opensmac

[Domin]

Very interesting! If you're able to get a clone working (especially with working multiplayer, and maybe too optimistically with the simultaneous turn dynamics the original has) that would be pretty amazing. Could you link to the github repository, maybe from the OP?

One thought I'd like to put out there for when you get to AI: Perhaps build the AI as a module which is somewhat exposed to the user, effectively a relatively easily editable script. This would allow users to quickly test tweaks to solve AI stupidity that they encounter, and make AI competitions fun. Have actions that could be taken by AI (including various cheating options) and information that could be gathered by the AI as built in functions, and a basic example AI which does the very basics mostly as an example (probably cheating even more than the official one to stand), then encourage people to tweak and improve, and share their advances.

I feel like having the AI's reasoning be open and adaptable (in a "there is a menu option called Edit AI script" way, not a "Get git set up, download the source code, make changes, reload program" way) would attract people who would be both great for the project and likely to have loved SMAC/X in the past. The game seems like a fun challenge for AI designers, because there's fairly simply structured information and input/output, but enormous complexity and possible advanced strategy in various directions many of which could be improved incrementally by a semi-technical user with good game knowledge if the scripting had good examples (terraforming, unit design, army vs development, research paths, diplomacy, etc). I'd especially love to see different AI's battling each other, and the ability to script your own governors for bases so they make sane choices reducing micromanagement lategame.

@Mart: I think you're thinking of new game design? Cloning a game is quite different, you pretty much just need programmer power for it.

I would certainly like to support pluggable ai modules/scripts. However I think I have not implemented enough of functionality to get good picture of what would the script/ai api look like.
And there is the problem of fog of war to which I don't have solution yet. The idea of script-assisted human gameplay has haunted me for years so I clearly see that there can be overlap between automation done on behalf of human player and ai in narrow sense of fully ai player.

That would help, and perhaps be advisable to think about a fair amount early on to reduce later need for restructuring, but again that's something a programmer (or at least someone with a good understanding of programming) is going to be needed for, unlike a normal game's design phase where you want designers primarily.

Edit: hm, found the github, it looks like you're not including SMAX elements? Is that a later plan (build basic game, add extras), or are you not keen on SMAX? I'd encourage you to aim for the full game, there's a fair amount of added depth to gameplay from the assorted extra features (more native diversity, techs, factions, facilities, SPs, etc), which has lead to most of the users here treating SMAX as standard.

At the time you looked at the source there was nothing specifically smax because I did not get there yet. Personally I don't have a strong opinion on smac/smax issue but have chosen to include smax stuff from the start. As there will be no ai for some time, I suppose the natural things to leave out at first are these which are so far in the tech tree that they rarely appear in multiplayer games. And alien factions - they are mostly useless, especially in this case.

[ete]

I would certainly like to support pluggable ai modules/scripts. However I think I have not implemented enough of functionality to get good picture of what would the script/ai api look like.
And there is the problem of fog of war to which I don't have solution yet. The idea of script-assisted human gameplay has haunted me for years so I clearly see that there can be overlap between automation done on behalf of human player and ai in narrow sense of fully ai player.

Cool that you're interested in the script assisted human play thing, and yep, makes sense to not start on that side until you've got a good idea of the foundation of the rest of the game, though worth keeping it in mind while building it up.

At the time you looked at the source there was nothing specifically smax because I did not get there yet. Personally I don't have a strong opinion on smac/smax issue but have chosen to include smax stuff from the start. As there will be no ai for some time, I suppose the natural things to leave out at first are these which are so far in the tech tree that they rarely appear in multiplayer games. And alien factions - they are mostly useless, especially in this case.

Okay, I imagine people who would want pure SMAC could easily enough remove most features with alphax if they wanted. And yea, leaving unused lategame stuff as low priority makes sense.

[Mart]

A thought, when making a game clone - can this be made without faction number limit? I mean, you could start a game having 12 or 14 factions in it.

[Kilkakon]

That would be pretty huge :O

[Domin]

A thought, when making a game clone - can this be made without faction number limit? I mean, you could start a game having 12 or 14 factions in it.

There is no technical problem with that.

[ete]

That'd be a major selling point for the clone, especially if it (as it looked from my glance at the code) you're loading from the original format text files and can load a bunch of custom factions alongside official to spice up a game.

[Ford_Prefect]

Are you still working on this project? 

[Domin]

I haven't done much lately as my life got in the way.
I do intend to continue the project.

I'm happy to see quite some progress on cvr files, I thought they would be a stumbling block.

[Ford_Prefect]

Haha yeah.  I was thinking of making a SMAC clone myself, then I found your project.    I realized no one had figured out how to read or edit CVR files... which meant that such a project would have to recreate the units from scratch.   I suck at modeling and so I thought it would be worth a shot at trying to figure out how the CVR files worked. 

I'm honestly wondering if it would of taken less time to recreate the assets.    But its been fun figuring it out.  (except for the normals)

[Domin]

I'd just like to let you know that I resumed work on the project.

[Impaler]

As long as your back now looks like a good time to mention that I grabbed a clone of your work so far but found that their was some kind of Python incompatibility with your version of pickle and possibly the version of Python your using too (I'm using 3.3).

If you can get this cleared up I may be able to contribute a bit, I've worked with Python and pygames before.

[Domin]

I'm using python 2.7 and that's the reason behind the problem with pickle. I've checked it - in python 2 you manually choose between pickle and cPickle (the fast C implementation) whereas in python 3 you just use pickle and it automatically uses fast version if possible falling back to python implementation. But that's just that one issue. Frankly I've never used python 3 and I'm not aware of further differences apart from the 'print' statement/function thing.

Edit: I've checked and pygame supports python 3 only in development version - there are no standard packages available across platforms.

[Ford_Prefect]

The cvr work I did was a partial waste.  Can't export the game models.  I did make it possible to "paint" the models with different colors. 

I think that it would be quicker to create the units from scratch.

If you are interested, here are the links to stuff on the caviar format
http://alphacentauri2.info/index.php?topic=7866 Forum thread for the editor.
https://github.com/uaEquals42/CVR-Colorizer  Code for messing with CVR files.  Written in python 3
http://alphacentauri2.info/wiki/Caviar  Everything I've figured out about the caviar format.

[Domin]

Are there mayor problems that prevent rendering models from crv files? I know about the normals which obviously makes rendered graphics flat and ugly.

What about relative positions of parts of multipart models? I've noticed that Colorizer displays them separately.

[Yitzi]

Haha yeah.  I was thinking of making a SMAC clone myself, then I found your project.    I realized no one had figured out how to read or edit CVR files... which meant that such a project would have to recreate the units from scratch.   I suck at modeling and so I thought it would be worth a shot at trying to figure out how the CVR files worked. 

I'm honestly wondering if it would of taken less time to recreate the assets.    But its been fun figuring it out.  (except for the normals)

I'd advise against figuring out how the existing files work when making a SMAC clone; most of the arguments for why it's ok to reverse-engineer SMAC for these sorts of projects relies on the fact that it's being used to enhance SMAC rather than compete with it; if you use it for a SMAC clone, I think you're on much more problematic legal ground.  (This is why I can't provide much help with opensmac, though if it gets going I'd be happy to help include an equivalent to my patch, and any other changes that aren't in the original.)