi and I created, then give me a call.

[F_Smith]

*Sigh*.

Well that's just swell. My code has been rejected before it had even been tried.

Rodrigo, this is still 90% of what you designed. There was just a problem that a spreadsheet couldn't show. The other half of the design team even spotted this problem early, and pointed it out to you. The problem continued to come up during testing. That problem may require a change in the system.

That's how it works, isn't it? Testing has to be the final answer, doesn't it?

My last proposal was to code it both ways, for a round of tests. Surely you are not against this?

* * *

This certainly seems like 'ownership' problems, I must say.

This is what I meant when I complained about 'not working with programmers'.

I feel I'm being treated as a lackey, to be told what to do. I'm just the worker bee, to your Queen.

What kind of programmer do you think is going to just 'do what he's told', even when he comes across a problem? Do you want anything coded by that kind of programmer?

Most any good programmer would have pointed out this problem, and offered suggested solutions.

* * *

Perhaps I should just drop this, then? Rodrigo clearly would prefer I forget it, he apparently has a different programmer in mind for this stuff. Should I halt development?

[Mark_Everson]

Hi all:

I've been out all day, and just have had a brief look here. I hope what I say will help rather than hurt... I haven't read much of what was said today, but I think I get the drift.

I saw F_Smith's post this AM about the negotiation process. While I thought his proposal to just drop the way we have been talking about was premature, I think his 51% solution has merit and Should be tried.

Since he has offered to code it up both ways, I think that's Better than one way or the other. We will then get to test the two methods for both fun and simulation of reality.

We of course should not make early decisions based on what is there now, because features are missing (riots etc) that will reduce the capacity of the player to just 'overbid' what they want. But we can start the process of examining the two in play!

So please lets pull together on this...

Please.

If the govt model is coded as designed and there are also a lot of little switches that allow alternatives to be tried, we will all have what we want I think... A way to test the Fun of the 'base' system and lots of alternatives. I don't see anything wrong with that.

Can we all agree that that is the way the Beast/govt coding should proceed?

[Richard Bruns]

At first I thought that the object builder was a design utility for people who were familiar with the models, but now it seems to be a demo for playtesting. I am not familiar with the government model, but I would be willing to playtest it. Would it help if I downloaded the thing and experimented with it?

If it would help for me to work on it, how do I download the thing?

[F_Smith]

Mark:

I'm sorry if I got a little bent.

Long weekend watching both kids, I guess. Yes, I'll do both ways up tonight. Heck, for an example of the '51%+' approach, just go to the beast and play with the tax rate.

But tomorrow, we'll be able to test both ways.

* * *

Richard:

Click Here for 'The Beast'

Note: All the new work is in the '4000bce' scenario.

Yes, please do check it out. The beast is a 'prototype' for the govt, population and some 'ruler' stuff.

It needs heavy testing by all concerned.

Be aware that the User Interface stuff is still very crude, all the work goes into the plumbing of the data model and the turn code. But this is a prototype of what will be the actual guts of the game code's architecture. Any substantive corrections/comments/criticisms are welcome, and will be fixed. Just be kind. Remember that there's thousands of lines of code here, so if a dozen or so are completely off the mark please tell me so with kid gloves! I will fix it anyway you want it to be, altho I may offer alternatives, too . . .

Prototypes are intended to be run thru 'iterations' of correction and debugging cycles, until you eventually end up with a final product.

That's the way software is built, now-a-days.
[This message has been edited by F_Smith (edited August 20, 2000).]

[axi]

Uh, oh, I told you so...

Ok guys, lets just relax a little bit here. I know that F_Smith wants to have a productive evening tonight, so let's try to be helpful tonite and from now on. Everybody should get in the trouble of checking the beast more often, because the more eyes we've got, the faster things will progress.

I agree with F_Smith and Mark that we should try both approaches, to see what's best. Specially since we can't reach an agreement simply by theorising.

I propose that CR, FP and RD (or ED) should be coded through negotiations, while TR, SL and ED (or RD) should be done the other way. As for the INPs, we must do them as a combo, so, after classes, ideologies, support shares, etc are put in, we can try both the negotiations procedure with the equilibrium point and the electoral system I had once proposed (or any other "majority rule" system). Of course we shouldn't write off the negotiations approach, until the riots model is added; only then will Rodrigo's approach have a chance to prove that it's workable. As for my personal opinion, I won't be sure about anything, unless I see that it works or that it doesn't. I have in mind too many arguments for and against both ways; it is not easy to foresee what's best.

All further theoretical discussion on this issue should be held in the govt model (LGJ has posted there already). Unless I'm told not to, tomorrow I will move all the relevant quotations to the govt model thread, because they are really OT here.

F_Smith, if you don't mind, I would like to ask a couple of things from you:

1) Please, whenever you have spare time, read thoroughly the govt model thread and also, whatever in the govtecon thread that might interst you. Whatever questions/comments/objections you have in mind, please ask them. Rodrigo and I have put quite alot of thought on this and it would be a waste if it got overlooked during coding. Rejection hurts, you know it better than me.

2) Also whenever you have spare time, please give us more explainations on what exactly your "improvisations" consist of (in other words, the system's actual design). It would be helpful if we could find faults in your "real" model, exactly as you find them in our "ideal" model.

3) At least fix bug #3! It is outrageous!

My watch shows 3:45, so I'll call it a night. You'll hear from me tomorrow.

------------------
"In a time of universal deceit, telling the truth is a revolutionary act."
George Orwell

[roquijad]

Well, I'll play the bad guy, I guess. It's not that I believe our model is perfect or anything like that, but the same way you, F_Smith, cannot work as a bee for the queen, I cannot work under this style of yours where just EVERYTHING is changed with little arguments behind. It's impossible for me to work asking "what did you REALLY code?" as Axi did in the last post. What good will I do checking the beast? I'll see a lot of stuff that I won't be able to understand and that I couldn't know how to link with the model we developed. I won't be able to help!

I don't believe in the "we have to test and do less theoretical work" statement. We all are not that stupid. We can foresee what pros and cons a system has. If we start coding without thinking ahead, then we'll end up with a whole system that unlikely will be what we want.

I'm not stopping you from coding whatever you think is best. I can't and I don't want to. It's just that in practice I can't help you with any invetion you come up with. I don't want to spend my time trying to understand your system and then trying to link it with the rest of the model (if possible). I don't know how to work in this "code first, then analyze" style.

I'm not upset. I'm just saying that I can't help you if you start your own "de facto" model. Since you're coding two models now, you can be sure I can help you in anything regarding the model Axi and I created. As for this attempt for an alternative approach, I'll step aside. I'm sure Axi will be very helpful.

Just keep on going, F_Smith, I don't want you to stop.

[Richard Bruns]

When I tried to modify the 4000 bc scenario, I couldn't do anything other than change the civ's borders. I think the trouble might have been in the connection; everything was going at a good speed at first but then slowed down horribly.

I was able to open the Create Culture window, but the culture did not appear after I altered it. There were a lot of nonfunctional windows, like Edit MapSquare. Generally, I couldn't do much at all. Do you have any idea what might have happened?

[F_Smith]

Axi:

1) I have read thru it all several times. I spent about 4 hours yesterday going thru it all again. That's how I come up with the problems I'm finding. It's a big, complex, far-reaching model. I've had to build a gameworld, build a population, religions, cultures, leaders, governments, and on, and on. I'm doing the best I can. But only ya'll know it as well as you do, so only ya'll can answer some of the questions.

2) I'm sorry, thought I had been clear: The only 'improvisation' in there is the voting method for changing a govt value.

Everything else is exactly what ya'll designed, as ya'll designed, or at least as far as I understand what ya'll designed.

The improvisation was the method for changing a govt policy. The ruler (or someone in the govt) 'proposes' a change to a policy, then all parties with political power either vote yea or nay, depending on the equations you and Rodrigo have put together. More yea than nay, it passes. Otherwise the change fails.

3) That's almost certainly because I'm basing window size on screen size, and I have a 19 inch monitor. My bad. It'll be easy to fix, but right now I'm working on the 'turn logic' code. I'll try and take some time to re-arrange that panel. I'm just really putting cosmetics on the back burner, for now. Later, all the GUIs will have to be reworked in detail.

* * *

Rodrigo:

No, no bad guys here.

Just a clash of cultures.

I'm afraid everything has to be open to change, if we are to end up with a great piece of software. Testing will be the only way to answer most questions. Some core assumptions will have turned out to be wrong. It always happens.

It's not that we're stupid -- it's just that programs are so complicated that there is *no one*, *no where* that can fully forsee how all the details will work out.

Testing is the only way to do this right. It's a basic rule of programming. We ain't so smart we can write all this code in our heads and have it come out right the first time.

Ever written a 5,000 line story, article, thesis, whatever? How much writing, re-writing, re-thinking, starting over, etc do you have to do?

This is like that, only times 100. I'm manipulating words to tell a computer *exactly what to do in order to simulate this gameworld*. I am not the 'Mozart' of coding that I can just write it from beginning to end in my head. There are a dozen ways to do most any of the hundreds and hundreds of tasks we must do in code. I won't know which is the best way to do each until I try them. Each will have a different result in the final product.

And you can help with this process. You *must* help, if the finished product is to be what you want. Ever seen software that programmers write when they don't prototype? It has nothing to do with what the client really wanted.

Your part shouldn't take 10 minutes a day. Look at what is there and comment on how close it is to what you wanted. Because that is the deal -- this is not 'my' system. This is my attempt to give you the system you wanted. There are going to be problems with your design, and I have to be free to suggest solutions. You have to work *with* me, not tell me what to do then send me away. We're members of a team. We work together. I need your feedback. It moves the process along 10 times faster.

That's the key. You're writing the requirements here, so you must work with the programmers to make sure they are clear on all the thousands of little points.

There are problems with *every* design, in the software world.

Only prototyping can give a finished, working product that is bug-free in a reasonable amount of time.

* * *

Richard:

Good point -- I haven't written any test cases.

That is a big part of all these problems, isn't it? I see. Richard, you have actually been extremely helpful.

I'll explain my stupidity in that only by saying that the datamodel is completely debugged. All the data stores, loads and changes flawlessly. The 'turnhandler' code is about 80% bug free. It's approaching being finished, as soon as I put in live numbers tonight.

The Gui, however, is full of bugs.

I'll write a test case pronto, next post, to help guide thru that minefield I've called a Gui.

Do remember, that Gui is just a tool for use to view the data. It's not going to bear any resemblance to the final Game front end.

[axi]

F_Smith: You were right about bug#3! I run at 1280x1024 with my 17" screen and when the editor window is maximised, the button shows up alright. We can live with it, for the time being.

If you want us to really try out the beast, please fix the "Edit power" function.

[F_Smith]

The 'edit power' button should work.

I'm debugging the last of the 'negotiated' policies' formulae now, so it will work exactly as Rodrigo requires. I'll upload it in an hour or two.

Then I'll put in a switch to allow the choice of either system, from a pulldown menu I'll call 'Game Options', so we can play the other 'govt' game, too.