Pruning image and sound assets distributed with the game engine
-
By the way, of course, not necessarily just a repository. Adding a "cavalry" unit for all players means that if a mapmaker wants to make a map like Revised, but with also a "cavalry" unit, it can do it without having to provide any units folder in the map itself. That would just mean supporting future games having "cavalry" units, like now you are supporting old and new games having "artillery" units (not requiring the mapmakers to make or grab artwork for such needs). Of course, I'm just talking theorically here, as clearly we don't have anybody that wants to provide such images, nor a developer that wants to accept any of them, looks like. No idea if you are, instead, thinking about a process in which, if a mapmaker would want to make a "Revised with cavalry" map, he could provide the units and ask for them being added to the assets right before adding the otherwise working map using them to download list; that would work too, but it would be a very conservative process, even if documented anywhere (which I don't think it currently is).
-
I think you guys are having this discussion in the wrong thread
-
@ff03k64 You're probably right; I'll move this and the latest 9 posts over to the thread you linked:
https://forums.triplea-game.org/topic/1536/pruning-image-and-sound-assets-distributed-with-the-game-engine/18
from:
https://forums.triplea-game.org/topic/1534/general-map-making-questions/47?page=3 -
@Cernel i didn't care, just thought it was funny
-
TripleA assets should provide a high coverage of any possible needs for historical WW2 games
I do not necessarily disagree but suspect you're conflating the perspective of map maker with that of game player.
(1) There are other ways to provide example assets without distributing it with the engine executable.
(2) The size/cost of those assets is not insignificant and is payed by every player whether map maker or not. There are lots of benefits to a quick download and install. 95%+ of players are not map makers.If you are not doing that, I don't think that maps' assets grabbing would be a good substitute of getting stuff from the surely free program's assets, if that is what we are saying here.
I'm not an expert in copyright. I would not necessarily call it "asset grabbing", but okay.. I'd point out that is the case for most maps, it seems the horse is out of the barn already. Given TripleA is GPL3.0, the maps and contained works, by being made available might fall under that license as well.
I guess that's your opinion.
You seem to be of the opinion that including assets has no costs. It does. I've measured the time to install and looked at file download sizes and it's correlated to distributable size. There are many components involved and the assets are just one, but a significant one. Perhaps you forget when TripleA was nearly 600MB, it was a long road to get it under 50MB and with Java bundling again we're somewhre ballpark 60MB.
I strongly believe that a "cavalry" unit should be added to all subfolders in assets,".
That you're suggesting this for the game engine is your opinion that you'd like for all map maker assets to be made part of the game engine. Why do we have a forum thread and not see those images landing in the game engine some years ago? In part the assets that were included were a dumping ground as people did not find a better place or consider the cost. I'd still like to hear where you draw the line of what to include vs not, lest we have a 7.2GB download as your bar seems to be anything that would be of any use at all.
TripleA would be definitively stuck supporting the needs of the past only, with no possibility whatsoever to redefine it for the future,
TripleA is stuck supporting the needs of the past, perhaps 70%~90% of development work is dedicated to this overhead and tax. It's one reason why a lot of contributions in the last few years has been to drop half-done and never used features to help alleviate that ongoing tax. Keeping maps working is something that we have to do and it adds a lot of extra effort.
with no possibility whatsoever to redefine it for the future,
This is hyperbole, it's possible to redefine, but not necessarily easy. Bundling with the game engine is a step away from that goal as you then require releases and can't readily change items. You also have to do mass analysis of all maps to identify 'blast radius' of changes. Given we have so many maps, it takes very non-trivial scripting abilities that I think only 2 people in the entire project possess (doing that analysis by hand is possible too, but very time consuming as there are 100+ maps).
, then now we should see in assets only the players and the units needed for WW2 Classic, so no artillery, no Italians, etc.,
This is getting to be non-pragmatic and ignores that the bar is to avoid repeated downloads. Some images are included in 100+ maps, rather than have all those maps be 25% or more large in size and duplicating images, they are consolidated so they are downloaded just once with the game install. If we were to ever update unit images, it would be more feasible. Updating identical unit images over that many maps is a non-starter, nobody has the weeks of free-time to do that. We do need to have focus on delivering the right features to TripleA, it is a shrinking community and spending time in the wrong places will not help.
Of course, that also means that TripleA would be definitively stuck supporting the needs of the past only, with no possibility whatsoever to redefine it for the future, cause if an image is not currently provided, then of course no maps are currently using it. If such a point of view would have been actually the official one for this program, then now we should see in assets only the players and the units needed for WW2 Classic, so no artillery, no Italians, etc., cause those were obviously added to assets at some point, then used by WW2 Revised, Pact of Steel, etc., instead of having to add such assets in the maps themselves. Notice also that in case of Pact of Steel (the Chinese assets are for that, not for v3), we are not even talking of "clones", but a custom map, like whatever custom maps may be made now on.
-
@LaFayette said in Pruning image and sound assets distributed with the game engine:
I'd still like to hear where you draw the line of what to include vs not
Obviously, that would be up to the developers to decide. You can decide if the "cavalry" or the "train" deserves to be there just like you decide if the "Canadians" deserve the same. In the moment something new gets added, that is just a bet on the future, more or less like adding a new feature. What I said is that following the concept to "provide a high coverage of any possible needs for historical WW2 games", especially with respect to be at least able to use in sensible ways all that the program currently offers (mostly xml) would be a better basis than looking at what is currently being used, as that is just mostly the consequence of what has been available so far and what the mapmakers happened to decide, hardly with coherence amongst maps, even for the basic WW2, as I detailed.
-
@Cernel said in Pruning image and sound assets distributed with the game engine:
Of course, I'm just talking theorically here, as clearly we don't have anybody that wants to provide such images, nor a developer that wants to accept any of them, looks like.
The idea of adding random artwork for the sake of future benefit to the game engine, to actual detriment (adds to maintenance cost), seems to be a bad idea. Particularly when there are other, easier to find and more appropriate places to place such artwork. It's not a matter of 'accepting' it or not, it's about adding assets to the right location and making them available properly.
If just a couple maps use images, let's say that the map is new and maybe will not be played by more than a dozen players, why should the thousand plus TripleA players all have to download that image when they could just download it with the map?
Given bundling with the engine means images are now dependent on more than one map and dependent on a game engine release, it does not seem to be a flexible or very desirable thing to do, particularly if it's just for the sake to distribute those images to potential map makers. It's much better to provide them in a consolidated context with documentation.
No idea if you are, instead, thinking about a process in which, if a mapmaker would want to make a "Revised with cavalry" map, he could provide the units and ask for them being added to the assets right before adding the otherwise working map using them to download list; that would work too, but it would be a very conservative process, even if documented anywhere (which I don't think it currently is).
I'm not thinking about as a process, no. 'Dutch' might be added to the game engine assets, but nobody has volunteered to do the analysis of whether all dutch players actually have the same images. Given the status-quo is to leave it as is, it'll probably be left as is. Particularly now that the maps download quickly, the need to bundle with the game engine is far, far less than previous when map downloads were very slow and took a very long time. In other words, because maps download faster, it's fine for them to be a bit 'fatter', particularly if it means the game engine is slimmer (which provides a benefit to all).
-
@Cernel said in Pruning image and sound assets distributed with the game engine:
Obviously, that would be up to the developers to decide. You can decide if the "cavalry" or the "train" deserves to be there just like you decide if the "Canadians" deserve the same. In the moment something new gets added, that is just a bet on the future, more or less like adding a new feature. What I said is that following the concept to "provide a high coverage of any possible needs for historical WW2 games", especially with respect to be at least able to use in sensible ways all that the program currently offers (mostly xml) would be a better basis than looking at what is currently being used, as that is just mostly the consequence of what has been available so far and what the mapmakers happened to decide, hardly with coherence amongst maps, even for the basic WW2, as I detailed.
You're ignoring there are other ways to distribute images, they 100% do not have to be downloaded with the game engine to distribute them.
It's also not a matter of 'deserve', it's a probabilistic assessment based on how many times we expect an image to be downloaded. The original roots came from SVN days when downloading maps was very slow, adding to the main download a single set meant maps using those images could be downloaded much more quickly. It was significant as waiting many minutes for map download, and repeatedly was a deal breaker. Over the 10 WWII maps, if they all each took 5 minutes to download, which at one point was the case before we migrated to a faster server, and before it was possible to download in parallel, that meant a player had to in serial download maps over the course of an hour to have the core WWII. That is the original reason why much of any assets were bundled at all, to avoid that scenario.
Now that is no longer the case, the reasoning to bundle assets with the engine is far weaker. In many ways it's simply better to include what is needed with a map and not create dependencies on other maps nor the game engine.
-
@LaFayette I'm not ignoring anything. It's obvious enough that assets are not required at all for any new maps that I don't believe me or you need to state it. At this point I wonder if the assets are just legacy, an nothing more, for you.
-
@Cernel said in Pruning image and sound assets distributed with the game engine:
At this point I wonder if the assets are just legacy, an nothing more, for you.
That sounds like putting words into my mouth. I'm not sure exactly what you mean by "legacy and nothing more".
-
@LaFayette That means that the assets are there just to support the existent maps.
-
@Cernel The reason they were there originally is due to download time.
See:
The original roots came from SVN days when downloading maps was very slow, adding to the main download a single set meant maps using those images could be downloaded much more quickly. It was significant as waiting many minutes for map download, and repeatedly was a deal breaker.
It does not seem worthwhile or particularly beneficial to reverse that.
I'd agree it's unlikely for a new map to contain assets that are going to become so widespread used that it meets that bar.
Do you consider it a good thing for assets to be downloaded with the game engine? As stated, beyond download and install time, there are reasons to not do that. Particularly as well for lack of ownership. Maintainers are not going to want to modify existing artwork, a random artist is not going to want to make a change that is probably going to impact many existing maps. And finally:
(1) There are other ways to provide example assets without distributing it with the engine executable.
-
@Cernel I feel like you're confusing whether to bundle or not is a value judgment of artwork or 'historical significance'. The reasoning to bundle or not hopefully has been made clear and it's not a value judgement of the artwork.
I hope the repository of sample assets does grow and will eventually include far more than it does today: https://github.com/triplea-game/map-maker-assets
As discussed late in this thread: https://forums.triplea-game.org/topic/918/map-maker-resources/ and for the reasons stated there, for now adding images to that repository of sample assets is the best idea we have so far for how to host the files and make them available.
-
@LaFayette said in Pruning image and sound assets distributed with the game engine:
Do you consider it a good thing for assets to be downloaded with the game engine? As stated, beyond download and install time, there are reasons to not do that. Particularly as well for lack of ownership. Maintainers are not going to want to modify existing artwork, a random artist is not going to want to make a change that is probably going to impact many existing maps. And finally:
(1) There are other ways to provide example assets without distributing it with the engine executable.
I think I've already answered this question, or at least at the moment I've nothing to say that I've not already said. As far as evaluating the costs goes, that is up to the developers, as it would be up to the developers, say, accepting adding a "train" image to all asset's players. Obviously, the only benefit would be in case, at any point in the future, there is a mapmaker that wants to make a map that features only players and units currently disposable in assets, plus also a "train" unit, and doesn't have to add a units folder to its own map because of that (as I said, it would be a bet on the future to guess what units may be important enough to be part of the default assets; obviously "infantry" is one of them, maybe "train" is not). The same thing if the assets would currently receive any kind of territory effects.
-
@Cernel Okay. Perhaps the perspective difference is whether bundled assets should be making a bet on what will become heavily used. Notably such assets can be simply included here: https://github.com/triplea-game/map-maker-assets, then when it's the case that an asset is very likely to always be downloaded by every player, then we can save the repeated download and bundle. The game engine assets should reflect the current, particularly when we have data available that can guide those decisions. It should also be considered that releases are easier to come by, and it's not necessary to predict many years out in the future - releases can happen more often and in theory every pre-release is itself a "stable". We are out of the business of pushing "unstable code" that is fixed months later all at once by the maintainers, each incremental addition is ideally solid (though TripleA makes that extremely difficult, so we're not perfect at it).
-
@LaFayette said in Pruning image and sound assets distributed with the game engine:
@Cernel Okay. Perhaps the perspective difference is whether bundled assets should be making a bet on what will become heavily used. Notably such assets can be simply included here: https://github.com/triplea-game/map-maker-assets, then when it's the case that an asset is very likely to always be downloaded by every player, then we can save the repeated download and bundle. The game engine assets should reflect the current, particularly when we have data available that can guide those decisions. It should also be considered that releases are easier to come by, and it's not necessary to predict many years out in the future - releases can happen more often and in theory every pre-release is itself a "stable". We are out of the business of pushing "unstable code" that is fixed months later all at once by the maintainers, each incremental addition is ideally solid (though TripleA makes that extremely difficult, so we're not perfect at it).
Right. Of course, the more time passes, the more maps will potentially not use assets that they might have used. Also, I believe that it would be wrong, then, to go removing any assets from maps on the basis that it is now provided by the program, while, on the other hand, I think it wouldn't hurt removing from the program an asset that proved to be unpopular, and adding it to the few maps currently using it.
I think that default assets should be used only by basic or simple maps (probably v4, v5 and v6 should have their current units folders removed, if actually useless (might be the mapmakers just didn't realize they didn't need to have them)), that don't really specifically care about the skin quality.
Still in my opinion, any really original map, instead, should fully feature all the assets it uses, even if currently some of them would be straight duplicates of what already offered in assets, so to assure the complete stability of the original skin for that map.
Just to let you know my point of view on assets usage. -
wow i read all this ... I get that your two philosophies are you both determining what is more important and time saving by determining where to store the old and eventually unused image files in the main game download or storing them in the map download
Can you have the main official release of the game run and then it notifies you there is new updates available and then you just click it and you're updated?
I'm just saying I'll download the official stable game once and then I want it to tell you when theres updates so you don't have to keep redownloading the main game of course -.-.
-
@Captain-Crunch in my simple viewpoint as always
They are both way to concerned about the minor things and always seem to forget about the major things …. But ignore me I am always concerned about the lobby and the players 
-
@prastle hehe well it sounds like Lafayette might know more about the technical side of it and making the main game download as small as possible and for some reason I enjoy these nerd discussions!
-
shouldn't all (as many as possible) heavy (sound/image) game assets be in the domain of maps?
