A simple improvement

[Loren]

PDF said:
I can't see how it can unbalance anything if it works by the rule (ie you need commander to carry the troops).
It'll just save mm so would be great, as would be waypoints for multi-turn moves, "armies" (multi-leader) allowing to better distinguish the groups by mid-late game, and several other things ...

I would permit it to happen without commanders. Soldiers don't really need commanders to be told "Go to city X".

[Loren]

Gandalf Parker said:
It would add quite a layer to the turn processing. Every province would have to be checked (1500 provinces) and then processed if need be.

It would add almost nothing to the turn processing. (Note: I've been making my living programming for half my life now.)

1) Checking even 1500 provinces would be a triviality on modern processors.

2) There's no reason to check them in the first place. The check would be made when new units are created or when units executed new-unit orders.

It does, however, affect the storage format. You need to add a flag to every province saying where to move to and you need to add a flag to units saying that they are executing new-unit orders.

[Gandalf Parker]

The game doesnt do things that way. It doesnt keep track of things. You can see in the logs that it checks each province for many things that would fall into the category of "if" checks. And those checks do take measureable time. Adding another would be felt.

Im not saying its an idea killer. Just that it would be part of the "con" side of it.

[PhilD]

Gandalf Parker said:
The game doesnt do things that way. It doesnt keep track of things. You can see in the logs that it checks each province for many things that would fall into the category of "if" checks. And those checks do take measureable time. Adding another would be felt.

No way. If the game programming is done half correctly, this would take no more time than processing an equal number of orders - exactly as much as it would take if each nation's player had issued the order, province by province, except they wouldn't have to bother doing it.

Giving more options, ways for players to automate their order-giving (like letting them issue multiple turn movement orders, and so on), would not make turn processing significantly more complex. It's more work to code in, sure, but that's another matter.

[Gandalf Parker]

Like I said, it sounds like a good idea.
But looking at the logs you can see that it processes thru all of the provinces multiple times for many different things. I suspect its the result of adding such routines to a game that wasnt originally meant to take things like that into account.

Its still a good idea but dont be surprised if it processes thru all the provinces to do it.

[Arralen]

It's late, I'm tired and have little notion to go into details - with the current game engine, there's no way to tell troops to "go somewhere" on their own.

Routing is triggered by battles (and during battles - check the turn sequence order) and is not available during unit freespawn or building.
Furthermore, routing is random - routing units spread into neighbouring provinces. If that was easily changable, we might already have "fire and retreat" orders for light cav which have it move back into the starting province without scattering.

We don't have waypoints for a similar cause - no way to store the destination and steps in between over several turns.

Keep in mind - the game engine is basically the one from DomPPP - and when they started to make that game, no-one thought about complex movement army orders and 1500 provinces ...

And if someone wonders if this will change in Dom4, or if there will be a Dom4 ... the devs would have to throw away nearly all the code done in the last years and start from scratch ... therefore chances are very slim.

[Gandalf Parker]

t's late, I'm tired and have little notion to go into details

But you do it very well.

[Agrajag]

I suspect its possible to code in multiple turn orders without throwing away all existing code, and probably without much impact on turn resolution time, though it would take some time to code, and UI work :O
Just make it so you can queue orders onscreen, which on the .2h file are represented as more order under a different sector, marked "should be done in X turns", and at the end of the turn resolution process, just make the block for the next turn's orders into the next turn's .2h file (and the blocks for "in x turns" into "in x-1 turns"), then just like you load a game now and get a choice wether to continue the save or erase all orders, you'll have the .2h file for this turn's orders.
My idea obviously has several drawbacks:
1) It assumes that invalid orders are disregarded, rather then cause an error.
2) Would make MP more cumbersome, since it would require lugging around both .trn and .2h files for each turn.
3) Would require quite some coding work, especially dreaded UI work.

[Loren]

My approach doesn't even need an ability to queue orders. It needs two data fields:

Units need a "new" flag that means they execute orders on the province. Provinces need a destination.

At the start of move order resolution you look at each unit. If the flag is set and there's a valid target (it won't move if the target isn't in friendly hands) you move the unit, otherwise you clear the flag.

[Loops]

Configuring auto-marching orders for troops sounds nice, but it isn't really thematically consistent with the rest of Dominions where commanders are required to really get troops to do anything.

What about another approach to issue - one more focused on the sorts of things we already see in the game. Suppose there were a new order for certain types of commanders ("Gather Minions") that would act like a spell that teleports all existing uncommanded troops from adjacent provinces to him. This would dramatically reduce the scale of the management issue without requiring anything really new. You could even get fancy and limit the effect to his leadership ability.

There are already commander abilities that generate troops and spells that can target uncommanded troops (not specifically maybe but hostile rituals can certainly hit them), so it wouldn't seem like there were any obvious hurdles. Although, I'm sure a lot of little complications would pop up, and this mechanic (or anything like it really) would be extremely powerful so it'd take some careful thought.