[Lunick]

How does DukeGDX load them all?

[necroslut]

oasiz, on 19 April 2020 - 05:04 AM, said:

Each textured enemy variant in carib would need to be a new enemy essentially in order not to overlap existing things.

You should be able to use mapart for that. Of course, you'd have a copy of the enemy art for each map, which would be very wasteful. It would be useful to be able to define art files per map through con, sort of how we define level music.

[oasiz]

Lunick, on 19 April 2020 - 08:52 PM, said:

How does DukeGDX load them all?

Not sure but I think it doesn't have full support for on the fly loading and even then I think it probably does a soft restart.

necroslut, on 19 April 2020 - 10:35 PM, said:

You should be able to use mapart for that. Of course, you'd have a copy of the enemy art for each map, which would be very wasteful. It would be useful to be able to define art files per map through con, sort of how we define level music.

I would say that per map/mod def would make more sense.
Ideally one would distribute a map that defines what things to load in/out per map, be it art/music/sounds/maps.
Ideally it could have a stub that allows you to define an episode as well so that map packs and mods can be made truly modular instead of overwriting.
A new map format could help resolve a lot of this since you could basically start referring to assetblob + index instead of just index.
Adding such indexes during conversion would be trivial and this could allow some degree of simultaneous loading even.
During map load each assetblob could have a list of valid hashses (version differences are usually irrelevant with assets) and it would search for said files and "mount" them.

Problem is that CON needs to be compiled and CON nowdays dictates a LOT of stuff besides just the stuff you see after "new game".
When defining enemies and such "again", the copypasta enemies would attempt to define many duplicate variables, states, etc..
You'd have to re-tool a bunch of it.

Duke was designed to be moddable but how you load these is rather "vertical".
Considering you load modified code and everything, it's not too far from even comparing the end result to a very user friendly romhack.
Maps and such are very naive, they just point to resources, which in this case occupy same indexes.

As a stopgap, something what WON HL1 did would be nice where selecting a mod/map would just perform a quick soft-reboot and sort of seamlessly allows you to jump between mods. I have a huge interest on getting something like this working as I believe that the rather difficult mod loading is one of the biggest reasons why people mostly stick to stock assets as it's impossible to keep track of what launch params each mod needs, reducing the amount of people who even bother to try. Someday maybe..

[Ninety-Six]

Maybe it could be a community-focused effort or something?

[oasiz]

I can surely be.
Map format however is not extremely simple, once a map is loaded in, the tag fields, extra and such work as rudimentary scratchpad ram during runtime for effects, actors, etc..
This has been extended over the years but that also just bloats the struct each time.

eduke devs have some ideas and it's roughly in the pipeline somewhere.
However this won't really solve the modding issue yet, but a first step.

[necroslut]

oasiz, on 19 April 2020 - 10:57 PM, said:

I would say that per map/mod def would make more sense.
Ideally one would distribute a map that defines what things to load in/out per map, be it art/music/sounds/maps.
Ideally it could have a stub that allows you to define an episode as well so that map packs and mods can be made truly modular instead of overwriting.
A new map format could help resolve a lot of this since you could basically start referring to assetblob + index instead of just index.
Adding such indexes during conversion would be trivial and this could allow some degree of simultaneous loading even.
During map load each assetblob could have a list of valid hashses (version differences are usually irrelevant with assets) and it would search for said files and "mount" them.

Problem is that CON needs to be compiled and CON nowdays dictates a LOT of stuff besides just the stuff you see after "new game".
When defining enemies and such "again", the copypasta enemies would attempt to define many duplicate variables, states, etc..
You'd have to re-tool a bunch of it.

Duke was designed to be moddable but how you load these is rather "vertical".
Considering you load modified code and everything, it's not too far from even comparing the end result to a very user friendly romhack.
Maps and such are very naive, they just point to resources, which in this case occupy same indexes.

As a stopgap, something what WON HL1 did would be nice where selecting a mod/map would just perform a quick soft-reboot and sort of seamlessly allows you to jump between mods. I have a huge interest on getting something like this working as I believe that the rather difficult mod loading is one of the biggest reasons why people mostly stick to stock assets as it's impossible to keep track of what launch params each mod needs, reducing the amount of people who even bother to try. Someday maybe..

This goes a bit beyond (and aside from) what I meant... I was talking only about art replacements, not code/sounds etc, and specifically for maps inside a mod/episode rather than per mod/episode. Now, I don't know if "mapart" even works inside an episode or if it just triggers when loading a usermap, but I imagine expanding it wouldn't be too difficult if not.
There are many cases of episodes or mods where, for instance, a reskinned enemy is used only in a level or two (such as a secret level), or just one episode out of several – sort of how Plug&Pray does it.
Currently this requires duplicating the entire enemy/character (both art and code), even if they're identical aside from art and even if it is intended as a complete replacement.
One might also want to use a feature like that to change the player's outfit between levels, like in Time to Kill/Zero Hour.
The reason I brought up con is that we already have user.con where episodes and levels are defined. As mapart is already implemented, adding something for per-map art in user.con would only slightly expand on what is already there.

I do agree about the way Half-Life handled mod loading though, with the Custom Game (or Select Game or whatever it was called...) menu. I think I've suggested something similar for ED32 way back... Still, that would only be relevant for changing between mods easier, not for handling changed content within a single mod.

[oasiz]

mapart is still a hack around the actual issue I think.
There should be a way to load in multiple assets from various expansions.

Sure you can remap the tiles but this will result in maps that require a very specific hacked tile order vs. just using "hints" (Use this tile from this GRP)

[Mark]

An audio clip from one of many little side actions you'll encounter while walking around in "Suburbs".

Attached File(s)

•  shootbitch.mp3 (118.95K)
  Number of downloads: 39

[Mister Sinister]

One more.
HEALTH, HUSH and HASTE starting rooms. May change a bit in the future, but that's what I have now.

Spoiler

https://sun9-37.userapi.com/c206816/v206816318/f2d44/0L1AsKPzdoA.jpg
https://sun9-40.userapi.com/c206816/v206816318/f2d3c/8PF8i0XtRu8.jpg
https://sun9-43.userapi.com/c206816/v206816318/f2d4c/rpZemlAZsnM.jpg

[Ninety-Six]

Oh, there's three now? Crap I must have missed the second one.

[Mister Sinister]

Not yet, but there's going to be three more.

[Sanek]

Mister Sinister, on 20 April 2020 - 07:19 AM, said:

Not yet, but there's going to be three more.

So is it going to be the episode or single user maps?

[Mister Sinister]

According to my plan it's a map pack with a little extra thrown into it.

[Ninety-Six]

Another pack on the horizon? sweet

[ck3D]

Oh man, now I'm kind of tempted to postpone playing Health till the whole pack comes out so I can enjoy the whole series the way it's intended to be!

Really consistent aesthetics and clean looks throughout and the concept is really cool - looking forward.

[Maarten]

Great to see so much activity in here! Looking and sounding good, all!

---

Meanwhile my map is 1350+sectors, 10800+ walls. Going steady ;) Lately the focus is on the gameplay as well.
Most of this shot was made a few months ago, but felt like sharing it haha

Attached thumbnail(s)

• 

[Ninety-Six]

Looking like 2020 is going to be a pretty good year for duke maps. Unless I'm entirely wrong about the release forecasts but I'll play them no matter what year they come out.

[ck3D]

Maarten, on 21 April 2020 - 10:35 AM, said:

Great to see so much activity in here! Looking and sounding good, all!

---

Meanwhile my map is 1350+sectors, 10800+ walls. Going steady ;) Lately the focus is on the gameplay as well.
Most of this shot was made a few months ago, but felt like sharing it haha

Every screenshot you've posted thus far has a very original look to the point where I can't even tell for sure if the current, relative absence of detail is a bad thing. Unique-looking style right there.

_{(I think I'm seeing misaligned textures on that planter and can't seem to refrain from posting remarks on WIP screenshots because it always still triggers my OCD)}

[Maarten]

@ck3D: that's a interesting thing you're pointing out about the somewhat lack of details ;) I guess my style has gotten a lot cleaner compared to my previous maps (Toxicity, ABBA). It's also a choice because the map is very HUGE (in fact it's one BIG outdoor location). I might be posting a 2D Mapster shot...someday. But for now, this is enough information :P

I`m also writing music which I want to record for this map - just like my previous ones :)

[11bush]

A couple pictures of an ally bot I'm working on for a Korean war map.

Attached thumbnail(s)

• 
• 

[11bush]

Changed some stuff to make him look more authentic.

Attached thumbnail(s)

• 
• 

[ck3D]

Map 2 progress has been steady but slow because the part I'm finalizing now happens to be a museum, the layout of which has been done for a long time now but I had never filled it with all the installations and whatnot till now, so I've essentially just been adding detail, having fun with sprites in that section is a nice way to save up on walls for a change and coming up with creative structures in the context of an in-game museum feels meta as fuck, it's a funny excuse to go a bit nuts with stuff. Once all those rooms are filled I'll have that one last underground section I'm thinking of to design (2500+ walls should hopefully be comfortable) and then I'll be done with map 2. I actually can't wait to start all the other ones, but if I start several of them simultaneously I feel like I might eventually just go fucking crazy (but maybe I should try).

[Micky C]

Are you going to be using any sloped sprites?

[ck3D]

I don't think so, or at least not just yet - as of now, generally speaking I'm thinking of getting done with the core design of all the maps vanilla style, before I finally update Mapster32 and see if/where I feel like implementing this type of new tweaks eventually. Just more simple of a workflow for me (so far at least, that might change as I progress), but you're right, sloped sprites in the context of a museum or just abstract places/levels in general would have infinite potential. The installations so far do utilize the vertical axis, though. What's funny is that coincidentally, the whole idea of that map is essentially being one big slope, so the gameplay itself is already pretty diagonal.

[Dukebot]

I am close to have a beta version of the map, right now 1501 sectors and 15956 walls. Still need to do some polishments and gameplay adjustments before there's a playable beta. The gameplay of the map it's about an hour, but after modifications and putting more enemies probably it will be longer.

https://i.imgur.com/GX8UFFx.png

[Merlijn]

Sounds (and looks) great Dukebot! An hour of gameplay, wow must be a big map. ;)

I'm also making steady progress on my map, currently at 1236 sectors and 9319 walls. At this point, it's probably save to say there will be a Shaky grounds part 4.
There's still a lot of ideas I have for this map and I want to do them properly. Also, as often happens with me, certain sections become larger than planned.
I was going to add a small alien hive section but it ended up being a lenghty and significant part of the map.

We'll see where it goes, but I already have things mapped out for this series and I don't think all of it will fit within this map haha.
And I see no reason to cut things when I can just extend the series.

[Dukebot]

Thank you @Merlijin! What you said it's great news, big fan of Shaky Grounds and your maps in general! Great to know that the series is going to be longer than you planned this is really good news!

[Merlijn]

Thanks for those kind words!
Funny thing, shaky grounds was supposed to be 1 map at first. But the theme gave me so much inspiration that it's becoming a series now. ;)

Part 4 should be the ending of the series, right now I'm thinking of releasing it as an episode/map pack at that point.
I think part 3 will still have its own release though, like the last 2.

[Ninety-Six]

Merlijn, on 02 May 2020 - 01:34 AM, said:

right now I'm thinking of releasing it as an episode/map pack at that point.

If that happens (or even if it doesn't, I suppose) I'd love to have another Merlijn map pack to play in rotation alongside Red.

Spoiler

Even if Red was never an episode but that's nothing some fiddling with the USER.CON can't fix.

[Slap McBald]

I've been kinda silent about Blood And Ice recently, but here is a new screenshot.

This is a boss fight room for episode 1 (Guess what it's based on!)



small hud edited from this

credit goes to triaxis (i don't feel good crediting triaxis because he's beta bryan's brother i think) and oasiz