The future of GGS?

[Leland]

Let's see if I can make a poll out of this...

[VetLegion]

Good options. I guess everyone has to decide for himself.

Project was not exactly born dead , but has come to a virtual stand as we speak.

Everyone has his/her own reasons, and I agree, two years without even barely playable demo is really, really bad performance.
And after chrispie stopped work on the UI, all coding seems to have stopped.

So when people voice their concern or leave, it is ok. If Dan and Mark want to shut down this section of the forum, they have the basis for it - lack of activity.

However, I will remain on the project, even if only on the mailinglist and the web. I have put a lot of time in it and I want to finish it. I can not make great promises. I am a coder of moderate skill and not too much expirience. That is why I hesitated with a lot of stuff, like event system and client/server framework, I thought someone expirienced might do it better. A lot of core stuff I dont really know how to do. I And chrispie made excellent UI , but events, after some discussion, were never done.
I dont remember exactly why, but I guess noone picked up the task and said "ok, I ll do it"
neither have I, so - guilty.

What is great are the ideas people had and posted. I dont think it is all very megalomaniacal. It does however require work to implement, and I for one havent done it.

So the basic line is, yes, programmers are the critical part here and will remain to be critical part, that is the fact. If you know a good coder, invite him to GGS

my option is with the shaping up then, I dont have intention of leaving.

[Leland]

On the programming side, I think it's pure lazyness. We did some serious pondering about our alternatives and even decided to look into the Zthread library to do our event handling stuff, but then the whole thing kind of came to a standstill. My excuse used to be that I didn't have the disk space for DirectX libraries, and that is not an issue anymore. Hmm.



So if we assume that in some bizarre alternative universe I decided to get the DirectX, chrispie's code and whatever we've got and just bolted it together with Zthread-based event/thread code, would anyone else be available to join the programming effort in the slightest?

[Nath]

So if we assume that in some bizarre alternative universe I decided to get the DirectX, chrispie's code and whatever we've got and just bolted it together with Zthread-based event/thread code, would anyone else be available to join the programming effort in the slightest?

As I've said, I can't program much, but if there's anything I can do, I'll try.

[chrispie]

After I voted The megalomanic option, I have to say that's probably a bit too strong for my opinion. I'm more of the opinion of our ideas being sound, we just need to get on and start making some of them work. After several months of having mislaid my pick-axe I now think I've found it again, and I'm ready to start hacking at that big rock ( metiphorically speaking of course!! ).

I'm by no means an expert programmer, heck I've never even worked on anything this big before - I don't think we need genius programmers. All we need is semi-able programmers, and more importantly a vision to make it all fit and work together as a whole framework. I think any level of programmer can take part, maybe with some guidance from the more experienced.

It's wholey possible that we got to caught up in pety details like that of the event system, which, although crucial to the system doesn't necessarily need -that- much thought, once we've agreed on the basics of it. I'm all for it being an expandable systems, which starts with the simple, and grows as we need it to. There's nothing wrong in programming with doing something, deciding it doesn't quite work, and changing it, we don't have to write it perfect first time - I think we all know that - we just need the confidence to write the things and not worry about failing. That's the big advantage of writing a modular system, it's much easier to see where the problems are when it's all being run in seperate parts, rather than building a bigger and bigger single 'thing'.

So anyway, lets all just pull our socks up (metiphorically speaking again! ), and start thinking about what we can do to fix this little set back. I'm gunna have a look at the UI now, see if I can get it a bit better, and maybe add some new stuff in there. I'll also check the other stuff I did, like the chat thingy and the event stuff, see what I can work out from it all.

Chrispie

[Leland]

I tried to get the old UI demo to compile today but for some reason it didn't work out. I'll have to try again tomorrow.

[chrispie]

I posted the up-to-date UI as is needed to work (in theory) at www.baird78.fsnet.co.uk/GGS.zip (it's case sensitive). All the code, exe, graphics everything in that zip. The whole shabango for the UI, and even a tiny bit for the map too. I think you're missing the png library possibly, not 100% sure.

[VetLegion]

Great to see you here, chrispie!

[Leland]

Alright, I finally cast my vote. I was at odds whether to pick "project organization overhaul" or "shape up", but finally I ended up clicking the latter. I can understand that some people are terribly frustrated by the lack of results, and are therefore not the least interested in the project. My personal viewpoint just happens to be a little bit different: I don't care about the destination, I care about the journey. To me, GGS provides a lot of food for thought, be it social modelling, playability issues, software design, object-oriented programming, project organization, technical writing, historical tidbits and a plenty more. It's kind of like an intellectual exercise, and I dare to say that even if the project died right now I wouldn't say that I have lost a possible future game, I would have lost a current hobby.

I wouldn't say that the project is dead, that's up to all of us to decide. It's not dead as long as there are a couple of crackpots making stuff up, and it seems that I am not the only one whose got a few loose screws. So what if we're the laughing stock of alt civ community! I am not here for social acceptance, damnit. Nevetheless I'm kind of surprised that so far only one vote went to the "born dead" alternative, that's a positive sign.

The second alternative, that we should move on and try to get our ideas working in some other projects is basicly the most viable way of getting something done. I agree what korn469 said in another thread about us being better off making a mod or something else if we really think we have something here. Again I only have my personal views to argue for: I am not interested in doing something easy, half the fun is that this is an ambitious project which strives to create something novel, and something excellent. World is filled with small little games and mods nobody cares about, and I have no passion for contributing to that kind of things. After all, this is a hobby, not work, I should be allowed to daydream a little bit and fight the blasted windmills as I see fit. The same reasoning pretty much applies to reducing the goals of the project... for me, that would just limit my own freedom of choice as to what to do.

I don't know if this makes much sense... well, it is just my personal rationalization. I think it would be a shame if GGS suffocates to death, but to me it would be the project that I'd miss, not the product. And hey, if my little contribution can help in creating something concrete in the side, all the better!

(It would be nice to hear opinions from those who voted items 1-4, just to get a bigger picture what to do. I think the results of this poll should define to course we are about to take.)

[Nath]

Is it possible to make the cursor transparent?

[Nath]

A question with lines 45-62 of MapManager.cpp:

long i=sector[l.x][l.y].pop;
// if not on huge list, then just add amount, and check if we need to be on huge list
if(i<40000)
{
	sector[l.x][l.y].pop+=(short)amount;
	if(sector[l.x][l.y].pop>=40000)
		addSectorToHugeList(l);
}
else
{
	if(sector[l.x][l.y].pop>=40000)
		huge_sectors[i-40000]+=amount;
	else
	{
		sector[l.x][l.y].pop+=(short)amount;
		addSectorToHugeList(l);
	}
}

I'm not sure I've understood this -- if (i<40000) is false, wouldn't (sector[l.x][l.y].pop>=40000) always be true? When would the inner else thing run?

[Maniac]

Hurray! Great speech Leland! Couldn't have done it better.

[korn469]

The second alternative, that we should move on and try to get our ideas working in some other projects is basicly the most viable way of getting something done. I agree what korn469 said in another thread about us being better off making a mod or something else if we really think we have something here. Again I only have my personal views to argue for: I am not interested in doing something easy, half the fun is that this is an ambitious project which strives to create something novel, and something excellent. World is filled with small little games and mods nobody cares about, and I have no passion for contributing to that kind of things. After all, this is a hobby, not work, I should be allowed to daydream a little bit and fight the blasted windmills as I see fit. The same reasoning pretty much applies to reducing the goals of the project... for me, that would just limit my own freedom of choice as to what to do.

Leland basically what i meant was taking an already working engine then changing it to meet the needs of the project so that ideas could be tested from the begining instead of trying to build something from the ground up

i think that disease, religion, social stress, and many other factors have had a major effect on how history turned out...if the native americans had of been immune to small pox and had of infected the europeans with a disease that spread to europe and killed about 40% of europe's population in a one year time span i think that things would have been very different and it would great if the game could model something close to that

the way Civ3 handles strategic resources is both fun and actually give the game more depth than civ2 had and if GGS could adopt in depth but easy to understand models (they don't have to be simple, but once a model does its job any any extra complication to it is a waste) then it would be a fine game

however until there is even a half way playable engine (and i'm talking about just prealpha peices of each part) then basically the project is only fun to particicape in if you are an artist or a programmer...so if this is still a year away then the project will most likely only grow as more programmers or artists get added to the project

i dunno, i don't wanna see the project die but i don't know what it will take to get the project to grow into a thriving alt civ game

[chrispie]

A question with lines 45-62 of MapManager.cpp:


code:--------------------------------------------------------------------------------
long i=sector[l.x][l.y].pop;
// if not on huge list, then just add amount, and check if we need to be on huge list
if(i<40000)
{
sector[l.x][l.y].pop+=(short)amount;
if(sector[l.x][l.y].pop>=40000)
addSectorToHugeList(l);
}
else
{
if(sector[l.x][l.y].pop>=40000)
huge_sectors[i-40000]+=amount;
else
{
sector[l.x][l.y].pop+=(short)amount;
addSectorToHugeList(l);
}
}
--------------------------------------------------------------------------------


I'm not sure I've understood this -- if (i<40000) is false, wouldn't (sector[l.x][l.y].pop>=40000) always be true? When would the inner else thing run?

Hmmm, err lol! Yep you are right there Nath, I seem to have wrote a few lines of useless code there, dunno what I was thinking. The MapManager stuff is still pretty much useless anyway at this moment, none of that is implemented as yet, just something I was playing around with.

Good to see people are reading my code at least, hehe.

I would be happy to hear peoples views on the UI in particular, if anyone has any ideas for improving it (code wise or graphic wise) I'd be happy to take them on board.

When you say the cursor transparent, you mean make it semi transparent like the windows or have holes in it?

[Leland]

Nath: Subscribed to the new mailing list yet? I'm also in the process of reading and trying to make sense out of chrispie's code, so if you've got questions or comments or anything it would be quite welsome if you could take it to the mailing list (or here, if it's easier for you... I still maintain that programming and design should be kept separate).

M@ni@c: Thanks for your compliments, I notice that after my "speech" there were two new votes to "downgrade expectations"...

korn469: Excellent points. Prototyping could be a very effective design method, the problem here is again our scarce programmer resources. It's fun to make things from scratch ("it might be junk, but it's my junk!") and learning an existing engine is another threshold to overcome, so at least I haven't bothered looking into any of those, especially because of the nagging feeling of having to throw the prototypes away at some point (or settle with an engine that might be inferior to our purposes). I suppose this is again a personal thing: I think Vet and Nath are more prone to just pulling up their sleeves and just doing something (or not, I don't know for sure...), whereas myself and chrispie prefer to endlessly gloat over implementation details (examples: chrispie writing a whole new general purpose GUI when MFC or some other existing windowing system could've worked just as well for starters, and me slowing down the whole endeavour with rather silly event-based architectural considerations).

Well, it all boils down to what you said about artists/programmers (and lately not even our only artist has had any fun...): the bottlenecks of this project are those who can code, and the designers are rightfully frustrated by the lack of results. I mean, the mere fact that we have programmers implies that when someone writes a detailed design and gets positive comments about it, the design has a fair chance of being implemented and seen in practise. This, however, has not happened, and I think that could be interpreted as programmers failing to live up to our name and making fools out of designers. This is not such a good thing, and I hope that it is not too late to redeem myself and actually getting some programming done soon (unlike last year or so of empty talk).

Anyway, if I look at the poll results it seems that a slight majority wants to reduce our objectives. My personal vision for the project is that it's heart and soul is that it's a "next generation" (that term is terribly overused!) game built from scratch and a lot of things other civs take for granted are thrown away. My take on this is and has always been to build it incrementally by doing the boring groundwork first (software architecture, map, population modelling, and a million other things before getting to actual playability), so I think I lack the ability to tell which goals the rest of you consider important and which are expendable. Any suggestions? Can we do this with a regular Civ map, for instance (my opinion is definitely NO, but I am asking anyway...)? What are our core principles that separate us from other alt civs?

chrispie: The transparent controls look really cool, the only problem is that they aren't really transparent, they just copy the background whenever it's updated... if you move them around the background stays the same and the illusion of transparency is lost. Hmm. I wonder what it would take to program true transparency, the kind of which isn't lost even when moving the controls around?

[Nath]

so if you've got questions or comments or anything it would be quite welsome if you could take it to the mailing list

Ok.

When you say the cursor transparent, you mean make it semi transparent like the windows or have holes in it?

The holes thing. The cursor is in a white box:

[chrispie]

From Nath:

The holes thing. The cursor is in a white box:

Hmm, that's not supposed to happen! It doesn't actually do that on my system, does it do that on anybody elses? On mine it is really transparent, just like a windows cursor, very bizarre.

From Leland:

chrispie: The transparent controls look really cool, the only problem is that they aren't really transparent, they just copy the background whenever it's updated... if you move them around the background stays the same and the illusion of transparency is lost. Hmm. I wonder what it would take to program true transparency, the kind of which isn't lost even when moving the controls around?

Hmm, the windows move? Lol, I think I've not quite seen your point here leland, do you mean the windows move or the background moves? When the bg moves, the picture in the windows 'should' change too. Though you saying this has made me think that maybe I could just ignore the pink colour of the fill, and draw a semi-transparent white block over that, that might work better than copying.

I really need to work through the whole UI, to be honest I'm not that happy with it now, some of the stuff in it I just don't like the way it's done. It might take me a few weeks, but I'm going to concentrate my efforts on improving the way it works, I want to be sure it works well for everyone too. So please start posting anything about it that doesn't work for you, and/or logfiles if it doesn't work at all.

Chrispie

[chrispie]

Also, I'm reviewing the use of messages in the UI. I've come to the conclusion that the message system may very well be a big waste of time. I think the only need for a message system is in queing, but it doesn't do that! I know I'm talking about some big changes here, but I want us to have a very good UI, which I'm by no means convinced is what it is now.

[VetLegion]

Also, I'm reviewing the use of messages in the UI. I've come to the conclusion that the message system may very well be a big waste of time. I think the only need for a message system is in queing, but it doesn't do that! I know I'm talking about some big changes here, but I want us to have a very good UI, which I'm by no means convinced is what it is now.

we use win32 queue if I recall? I think we should switch to DInput. queuing is absoutely necessary, I have learned so while I was doing UI without it. many mouse states get dropped if you poll for them in a loop or something.

about messages, why drop them? they work, just adapt them not to use win32

[chrispie]

Hey Vet

I'm sure the UI did originally use DInput, but I took it out because it was slow and you had to call update to make it do anything, meaning it's messages are heavily attached to the code of the program.

The win32 message system is used now because it runs independantly of the UI thread, meaning that in theory, it shouldn't drop messages. Though my UI has a seperate msg system - that's what I was gunna drop until I realised it uses functions like sendMessageToAllChildren, virtually impossible to duplicate in the form of normal functions.

[Leland]

Originally posted by chrispie
Hmm, the windows move? Lol, I think I've not quite seen your point here leland, do you mean the windows move or the background moves? When the bg moves, the picture in the windows 'should' change too. Though you saying this has made me think that maybe I could just ignore the pink colour of the fill, and draw a semi-transparent white block over that, that might work better than copying.

I meant that when you move a semi-transparent control around, it doesn't change: the only time it is transparently drawn is when the background changes. I attached a screen cap of a window to give an example of what a window could look like after a little fooling around. Another thing I noticed is that controls that are covered by other controls still seem to catch (some?) mouse events.

About queing, we'll have to include it in one form or the other anyway due to communication between different computers/threads, so perhaps we could refashion the current message/event system to use a queue? As for the message queue of Win32, I see no problem in having it in the background like they are now, that's an implementation detail, IMHO.

[chrispie]

Hmmm, I changed it so that it does like you say. It works fine as long as it's just plain transparent, but when I apply the alpha to make it see through, it slows down the drag quite a lot, I think it's just asking too much from most systems. I'll look at the code to see if I can change it, but right now I don't think it's very feasable.

[ElmoTheElk]

Good to see that something is happening again.

As I stated on the site: I am not really interested in keeping up the site (maintaining it). However, as not really good explained in that same post), I am happy to draw more controls, if needed, for the UI. (Altough a totally different version is a lot of work...)

About the transparence: The way I did it in Adobe (Photoshop) was the same as you desciped above; and it was the way I thought you implented it: Draw a white, semi-transparent rectangle (as a background for that window). This would prevend the problem some of you might have (as stated in this post). Don't know if we should change it... Maybe we should go for the less processor-consuming option.

Elmo

[Leland]

Maybe someday we have the luxury of using hardware acceleration for the transparency effect, but for the time being it's not such a big deal. How about making it plain transparent (or filled) when dragging, and semi-transparent when they're still? Would it look odd?

[RonHiler]

I'm disturbed by what I read here, guys. It looks to me like you are losing hope. This can't be good.

I will offer some advice. You can ignore it if you like. But this comes from experience, and maybe it will help.

I have programmed MD all by myself for over three years now, going on four. About a month ago, I took stock of where we were at in development. The picture was not pretty, I must admit. I suspect this could be the same place you are now, judging by what I am reading here.

In my case, MD was pretty far along. But looking at the list of things still to do, I figured we had another year left, easy. The plain and simple fact was that development was too slow. Interest was waining, both on my part and on the part of the fans of our project, and something had to be done.

A few weeks ago, the MD source code went semi-open, and I recruited help from our followers to complete the program. I tell you in all honesty I had many misgivings about this. We're talking about code I had spent over three years working on! It seemed a huge risk.

But it turned out to be the best thing I could have done for the program. I am once again confident that MD will be complete, and sooner than a year from now!

Why do I tell you this? Because the model we are following is working superbly. I will tell you of it, and perhaps you can adapt it for your needs, and maybe GG&S will continue to thrive.

I think the coding model you are using is flawed. Correct me if I'm wrong, but I *think* you have several coders, each of which is taking a particular role. One person codes UI, another is doing economy, another is working on the map, etc. This works in a standard office environment, where each person is focused on doing their job for 8 (or more) hours per day, but in a less focused environment, with volunteers and people that come and go, I don't think it works as well.

For the model that MD uses to work, you need four things.

One, which I think you may have (not sure) is a set of coding standards. Everyone *has* to code in the same format, because everyone is going to participate in creating each bit of code. For that to work, you have to agree on how the code will look. MD has it's code standards posted publically on our board.

Two, you need a code master. This is someone that will collect all the bits of code after a subject has been completed, and put it together into the master code base. In the case of MD, that is me. In your case, it needs to be someone who is pretty good with coding and pretty committed to the project, because that person will be heavily worked

Three, you need to post the entire code base somewhere, publically, or privately, however you wish. But the coders *must* have 24/7 access to see the current code. The code master must update this base on a regular basis. Preferably, it should be in sections (for MD, each class has it's own section showing the header and cpp file (and DFM file, if its a UI class)). That way, only those classes that have changed need be updated.

Four, you need to be able to write small and to-the-point routines. The absolute best way to do this is to use an OO methodology. It *can* be done with procedural code, though it's harder. In any event, you cannot be writing fifty page routines. They must be small enough that everyone can follow the routine easily and pretty much at a glance.

So, how does this work? Pick a subject. Make it *specific*. As a for instance, we are right now working on "resource placement on the world map" for MD. We just finished "closed economy resource regeneration" and "open economy resource regeneration". These are very particular routines, and, most importantly, can be coded in under 100 lines.

When a subject is chosen, the entire team focuses on that subject. One person may suggest an algorithm. Another codes it. The algorithm/code is posted on the board. Others suggest changes or enhancements. Changes can be posted as fragments that fit in or replace parts of the original post. This shouldn't take weeks. In my experience, two or three days from start to finish should do it. If it's taking longer, your topic is too complex. When you are done, you have a routine that does some very specific function very well. Once all changes are complete and everyone agrees the routine is done, the code master takes the final version (sometimes having to peice it together from the posts in the entire topic) and plugs it into the code base. There may (probably will) be a bit of extra coding to integrate the new function into the overall game loop. And viola! Subject done, you have a new feature, move on to the next one.

One great thing about this system is that even non-coders can participate. C++ is not all that hard to follow. A coder must do the initial coding, but once thats up, pretty much anybody (even artists! j/k ) can follow what is going on and make minor modifications or suggestions to the code.

Or, even non-team members could pop in from time to time. I might look in on a bit of code and make a suggestion here or there, or Joe Shmo might offer his two cents worth somewhere else.

It seems to me, you guys started off pretty well, make a nice overall design (which is certainly important). But the next step, the coding of all the little details, stopped the project. Perhaps this might help in getting it going again.

Or, you can ignore me and just do what you were doing. Maybe you have other ideas on how to get things moving again.

Ron

[Leland]

Well, our problems aren't as much in "everybody working on their own" as it is "nobody working at all" when it comes to coding. There has never been a proper leader in programming who would've pushed it forward (at least not in the past year or so), though the designers have been a little bit more active. As a result, the threshold for programmers to split has been very low, there has been a certain lack of direction and motivation haunting us. At least MD has one committed coder.

But hopefully we are changing. I've personally taken the resolution to set up CVS repository at sourceforge for our code, and maybe coming up with some standards on how we're going to use it. I'm not so sure if we could use the MD model directly though... there's no guarantee that any of the coders are permanently committed to the project so we'll need to fix things so that(ideally) no one person is critical to the process. You did make some valid points though:

• It's good to have more than one people working on each code snippet. But on the other hand, I don't think we would be motivated to do one thing at a time. It's like saying "you can't do this now, finish your vegetables first!". But anyway, even if we have several things being coded at once, I still think that whatever we do, there should be at least two programmers tackling the same code, and all the rest following the work through the mailing list and such. There may not be much difference between the results of two and three people, but there's a world of difference between working alone and working in pairs.
  
• Code has to be in proper sections. That's almost a prerequisite for doing things in parellel, in an ideal situation the two pairs/groups working on different things wouldn't need to torture themselves with heavy synchronization of their results. Also, the designs of our software architecture we have so far have been made with modularity in mind.
  
• Specific routines. This is very important, because when you dwell into your own code too much it can get very convoluted. Also, each model/algorithm and whatever we do should be, in my opinion, done in a lot of small steps. Object orientation helps a lot in this respect and we're not even dreaming of going procedural (well, maybe some of us are, I can only talk for myself...). Here, we stumble into the lack of design we are suffering from... we have the design doc, but that's definitely not enough. I hope that once we get the basic software skeleton finished, it's going to be easier to develop the models and code them somewhat simultaneously, which should motivate the designers.

It's nice to see that we're not considered hopelessly lost by our fellow alt civ projects. It's not going to be easy to get out of this ditch, but we'll certainly give it our best shot. Maybe you'll see some results in about a month or so, hopefully in form of working demo.

[S. Kroeze]

Don't mind us, we are just trying to die here! (Seriously, it would be nice to get the designers at least moderately interested in this project again, now it seems only the programmers are excited.)

There was discussion about turn order and technology models, the latter seems to still be underway but I think it's safe to say that we've got the turn order covered as well as is appropriate at this point of development.

Since ten people still seem to believe we can still accomplish something -however modest- I think we shouldn't give up yet.
Personally I often do not have the slightest idea what is actually going on, though I read the boards quite frequently. I think it is very important that someone -definitely not me- with clear understanding of the programming process, decides on which aspect we should concentrate. Nor should we mistake discussion for progress.

In the past I hope to have made some useful contributions. I am rather uncertain about it, because to me it has always been obscure who take(s) decisions and what was done with my few intelligent and possibly effective ideas. Yet I have some experience with designing games -though not professionally- because I have spent a lot of time designing Diplomacy variants or similar games and have played many differently structured board games. Basically, designing a computer game is -in my view- NOT an essentially different activity. A computer is just a lot faster than the human brain, especially when calculations have to be made.

Before one can become truly creative, one needs some possibly working game structure (read: a game one can play with one another) -however crude- to which features can be added and/or modified. Unfortunately this basic structure is not there yet.
I think we need some document with our "Rules of play".

Something like: "GGS is played on a map containing 600,000 hexes. The goal of the game is to score Victory Points(VP).
There are several ways to win points. When a player has accumulated 100 VPs the game automatically ends and this player wins the game.

The game is divided into megaturns which represent 100 years. When a megaturn has been played out completely, all participating players can score VPs, if any....."

This is of course just an example. Yet I hope this will force us to design in a more structured way.

[The Joker]

Ehh, haven't some posts been lost here? I was here a few days ago, and there certainly were posts made after Dec 24.

What's happened?

[VetLegion]

hi Joker

yes, you niticed correctly, at least 4 posts dissapeared, two of yours and two of mine.

I ll try to find mine in explorer temp folder, if I find them I ll repost them.

hows things? you still onboard? drop in a chat sometime, havent chatted with you for a while.

[The Joker]

Yeah, I'm still on board. As much as I can be between studying for my exams (as I wrote in the lost post).

I hope to be able to do more good soon, but the next 10 days I will mainly focus on the exams.

Anyway, are the chats still thursday nights?