@LaFayette said in TripleA Rules Sets: What would be ideal is to have a table describing different behaviors by rule set. We are actually on our way there. We have example maps that follow each rule set, all we need to do is to start teasing apart the behaviors that are different and then answer for each map, "does this follow that behavior, yes or no?" If we find that some rule sets are only very slightly different, simpler is better. The 'v3' moniker will be difficult to drop, if we can keep it simple and say "old edition", "revised edition" and "new edition", it could be easier to adopt. At this point though, without knowing what the rule sets actually are, it seems to be a first priority to know which rules are in each rule set before we do much of anything else.

@Cernel As we unzip a file, we will be doing a lot of checks (are all the required folders there? is there a centers.txt, etc.?) and as part of that process, we could enforce a casing convention. I have a tool that fixes capitalization. When I got image assets from @GenerationKILL sometimes the casing was a bit off, so I would just run this tool and it would conform every unit to a convention like this: Unit-Name.png regardless of the OS and regardless of the original naming. We could implement a similar normalization process at the map parsing level (though it might be hard to get map makers on the same page about which casing convention should be used). https://github.com/jdimeo/triplea-map-tools/blob/master/src/main/java/jdimeo/triplea/util/FixFileNames.java

So, would you be ok with removing some locks if it can be shown that the data is actually read only? I'm currently looking at ResourceList which is what GameData#getResourceList returns. The class has only one setter and it is only called during the game parse process and a test method. The data that ResourceList stores is a map of Strings and Resources. And Resource is a value object. ResourceList doesn't appear to need any sort of read lock since I don't see any way for it to change. If I see locks around GameData#getResourceList, can I just remove them? Would you like me to tag Resource and ResourceList somehow (such as @Value) to indicate that they are read-only objects?

@LaFayette Dont forget to add triplea.log to the list of files to uninstall.

That said, is there a particular reason (other than a chunk of code not really necessary to functionality) that save game files don't work that way? Java serialization was once popular, it was an initial design decision to use it, it was somewhat easy. We would actually like for save games to work more like you have described. It's a very large project, that is all, and we have a large backlog of small and large projects that we are working on and collectively there is not that much development labor available. Re: performance requirements These are the wrong things to focus on. First focus on functionality, what are you going to built? What value will it provide, how will it be used? You need a good product for anything else to matter. Then I'd encourage to stop worrying about specs. Simple and well done architectures are fast and perform well. Simple code is usually fast code. Once you have a reasonable design in place and functioning, then you can start analyzing where the performance bottlenecks are and where to go to optimize. It's a ways to get to that point, and once you are there the process is to measure, change, repeat. Humans have generally very bad intuitions for where performance problems will arise. It's also easy to try and fix all performance problems as optimization problems when they are really design problems. One can look at a block of code and wonder "how can I make this faster", and sometimes the right answer is, "how can I avoid needing to execute this in the first place?" That is why keeping things simple and working is very important, it lets you reason about what your code is doing. For example TripleA does a lot of calculations twice, why? Because there are 15 layers of call indirection separating the two which creates several thousands of lines of code in between the duplication. Bottom line OUr time is short in just about every way imaginable. I'm pretty sure and somewhat regrettfully going to say, if you want something that interoperates with TripleA, it's potentially one of and likely all of: untractable and/or a truly huge project We have wanted to for example to allow TripleA to run on web browsers, or phones. That project is probably 3-8 years away from being completed if it were started today.

Its a good idea, I like it.

I think you may have to look at which maps have the combination of properties set to see how it is used. I can imagine a few different variations of how such behavior would be used. Generally, if the intent were to be able to shoot down a multi-hit-point air unit, it would make more sense to me for the AA to do more than 1 damage.

Sounds right, though I think this is a simpler way to think about it: before a unit rolls, it looks to see if there is an available support, if so, it consumes that support.

@alkexr , thanks for those forum posts. It does look like this warning is no longer valid.

@LaFayette said in I would like to start helping out with tripleadevelopment but don't know where to start: Are you familiar with doing feature branches in git? Were you able to compile and run the game from scratch? No to both of those. I guess I need help.

@RoiEX I have read your specs, Issue seems to be closed at least for now. How do we proceed from that? Any Idea? I have already my questions about your navigations and other annotations. Regards @RoiEX .

There are three ways to merge branches: merge fast forward merge rebase git merge <remote> <branch> and git pull <remote>/<branch> AFAIK are the same thing and are both merges. git pull --rebase <remote> <branch> is a proper rebase. This is the one you should prefer to avoid having 'merge commits'. A fast forward merge is really similar to a rebase, and you do that with a -ff flag to merge, ie: git merge -ff <remote> <branch> There is also a proper 'rebase' that can be done. AFAIK these two are the same: git pull --rebase <remote> <branch> and git rebase <remote> <branch> Unfortunately it is fundamentally a bit confusing as it depends on exactly how the rebase is done and with which flags. This is a place where a GUI can get in the way.

@Trevan I am glad you know your way around! I never would have found that.

@TheDog Thanks. Quite handy.

Awh shucks (looks down & shuffles feet)

@redrum so here's stab #3 AI got moscow again in stab #2. this is v3 btw figuring that the most advanced version of the game that AI can play effectively. playing as allies with no bid maybe i should started as axis but i wanted to see the AI's initiative and figured giving some kind of handicap to AI was probably warranted. noticed that the AI is definitely stronger than before but still seems to have some weakness in the area of protecting transports. i know we've discussed it before but seems like there must be some way to prevent AI from spending all it can on facility repairs as well makes the AI very susceptible to strat bombing. autosaveBeforeEndTurn.tsvg

@Trevan Yeah, this is a very old problem. V1.III is not something I can be sure, but I'm almost sure you should be able to submerge regardless whether or not you are able to retreat. As far as "Submersible Subs" goes, my suggestion was to keep the name as an archeproperty and split it into a series of new properties.

@Trevan I suggest you not to make changes if you don't know what's the matter about it, especially in the moment that the change is not going to fix any known problems. The auto-elimination steps should rather be generally revised, maybe from scratch, so to solve the known problems and creating a system actually enforcing the correct rules (that will need to be interpreted, to some extent). Thinking hard about this, the only situation of only infrastructures remaining needing auto-elimination of non-infrastructure units I can think of is the case of defending units refusing to partake in air battles.