Fractal Softworks Forum

Please login or register.

Login with username, password and session length

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - xenoargh

Pages: 1 ... 3 4 [5] 6 7 ... 338
61
General Discussion / Re: The Phillip Paradox
« on: May 15, 2023, 10:26:29 PM »
IRL, the Sindrians would be a client state of the Perseans, most likely. Too small to defend itself, but valuable enough to be protected.

The mystery is why there aren't Tritachyon AM production facilities directly competing to supply the Sector's gasoline; maybe they got blown up during the AI War?

62
General Discussion / Re: Colony Threat event
« on: May 15, 2023, 10:01:37 PM »
Note: I haven't played this version yet, but here are some thoughts:

1. The problem with the Path and Pirates is that there isn't a solution. They're not like, say, Barbarians in Civilization, where they can be a game-ending threat but are ultimately something the player can get rid of. In game design terms, they're just a really fancy form of constant, fluctuating debuff. The mechanics have changed three times now, but the problem is basically the same.

2. During endgame, the player, unless they're doing a terrible job, is rich beyond all reason. Again, haven't played 0.96 yet, but I presume there are plenty of ways. There's still nothing to spend this cash on after a certain point.

So... why not let the player dump their millions of credits directly into their own military, via some sort of economic mechanic like the guns-vs-butter controls in Master of Orion 2? In addition, have some method to set their fleets to be aggressive, rather than loll around in their Systems. Even, heck, allow the player to target other Systems, like the Factions do.

Something like Hazard Incentives- "Military Budget" or thereabouts; a lategame mechanic where the player can dump whatever sums they want into the economy, and the result is offensive military capability that effectively ends Pirate predations? Or, heck, just let the player dump all those Onslaughts they've just captured onto their home Markets and boom, they now have a fleet worthy of the name. That the economy is neatly abstracted is cool, but this is an area where it could be a lot more interesting. Owning all the toys... but never really using them... isn't very satisfying.

3. It appears from posts that near-elimination of the Path and Pirates is now possible via Adventures. This is great, but should it be the only route? Why can't the player use Raids to effectively end the incursions, by crippling the Markets involved? I understand that "no capture is possible" is still the official game-design line in the sand, but surely it should be possible to drop 1000 Marines and 10K Supplies on a location and tell them to keep disrupting the place until further orders, and then they're an event that, while it will eventually run out if not "refueled", is effective? This shouldn't be practical with larger Markets, I think- anything above 4, say, should start costing too many Marines and Supplies- but that would still allow players to take a very active interest in ending the Path / Pirate threat at the source.

Or, heck, just let us nuke them without everybody else declaring war. I mean, they're the Pirates; no love lost there, you'd think.

The Path's more complicated, and deserves more nuance; presumably the Luddites would get... upset if you drop tankers of AM on them. They probably need to be handled through political solutions or perhaps a "strict immigration control" that severely nerfs colony growth (in the Future, there's still too much paperwork involved) but also keeps Path cells largely unable to form.

Another idea: instead of merely destroying Pather bases, let the player choose to launch a special Raid that gets intel on the cover names of the cell members on their worlds. That sounds cool- a raid that results in actionable intel is very real-world.

Anyhow, just food for thought. Players might not want to destroy the Path, but rather manage them until they can figure out a political solution- perhaps they agree to dedicate a few Terran worlds to Path-approved programs that lose money but make them happy, or other options. But destroying them should probably be an actual option.

4. A mechanic that makes these endless enemies fear/respect the player's Faction should be introduced. Heck, it should work on the major Factions, too. I presume that inter-Faction diplomacy still isn't really a thing yet, outside special story events but at the very least, players who've managed to grow to X size or have defeated Y DPs of fleets from <Faction> should probably see less constant attacks. This was one of the hidden variables in Mount and Blade: Warband; if the player had a fearsome-enough reputation, enemy parties tended to stay away unless the player committed to a siege.

As it is, in my last long-term game, I burned Story Points every time I wanted the Factions to go away (which was only practical in a modded game where I could get XP quickly enough). In any unmodded game, I'd probably cripple everybody's economy with AM strikes before founding a colony, just to keep them too busy and weak to pose a real threat. If they're going to all end up hating my Faction anyhow, it might as well be after they're no longer really dangerous. Surely there should be more... constructive ways to get along with your neighbors, lol. Don't get me wrong, it should probably be impossible to have happy relations with every Faction, but surely it should be possible to make nice with the Perseans and Sindrians, be neutral with Tritachyon, leave smoking holes where Kanta used to hang out, and make the Hegemony's fleet admirals nix any mission into my space out of pure terror of the consequences, lol.

If you're using these mechanics to simply prevent a win condition, I think that's fine for testing now, but it's probably not satisfactory for the final game design. Let people achieve victory, then start a new game where the Factions are richer, have more colonies, are more aggressive, get bonuses, etc. and see if they can survive. That seems like it'd be fairly easy to achieve procedurally and it'd be more satisfying, imo.

63
Suggestions / Re: Changing Fast Missile Racks
« on: May 15, 2023, 09:13:58 PM »
It would pretty much mean the end of that alpha-strike tactic.

Seems like it'd be easier to just fix the System AI.

64
Suggestions / Re: Make Neural Integrator cheaper
« on: May 15, 2023, 09:12:22 PM »
Does using X to take control of [REDACTED] no longer work? It did.

65
General Discussion / Re: The Phillip Paradox
« on: May 14, 2023, 09:03:13 PM »
Maybe, instead of a perma-debuff... something that's more of a tradeoff?

Seems like this just turns Andrada's fleet into a running gag, rather than a viable sidegrade faction.

IDK, I'm speaking from the perspective of somebody whose modded game has all-unique ships for the Diktat, so my perspective may be a bit skewed. I think the Diktak's interesting, in the sense that they're basically the oil-sheik faction, and I went with the idea that they were using Hegemony-adjacent tech, but weren't totally incompetent... just, uh... enthusiastic for the Leader's ideas:
Spoiler
[close]

66
Interesting idea: maybe make the falloff for these things gradual, with < 10LY being ideal? It's not like the people you've awoken or the stuff Hypershunt produces will magically disappear.

Now, that said.... I have yet to have a game where I've bothered getting Hypershunt working, largely because it always seems to spawn in an inconvenient place. I look at the math and I'm always like, "not worth the trouble", especially since by the time you can afford to turn them on, you're pretty much done with money for all practical purposes.

If we're ever able to non-destructively attack other colonies (i.e., actual capture / control, the Core powers can expand, etc.) then something like that might make sense as a necessary evil to allow an economy large enough to win a war, but as it stands, Cryosleeper's the win, largely because it's so useful for early growth, whereas Hypershunt always feels like a bunch of work (even just collecting enough Atlases is time-consuming sometimes).

I'd really like to see more ridiculous endgame stuff like that- building Dyson habitats and so forth- but until there's a real point to building super-economies, it's just an endgame toy.

67
Mods / Re: [0.96a] Hiver Swarm - V1.08b - 05/12/23
« on: May 13, 2023, 06:47:08 PM »
Just a quick thought: they might work with RCW if using something like EZ Faction to populate them as desired, after Nexerlin runs, rather than being entirely at the mercy of the generators. That might require a fire-once script; I've never looked at how RCW is implemented, but I presume it's much like early Exerelin's system, so it should be possible to run after that, since I presume it's largely just extending procgen.

IDK what state the public build of EZ Faction is in re: 0.96a+ (i.e., don't yell at me if it borks due to economic-engine changes- there's a reason why I haven't bothered updating most of my stuff, and won't until Alex says he's finally done with the core economy, lol). But it might provide a starting-place for this kind of thing.

68
Suggestions / Re: Uniquify the player faction (with few more ships)
« on: May 13, 2023, 06:31:28 PM »
Overall, I'm in support of the player having some special stuff just for them.

I'd like to see some kind of procedural paintjob adjustment for different factions, though this may be extremely tricky.
I've often thought that was the right way to go, using a secondary shader pass, along with using a shader for the damage system.

The 'mech shader for Mechwarrior 5, for example, uses a system where the base texture is mixed with "paint colors" (with about a half-dozen other textures mixed in; normalmaps etc.- it's massive overkill for SS).

Damage is handled via a damage texture and "heat" texture to produce a final result via a shader. This means that instead of managing a bunch of quads, the system just has to allocate three textures to the shader, plus the damage texture for <ship style> during the render pass. If ships were organized by <sprite,style> in the rendering pass, it'd be reasonably efficient, in terms of texture loading operations per frame.

If done this way, the ship would have two render textures for these special textures, that would use the ship's base sprite as a translucency mask and also render off-screen with a shader; each time they took damage, if taking a hit heavy enough to inflict visual damage, the damage map would get updated (essentially, just a black/white mask used for the final render pass) and the heat map would get updated every frame that some value <damageheat> was still above zero, to keep processing loads down. Damage map would be permanent changes, where white values indicate that the damage texture's almost overriding the base, kind of like how it looks now (although, fun fact, repairs to Hull/Armor could be reflected by simply running the hull damage operation in reverse), heat map would keep updating until <damageheat> reached zero.

Final shader operation would look something like:

Colorvalue = base_texture_sample
Colorvalue = lerp(Colorvalue,damage_texture_sample,damage_amount_sample)
Colorvalue.rgb += damageheat_sample.rgb
Colorvalue.rgb = vec3(min(Colorvalue.r,1.0),min(Colorvalue.g,1.0),min(Colorvalue.b,1.0)),

I would think that this system would be, on average, cheaper than the current system, because while it costs some CPU / GPU per hit (to re-render the damage amount texture and sometimes the damage heat texture), it's not having to do a whole bunch of rendering passes / additional quads / etc. per frame otherwise. It would cost some more texture memory, maybe. Remember, none of those textures have to get allocated for anything off-screen; they don't even need to be updated visually until they're on-screen again- a record of damage hits could be queued instead, for example, so that damage visuals aren't even updated until the ship's rendered again. Also, we are in 2023, and even low-end gaming GPUs have 6GB or more, so it's less of a technical constraint.

IDK how the current system is implemented exactly, but I've always gotten the impression it's using layers upon layers; this might well be significantly cheaper and offer some opportunities for visual uplift as well.

Also, if the shader was exposed, the current modded system people use for normalmaps, etc., could be converted to operate within that framework pretty easily, one suspects, improving performance substantially.

69
Suggestions / Re: Contact Improvement
« on: May 09, 2023, 01:57:15 PM »
That's a great idea. That Contacts don't ever improve just makes me seek out the important ones and ignore all the rest.

70
@Big Bee: I'm glad you get it. What you've proposed is a workable compromise.
Spoiler
I am willing create an AI sub-section, per artist, when they submit works made in that way.

Caveat: I can't create new blank posts in the OP to make it into some special animal all its own. This is a technical limitation, not me being unwilling. IDK whether even Alex can do that. I'll consider asking him, but I hate asking him for favors like that, because he has better things to do with his time.

When I started this, I never imagined the OP's would get so big, frankly, so I didn't reserve enough posts for this to go on for, well, we're getting nearly to a decade now, lol. At some point, the posts will hit the character limit and  I'll be forced into consolidating artists' submissions into big PNGs, which will involve more work and make it a little harder for people to use.

This compromise, of course, would require that artists say that they used AI art tools in the first place. So that's going to have to be on the honor system. If you haven't already figured it out... it's going to be nearly impossible to tell, when talented artists use this tech as a base and spend some time on polish.

Since there's <counts fingers> only one person doing this right now, that won't add much new work for me, other than creating a new sub-section under the artist's other works.

Note: this isn't a blank check on my time. The policy would have to be reviewed if there was an actual flood of polished AI-gen ship sprites from a significant number of artists (like, if five artists all whip out 100 beautifully-polished ships in six months... I'll both be delighted that people have chosen to give so much work here, and horrified at the workload, lol).

If people just let the AI do all the work, then submit without polish (turret positions, etc.) I may have to adjust policy there as well, since in theory some kid could use MS-Paint and Easy Diffusion to make a bunch of really awful "ships". But I'll err on the side of, "takes everything" for now, because the whole point of this gets undermined if I feel like I'm having to gatekeep.

But that probably won't happen. In my experience, we get somebody really talented like Bananana who submits a bunch of stuff in a pile, because they have too much sitting around for their own projects... and then nothing but the occasional single sprite for weeks or months.
[close]

Meanwhile, a new ship appears!
AI Image Source Note: these images were all made with img2img prompts and Stablity AI's Stable Diffusion, using the model known as Dread_V3. Text prompts were probably something like, "spaceship, ((aerial view)), panel lines, greebles, (((detailed and intricate))), rocket engines, jet engines, small details, dirt, hydraulics, hatches, struts, grunge, rust, oil stains, post-apocalyptic, low-tech, steampunk, dieselpunk".

Spoiler
The Horse-Thistle class, a new "pocket battleship" design, has yet to be seen outside the occasional Merc fleet, as the sheer expense and specialized role make it unlikely to see military procurement.

The ship is, like the Pirate's Acanthus-class, an unusual "flying wing" shape, which permits it to use an unusual amount of frontal-facing weaponry. However, this design imposes the terrible limitation of being ineffective against flanking opponents, and it's woefully under-equipped for anti-missile defense. Only its built in Plasma Jets allow it enough mobility to strike hard, then run away before being overwhelmed.

Artists note:  this was made with a simple grayish flying-wing design. Why it came out green-blue, IDK, but I went with it. It's really, really hard to get inference AI to make this kind of design well; this took several tweaks to both the text prompts and the base art.
[close]


71
If somebody wants to host a thread where only AI-generated stuff is allowed, great!
Spoiler
That's not sarcasm, I promise.

I'm totally serious. If people want to do that much work for the good of the community, I'm 100% supportive of that.

So, are you part of this posse and you want to volunteer? Great. A place like this has to have a bunch of things:

1. Somewhere to park all the data that's 100% reliable. Up for eternity on a web-host where you have FTP access, basically. Not a Google Drive, Dropbox, etc., but a real host. It's cheap these days.
2. Understands that this only really works if it takes all entries (i.e., no gatekeeping, no elitism, etc.) while dealing with the occasional copyright drama.
3. Has a person in charge who wants to put in the hours of work it takes to update the OP (because otherwise, the thread's useless, the images get buried).

If somebody really wants to do that for AI-generated spaceship sprites, I'm totally supportive. I'll even post there :)

But... not me. This is my little job, and it's all I'm able to find time for.

The people who think this is "just a thread" have never, ever maintained something like this one, and might want to stare at the OP posts for a while. It took something like 3 hours out of my Saturday afternoon to finally collate, label, sort and upload the last batch (roughly 8 new artists, 60 entries), then make sure everything was working properly with the posts, etc., etc. Maybe longer- I tend to lose track of time when I'm playing Sprite Librarian (most boring game ever, -10/10).

So if you want that vision and are willing to do the work, it's totally cool with me. Get it off the ground, and I'll happily jump on that bandwagon. Again, not sarcastic- just be aware of what you'd be expected to do. It's a big deal. Keep in mind, folks... I took up this job from Medikohl, after he'd been doing it for 3 years. I've been doing it for 6. So, this is a very serious commitment. When I say, "basically forever" on the OP, I meant that.

But, if somebody wants that level of commitment... I'd be the first one to give them a massive thumbs-up.
[close]

 Meanwhile, here's a spaceship:
AI Image Source Note: these images were all made with img2img prompts and Stablity AI's Stable Diffusion, using the model known as Dread_V3. Text prompts were probably something like, "spaceship, ((aerial view)), panel lines, greebles, (((detailed and intricate))), rocket engines, jet engines, small details, dirt, hydraulics, hatches, struts, grunge, rust, oil stains, post-apocalyptic, low-tech, steampunk, dieselpunk".

Spoiler
The Longhaul-class freighter is frequently seen making the shorter runs between the Core Worlds. It's a fuel-hungry, but relatively fast cargo hauler with excellent tonnage capacity inside its modular cargo bays. It's poorly armed, but moderately well-shielded and reasonably quick, so it's sometimes converted into a drone-swarm assault vessel by Pirates, who regard such a design as an excellent first chance for a younger captain to prove themselves.

A little irony here: this is "AI-generated" in only the loosest sense. About 95% of the pixels there are me, because I had real problems getting the AI to figure out the cargo pods. Next time I'll do that as a separate job and see if it can do better, lol.
[close]

72
Discussions / Re: Now That's Entertainment!
« on: May 07, 2023, 10:49:30 PM »

73
Suggestions / Re: Add a cheat subcategory to the mod index
« on: May 07, 2023, 10:18:59 PM »
Nah, it's just too old; I thought it was a comprehensive list back to the beginning of time, is all. Sorry, my bad.

74
Suggestions / Re: Add a cheat subcategory to the mod index
« on: May 07, 2023, 10:10:48 PM »
<ignore this, I was mistaken>

75
Having an odd issue that only arose with 0.96a - I made it a habit of loading JSON/CSV data in the mod plugin, and storing said data there to be checked when required (e.g in a BattleCreationPlugin or an EveryFrameCombatPlugin), but it seems like the data is somehow made null before it gets to that point. Easy (if less-than-ideal) to bypass by just loading it every time it's required but I imagine it'd be ideal not to have to do that, and I'm not sure what the cause might be since the patch notes don't seem to mention anything relevant.
Maybe just write a generic storage class where you can expose whatever you've put in during initialization into public statics? I've done that for a few things.

Also, are you absolutely sure that whatever is returning null isn't because, say, a JSON read failed during a Try-->Catch or something to that effect?

Pages: 1 ... 3 4 [5] 6 7 ... 338