Interview with Schema - LFG Podcast

[DukeofRealms]

​

We met Tritemare (one of the LFG Podcast panelists) at TwitchCon! Here is an interview with schema , Criss and the LFG Podcast crew.

 

[Gamel]

Watching now, seems cool,

but it is sad that schema don't have a webcam feed that captures a cat on a keyboard:p

 

[jzimmerman4]

Schema needs her headphones haha

Also while watching and based on what Schema said I feel like faction points and credits could be combined to be the same thing. I am sure there would be some repercussions but overall it seems like it would work effectively.

 

[Keptick]

Schema needs her headphones haha

Ahahaha, I didn't even notice them until reading this post. Schema definitely needs a pair!

"And then there's of course the crazy ones that make even bigger ships. They've been working on them for a year or more" -criss

>_> ... <_< ... ¯\_(ツ)_/¯

Ps: Next time you want to REALLY impress someone show them the tigerlily II by refirendum

 

[Nauvran]

Ahahaha, I didn't even notice them until reading this post. Schema definitely needs a pair!

"And then there's of course the crazy ones that make even bigger ships. They've been working on them for a year or more" -criss

>_> ... <_< ... ¯\_(ツ)_/¯

Ps: Next time you want to REALLY impress someone show them the tigerlily II by refirendum

:U
Ships don't always need to be big to impress. Small ships needs love too!

 

[SmilingDemon]

i had seen the Twitch vod already.
nice format that podcast has. i did like it

... Next time you want to REALLY impress someone show them the tigerlily II by refirendum

thats a beauty, but would have probably killed the stream or the server or both ... not that good advertisement

 

[Keptick]

i had seen the Twitch vod already.
nice format that podcast has. i did like it



thats a beauty, but would have probably killed the stream or the server or both ... not that good advertisement

Naaah, it's around 500m long, not enough to kill a server. Schema was blowing up a planet with the OP rocket launcher and the server wasn't budging, so it'd be fine :P

 

[Valiant70]

Schema mentioned that there are/might be plans to completely redesign planets. I have no idea how he's going to do that, but I hope whatever it is works out and "outschines" the current ones. This comment combined with his previous mention of level-of-detail and planet optimizations have got me really excited and curious for the future of planets.

 

[Zackey_TNT]

Schema mentioned that there are/might be plans to completely redesign planets. I have no idea how he's going to do that, but I hope whatever it is works out and "outschines" the current ones. This comment combined with his previous mention of level-of-detail and planet optimizations have got me really excited and curious for the future of planets.

Pancake planets, WE want them back!

 

[Nauvran]

Pancake planets, WE want them back!

Nope, we don't

 

[Zackey_TNT]

Nope, we don't

I miss them, they where simple and easy, you could make hangers in the sides too!

 

[SmilingDemon]

downgrade to an old version and get them as a sector export then ... even though they will generate with the old resources then

 

[Jay_do]

Schema mentioned that there are/might be plans to completely redesign planets. I have no idea how he's going to do that, but I hope whatever it is works out and "outschines" the current ones. This comment combined with his previous mention of level-of-detail and planet optimizations have got me really excited and curious for the future of planets.

Maybe he is gonna scale up the planets a lot and only have the plate your on render but from far away the planet will be maybe a generated image? Also maybe increase the side count to 20 so we can have 1d20 planets.

 

[karkinosz]

Maybe he is gonna scale up the planets a lot and only have the plate your on render but from far away the planet will be maybe a generated image? Also maybe increase the side count to 20 so we can have 1d20 planets.

I don't know about the rest, but the current planets are 12 sided so to strike a balance between roundness and plate size. After all if the plates are too small that'd be useless. Also a 200r planet has around 15 million blocks. I think for a 1000r planet a single plate was in the ten-twenty million blocks area.(I tried to spawn one, yes.)

Now something unrelated, my singleplayer has none of these huge planets. Only a few 150r ones and my current builds. It's a couple GB in size data wise. For a server like peak elwyn or the current starside. The databases get extremely large. (For at least normal players) into the hundreds of GB.

Now starmade has issues rendering it, namely because it's trying to render the other twelve plates. As well because theres a limit to the number of chunks starmade can load.


That is a 1003r planet plate. It is 22.9 million blocks. That planet has 11 other similarly sized plates.

... There's a a few hundred planets in a galaxy right now. And my computer is currently boiling eggs on the cpu.

No amount of "optimization" will change how stupid the amount of blocks is on this plate.

Multiplayer servers would perish if they had such large planets.
complete plate for reference sake.

So im going to say scaling up planet sizes? probably not. The generated image isn't a bad idea. Increasing plate numbers isn't likely so long as plates are small, which will be for the forseeable future.
[DOUBLEPOST=1452783856,1452783755][/DOUBLEPOST]Oh and I gotta add, im standing on that planet. No gravity. I guess large planets have bugs too.
[DOUBLEPOST=1452784114][/DOUBLEPOST]I hope schema can pull off the miracle of making large planets work as much as the rest of us :|

 

[Zackey_TNT]

downgrade to an old version and get them as a sector export then ... even though they will generate with the old resources then

That breaks the game, you cannot get the old cake planets. exporting from what I hear doesn't even carry the planet over.

 

[Nauvran]

That breaks the game, you cannot get the old cake planets. exporting from what I hear doesn't even carry the planet over.

sector exports works.
How do you think the skid-racing tracks were made?

 

[Valiant70]

I don't know about the rest, but the current planets are 12 sided so to strike a balance between roundness and plate size. After all if the plates are too small that'd be useless. Also a 200r planet has around 15 million blocks. I think for a 1000r planet a single plate was in the ten-twenty million blocks area.(I tried to spawn one, yes.)

Now something unrelated, my singleplayer has none of these huge planets. Only a few 150r ones and my current builds. It's a couple GB in size data wise. For a server like peak elwyn or the current starside. The databases get extremely large. (For at least normal players) into the hundreds of GB.

Now starmade has issues rendering it, namely because it's trying to render the other twelve plates. As well because theres a limit to the number of chunks starmade can load.


That is a 1003r planet plate. It is 22.9 million blocks. That planet has 11 other similarly sized plates.

... There's a a few hundred planets in a galaxy right now. And my computer is currently boiling eggs on the cpu.

No amount of "optimization" will change how stupid the amount of blocks is on this plate.

Multiplayer servers would perish if they had such large planets.
complete plate for reference sake.

So im going to say scaling up planet sizes? probably not. The generated image isn't a bad idea. Increasing plate numbers isn't likely so long as plates are small, which will be for the forseeable future.
[DOUBLEPOST=1452783856,1452783755][/DOUBLEPOST]Oh and I gotta add, im standing on that planet. No gravity. I guess large planets have bugs too.
[DOUBLEPOST=1452784114][/DOUBLEPOST]I hope schema can pull off the miracle of making large planets work as much as the rest of us :|

All of the things you mentioned are reasons that planets need to be completely redesigned. Most people don't seem happy with the performance of planets as they are now, and many are also not completely thrilled with their mechanics (although there are many fans of the current dodecahedronal planets). As of now we have no idea what Schema has up his sleeve, but it could include some major engine re-writing and arcane computer wizardry that ends up treating planets very, very differently in the game program. Or, he might just have something simple and clever that no one else has thought of yet. Perhaps it's some combination of the two. We'll just have to wait and see if he reveals any more about his ideas. Right now the cat's in the bag because Schema doesn't want to promise anything.

My best guess is that the bulk of what Schema will do to planets will amount to out-of-the-box optimizations, in other words doing a better job with fewer actual blocks in memory at once. There are a number of ways to do this, and it's certainly something that needs to be done if we want really big planets like many players do. It's entirely possible Schema has other wizardly tricks up his sleeves that we can only dream about now, or that the redesign he's considering is much simpler and more clever method of solving the problems with current planets.
[DOUBLEPOST=1452818955][/DOUBLEPOST]As I'm listening further through this podcast, I hear a lot of stuff that excites me, not the least of which is the possibility of NPC crew R&D (research and development) with the upcoming crew/quarter system. Schema didn't say anything about including that in the initial release of the quarter system, but it's something to look forward to in the future whenever it comes!

 

[NeonSturm]

@[33:30 to 34:00] I heard a statement I can't agree with - that there are only 2 ways of detecting collision.

The "strangely shaped object" schema mentions is not more than 2 graphs with 2 lines each with positive and negative limits for each x. Most objects do not even come close enough to each other to require the full check (usually all but 3 to 5 when using 3D flexible "half-pyramid" grids to determinate collision partners).

@[planets]

Alone the option to have not-editable canyons to fly to your secret underground-base … how can you screw that?

 

[Ithirahad]

Now starmade has issues rendering it, namely because it's trying to render the other twelve plates. As well because theres a limit to the number of chunks starmade can load.

Just don't render the other twelve plates, or render them as flat green pentagons.

... There's a a few hundred planets in a galaxy right now. And my computer is currently boiling eggs on the cpu.

No amount of "optimization" will change how stupid the amount of blocks is on this plate.

Multiplayer servers would perish if they had such large planets.

Consider the stupid amount of blocks in a Minecraft world. Then consider that you can store any world in a simple string of text - a generation seed - along with a very small file that holds all the modified chunks. Unless someone goes and builds across the entire damn plate, much of it can be optimized out, as far as hard drive space. Hell, using a stored seed for procedural generation would mean you don't even have to load proper block data for unmodified chunks a certain distance away from you. Just generate a lower-resolution heightmap from the seed (or don't even load ANY data for the chunks if they're obscured. This is hard to determine though) until a player gets close enough to actually see detail... And suddenly the RAM issues are gone too. That just leaves CPU and collision, which is above my understanding but I suspect that some things can be done to make things work a bit better in that department as well.

Again, this system breaks down slightly if someone builds a city across an entire plate. But seeing as those things are like 750m across in your image, it's a ridiculously large undertaking.

 

[Drakkart]

go schema go finally build us a game