TripleA development

[djabwana]

But yes- contributing to TripleA has been hard for me, because of the code base as well as a difficult PR review process, either due to a focus on incredible detail (that's what happens when a bunch of military strategists maintain an open source project I suppose ;-)) or because of inactivity of PR approvers. I certainly understand both reasons, but it has been discouraging me from contributing more if I know it will be a huge time investment to even get a change merged in to the next release.

[MagicStyck]

@djabwana said in TripleA development:

But yes- contributing to TripleA has been hard for me, because of the code base as well as a difficult PR review process, either due to a focus on incredible detail (that's what happens when a bunch of military strategists maintain an open source project I suppose ;-)) or because of inactivity of PR approvers. I certainly understand both reasons, but it has been discouraging me from contributing more if I know it will be a huge time investment to even get a change merged in to the next release.

This. All of this is exactly why I have done limited contributions so far.

[MagicStyck]

@djabwana said in TripleA development:

The biggest issue I see is the game file format itself. XML is fine, but we should use an XSD not a DTD and have the structure of the file reveal more of the available options instead of relying so much on generic properties with weird operators (like splitting a string on colon) that require extensive documentation.

The biggest benefit too is that map makers can leverage xi:include to include fragments of common elements (like units) into mods/skins/whatever.

Then we can also code generate the map saving/loading code with JAXB, removing a ton of boilerplate and and making games backwards compatible instead of using Java serialization.

The use of JAXB or JSON for serialization would help to fix a lot of the game save issues. The serialization that is currently being done to OutputStream makes sense if you don't want an extra dependency or you are looking for an easy way to keep users from editing saved games.

However, I don't see that trade off as worth it given that JAXB or JSON are far more portable, easier to troubleshoot, less prone to "mysterious" corruption and can just as easily be secured using encryption.

[RaiNova]

@magicstyck @djabwana If we work on accelerating the PR process, how much time would you then volunteer to TripleA improvements? What would be required to make these improvements worth the time for you? Would you be satifsfied with purely technical improvements like fixing the serialization and cleaning up the code base?

Would you expect that we can marshal the required dev capacity by merely improving the dev process?

Would it be worthwhile to discuss other ways to make development more attractive? If so: Would it suffice to adress the developers currently active in the TripleA community or should we try and recruit developers who are currently not part of the TripleA community?

[RaiNova]

@panther said in TripleA development:

IMHO the "conduct combat phase" is [...] about resolving the conflict represented by the battle-board. To me this is somehow the highlight of every turn.

When you deal with this conflict: Do you rather think of it as a battle or of a struggle between opposing military formations during a 3-6 months period (i.e. the duration of a turn)?

Some rules lean towards the former - e.g. the submerge capability or the fact that you cannot partially retreat (in general).

Improvements in the display of the conflict may depend on this preference.

[RaiNova]

@panther said in TripleA development:

What would "zooming in" show then?

It may show the bombarding ships and the related transports on the edge of the display.

The map could be turned in a way that the invader is coming from the left.

Infrastructure units of the territory could be displayed behind the defender, walls in front of them.

If artists are interested, it could actually show a bit of landscape.

These are just the four of my ideas I can most easily sketch out. Once we prodvide the opportunity, artists may come up with better and more.

I am also thinking about a more modern, drag & drop UI.

[RaiNova]

@panther said in TripleA development:

IMHO the "conduct combat phase" is not about territory sightseeing . [...] Napoleon [...] is not an engine topic but a quick fix within the map repository [...]

Would you be keen on better music/sound effects?

No. That's just distracting bling-bling in my opinion. As well as eventual graphic "enhancements".

This gives the impression you are playing TripleA mostly for the game logic and hardly for the atmosphere of the time, the setting, the personalities. Does this describe your approach towards the game?

@ all: What is important in TripleA for you?
(For me personally it's sometimes the logic, and sometimes it's the "story". For example, when I played "Total War Shogun II", I really liked it when the general gave a speech to his troops, expressing his sorrow for the enemy soldiers.)

If we start anew: What shall be kept? What shall evolve? And by how much?

[Panther]

@rainova said in TripleA development:

@panther said in TripleA development:

IMHO the "conduct combat phase" is [...] about resolving the conflict represented by the battle-board. To me this is somehow the highlight of every turn.

When you deal with this conflict: Do you rather think of it as a battle or of a struggle between opposing military formations during a 3-6 months period (i.e. the duration of a turn)?

Some rules lean towards the former - e.g. the submerge capability or the fact that you cannot partially retreat (in general).

Improvements in the display of the conflict may depend on this preference.

It is simply one battle of - sometimes more sometimes less - battles.
Resolving the batttles by the outcome of the dice rolls is most important.

[Panther]

@rainova said in TripleA development:

@panther said in TripleA development:

What would "zooming in" show then?

It may show the bombarding ships and the related transports on the edge of the display.

The map could be turned in a way that the invader is coming from the left.

Infrastructure units of the territory could be displayed behind the defender, walls in front of them.

If artists are interested, it could actually show a bit of landscape.

These are just the four of my ideas I can most easily sketch out. Once we prodvide the opportunity, artists may come up with better and more.

I am also thinking about a more modern, drag & drop UI.

What for? If I wanted animated pictures to create atmophere I would watch films or buy other computer war games.

Imagine people playing in leagues or ladders - sometimes playing five or more games at a time. They usually don't want to see those animations. It is distracting and - depending on the animation - a waste of time. They are busy developping strategies, see how they work out and continue from there.

[Panther]

@rainova said in TripleA development:

This gives the impression you are playing TripleA mostly for the game logic and hardly for the atmosphere of the time, the setting, the personalities. Does this describe your approach towards the game?

Correct. There are other games having this atmospheric approach.
To me it is a strength of TripleA to have a priority on the game mechanics - while allowing for a certain amount of customization including some atmosphere.

Other approaches - such as Beamdog's 1942 Online - are maybe more "atmospheric". But you can read reports of people who are enerved by the bling-bling shipped with it.

Really - for atmosphere and animations and whatever there are countless alternatives on the computer games market.

I, too, love scenarios like the GoT or WoW approaches, also historical scenarios. It is great how the map makers succeed to bring atmoshere to the games with the tools TripleA gives.
I am not against optimizing or careful enhancement.

But please let us not bloat up and by this reduce the straightforwardness of our TripleA software.

[zlefin]

I just hope it doesn't take too many more years for the unit transfer bug in bots to be fixed, so global and TWW can work right without editing.
sadly there's nothing else I can do on it.

In general I like the plan of just doing basic/essential fixes for now, while having a long term plan for a 3.0 that reworks the code base from the utter mess it is now.

UI-wise; I'd say triplea is fine without bells and whistles; but having optional add-ons for more atmosphere should be fine, as long as it's well compartmentalized. Much like Dominions, the user-base is a niche one that's fine without fancy graphics because of the good gameplay.

[wirkey]

@zlefin said in TripleA development:

I just hope it doesn't take too many more years for the unit transfer bug in bots to be fixed, so global and TWW can work right without editing.
sadly there's nothing else I can do on it.

In general I like the plan of just doing basic/essential fixes for now, while having a long term plan for a 3.0 that reworks the code base from the utter mess it is now.

UI-wise; I'd say triplea is fine without bells and whistles; but having optional add-ons for more atmosphere should be fine, as long as it's well compartmentalized. Much like Dominions, the user-base is a niche one that's fine without fancy graphics because of the good gameplay.

Don't ask me when and why but it seems like the unit transfer bug in TWW was solved. I recently played a game in the lobby within a bot and encountered no problem.

[zlefin]

@wirkey

I just ran a test in bot 204, and the bug is still present, in both Global and TWW. Do you remember which bot you used?

This fits my long-standing theory, in that the problem is that most of the bots have not updated to v 2.5.22294; and the solution is to get the bots to properly update the version they're using, which is something I do not have access to.

[wirkey]

@zlefin I don't really remember, but I usually try to play the 900s

[frigoref]

@cernel said in TripleA development:

Generally speaking, I personally believe that everything should be always visualized, nothing left to the memory of the player. This said, there are far more important missing matters, like the fact that TripleA fails to show retreat ways and which units can retreat until after you start the specific battle (and, in the current 2.5 stable, it doesn't even do that during the battle!), or the fact that the remaining movement of air units is not displayed, to name just a couple of such problems.

@Cernel Do we have feature requests for them on github? If so, could you please link them?

[Cernel]

@frigoref Not that I'm aware of.

Most of the times I open an issue nothing comes out of it, so I tend to open issues only for matters that I consider very important.

What I mentioned are some of the most important matters amongst the least important matters, as the game failing to show something is not much of an issue for experienced players.

For example, if I would open an issue for anything right now, I would probably ask to support the ability (which should exist in every basic game) to withdraw submarines individually (as TripleA incorrectly limits the decision to all or nothing).

[Cernel]

To be clear, when I said "there are far more important missing matters", I meant "there are far more important ones of such kind of missing features" because I was talking of "merely visual aids for the players", specifically the "just conquered" markers of 270BC Wars mentioned by @RaiNova (which just show what are the territories conquered on the current turn: a very scarcely important matter, in my opinion).

[frigoref]

@cernel
I think adding them as a new topic in Forum - Feature Requests & Ideas is the best so we do not lose the idea, the community can comment on it and we can sort the feature requests by votes.
This process we should try to keep and follow even though (or maybe especially because) we are not having a lot of development capacity right now.

[Cernel]

@frigoref Unless I'm missing something, I don't think currently we are even clear if and how (and where: here or GitHub) anyone is tracking feature requests at all, really.

This is how feature requests used to be tracked:
https://forums.triplea-game.org/topic/182/guidelines-and-feature-request-list

[MagicStyck]

@rainova said in TripleA development:

@magicstyck @djabwana If we work on accelerating the PR process, how much time would you then volunteer to TripleA improvements?

Like most devs, my time fluctuates based on work and family priorities. I could reasonable volunteer time of 3 hours a week or more.

What would be required to make these improvements worth the time for you?

PR reviews in a timely manner would go a long way. PRs that sit for long periods of time and never merge are annoying. PRs that are for documentation changes should be reviewed quicker than technical PRs.

Also not holding up the merge of PRs which are purely technical with non-technical objections. For instance: rehashing over and over the wording of comments or objections that it might affect something completely unrelated to the PR.

Would you be satifsfied with purely technical improvements like fixing the serialization and cleaning up the code base?

Absolutely. The goal, for me, is improvement of the software. Cleanup, bug fixes, closing gaps, is all part of that.

Would you expect that we can marshal the required dev capacity by merely improving the dev process?

Yes. In my experience, when the process is easy to adopt it is easier to inspire more people to help out.

Would it be worthwhile to discuss other ways to make development more attractive? If so: Would it suffice to adress the developers currently active in the TripleA community or should we try and recruit developers who are currently not part of the TripleA community?

It's always beneficial to discuss ways to attract others to help out, whether they are part of the TripleA community or not. It sounds like the current developers are burnt out and need help to move things forward. If someone is offering to help, whether they are in the community or not, let them help.

Making development more attractive comes as part of the cleanup. It has to be easy for a developer to setup an environment quickly and able to debug what is happening. If they have to jump through a thousand hoops to get started, your going to lose them.