Puzzle Game Water Rendering, discussion with Casey Muratori

[@colterwehmeier7258]

I absolutely loved Casey's blog articles about the Witness! Happy to see him return.

[@MidnightSt]

this video would actually be a great explanation/demonstration to people about why gamedev is so hard and takes so long. just sending them this video and saying "all they do is talk about how to best do it for the water to look good. they don't even program anything, just try to come up with a way to do it which would work for that game. now imagine that this is one half of the process (the other being actual implementation), which basically needs to be done for every single element of a game".

[@buoyancydabl]

Yeah, exactly. Programming is not violently mashing the keyboard. Its a brain storm of designing a feature, implementing idea, debuging, fixing, managing, thoughts about proper implementation, restrictions etc etc. Keyboard is just an instrument, like a piano in the hands of pianist, or a pen of architector.

[@magnuswootton6181]

JBlow is really good, I hope he hasnt quit and hes making something even more monstrous now!!!

[@benjaminpedersen9548]

He hasn't.

[@4.0.4]

This idea of painting water direction vectors sounds awesome. That, + a low-res fluid sim (to drive it) would be great for a lot of games.

[@SETHthegodofchaos]

"I dont want to be type cast, man..." Haha :'D

[@KaiwenTeh]

23:20

[@flamendless]

Imagine creating your own programming language and making a high quality game as a test for it

[@zephaniahnoahmusic]

Blow, "I don't wanna be type cast, man." Casey, "Too late!" 😂

[@Elrinth]

I think this looks visually pleasing and seems to have good controls, but hard to tell without actually playing it. Great hearing some poetic talk about the development stuff :)

[@fromatic2]

I would say it's good art direction

[@lexxandera]

The Witness is my favourite puzzle game. LOVE YOUR WORK!

[@junosoft]

Last time I was this early I was working on my game

[@hatema]

on*

[@junosoft]

@@hatema fixed

[@koktszfung]

what stops you from working on your game lol

[@EmDee3d]

I really hope he consults some good concept artists to flesh out what they're after. Seems like that would go an awful long way to deciding what to make

[@bloergk]

I'm guessing the characters, at some point, escape the grid structure and start moving in a continuous way, changing the gameplay in unexpected ways (easier in some cases if you can just push stuff at any angle and stop being constrained to the 4 neighbouring cells, but harder in some others like pushing some rock along a 1 cell wide bridge becomes tough if you're not automatically parallel to the bridge)? Neat idea Jon! And it's a cool representation of an existential crisis over your role in the world, and norms constraining your thoughts...

[@VolcanicPenguin]

That does feel like a very Jon Blow thing to put in your game

[@mvargasmoran]

11:09 hey! I also hate with passion everything that has software in it. it works correctly or as expected.

[@isaiaslafon9329]

that "mhm" from Casey drived me crazy!!! hahahah... nice video tho!

[@koktszfung]

what is realistic may not always be aesthetic

[@The1wsx10]

what if you change the viscosity of the rivers?

[@vast634]

Graphics during development: I intentionally keep them simple when developing the mechanics and prototypes. Using greybox, low poly, untextured / prototype textures. The menus being just plain textline and plain buttons. Simple lighting. Adding pleasing visuals too early can define the artistic style too early too. Unconsciously you would try to iterate on those early visuals, wich might not be the best direction to go in. For visuals I rather make a mockup level without complex mechanics, that is all about nice assets, lighting and post processing. Its easier to do right, once the mechanics and parameters of the world are defined.

[@barbatruc6344]

Jon, I noticed that you sometimes mention the « moment you became very good at game design » in several streams and videos, and as a rookie game designer trying to learn how to become very good, I wonder if you made a stream or video similar to this one about how to become a good game designer ? And, if not, would you consider making one on that subject ? I mean I’m sure it’s such a complicated question that maybe it’s almost impossible to answer it properly but that would be so interesting to at least hear about how you actually became better and what kind of work you achieved in order to do so ! Honestly that would be great. The game industry lacks of deep games as the ones you make and learning a bit from how you work, I believe, would benefit to a lot of us. Anyway, nice video !

[@barbatruc6344]

For those interested, I asked the question to Jon during the last stream and he answered that he got no particular advice except to improve the game until it’s very good.

[@narnbrez]

@@barbatruc6344 bang on the keyboard for many years

[@p4nc4ket1me]

Great talk! Just a note, the audio is quite low :)

[@charlesalexanderable]

I think you should maybe not seed the velocity at corners and junctions, or seed them based on precalculated templates for all permutations, as is it looks like flow is coming from somewhere like the ground tiles instead of adjacent water (only watched the first 20min so far or so so sorry if it is addressed)

[@joelodlund6979]

I like the book shelf, kind of open and irregular. Anybody know what it is?

[@ar_xiv]

that is... two or 3 bookshelves next to each other lol

[@ctx4241]

Good point for first part of graphics being un/important is that nowadays graphics are being overvalued rather than undervalued. And that is not opinion, ask any gamer what their top5 favourite games are. Also, Indies are able to base their game on gameplay and make very well, giants have games looking better than movies and I wonder in how many of them are featured in those top5 lists.

[@saniel2748]

Yet game needs to have style. Graphics is like illustration in comics, they may be cheap, but if they constantly look tasteless or just bad you'll have much worse time with the story

[@SimGunther]

Last time I was this early, Jonathan was working on a compiler back end to directly emit machine code. Any projects or resources that embrace this idea?

[@32gigs96]

Does he have a backend that optimizes the code or is it llvm or something

[@johanrg70]

@@32gigs96 He's using LLVM for that.

[@benjaminpedersen9548]

@Simon Farre It is not pointless for two reasons. 1) It compiles much faster than LLVM. Thus it is good for testing. 2) He plans on adding optimization features to it in order not to be dependent on LLVM.

[@SpaceManCharles]

Does Jonathan use any version control for his games?

[@sherwyngolgi7197]

he uses subversion

[@simplesmilerwastaken]

He uses Subversion.

[@SpaceManCharles]

Thanks guys

[@jay.rhoden]

Did this become a game? What game is this? It seems hard to tell from just this early work.

[@sada4483]

Still in development. Search up "Sokoban Jonathan Blow" for more footage. There's even levels with finished artwork shown.

[@jorava8768]

Jonathan's team has been working on it since The Witness came out. He's also creating the programming language and the engine for it, so it's taking some time.

[@axeldahlberg7043]

How many people are actually working on the game?

[@adriancardenosocerezo]

Please!!!! I love Braid. I have a Mac and I have had an actualization for Catalina. Braid don't work in Catalina!!! Please!! Please!! I hope in the future that Braid can be played in Catalina!! Thanks for all yours games (Witness is one of the beautiful art experience that I have lived)... Sorry for my english and thanks again.

[@julkiewitz]

Apple graphics API is a complete mess. They just can't settle on anything but keep changing everything every couple of years.

[@pajeetsingh]

Video games are pillars of civilization.

[@stuartbrown3070]

At around the 2hr mark you talk about your experience relying on claims published in academic journals - this is a problem seen in many fields and arises from a careerist race to publish.

[@syzygy6]

if you’re setting up a streaming box you have no reason to waste your time installing Windows so that should save you headaches in the long term (at the expense of short term headaches with Linux or whatever)

[@Technomancr]

*The "Fluid Dynamics Chat with Jonathan Blow" video is just a copy of this video with audio not synced up properly. The comments are disabled there, so I am commenting here.

[@jessemccarthy1893]

Ya it was weird when there were two Caseys and one was five seconds from the past

[@adruvitpandit5816]

What gaming engine is that? It has in game scripting commands! Woah.

[@peeling]

His. Written in his programming language.

[@guyug6940]

It’s unity, jon’s favorite engine!

[@KU-mg9el]

both rub me the wrong way

[@MidnightSt]

Even though it's obvious what you mean when you say "i'm going to make it a 2d game because I don't want to deal with 3d", still, referring to Braid with all its timetravel as a "2d game" feels weird :)

[@azizas9366]

Is Jon doing this game in C++?

[@JesseStiller]

In Jai, as a test-case for his language

[@rauchu5861]

bruh i thought that was devin nash for a moment BAHAHAH XDDD

[@logixindie]

First time I see someone hates the word "mechanics"

[@alexnoman1498]

It constricts you tremendously in thinking about the game, if you only think about well-defined mechanics. Imagine Lucas Pope built any of his games with known mechanics . They are unique and expanding the horizon. As soon as you think about mechanics as legos to build your game out of, you're quickly in the realm of the mundane/known. Also you then have to contrive the message/goal into these mechanics... When I think mechanics, I think Ubisoft. Just throw in stuff everyone knows and isn't too bewildered by, including the story. I'm just rambling at this point.

[@VolcanicPenguin]

A mechanic hit him with a wrench as a kid, can't blame him for feeling that way

[@booates]

arent they both art game hipsters? its funny but not that surprising, lack of mechanics is probably the main criticism of what they make/like

[@VolcanicPenguin]

@@booates wait, how does Jon Blow's games not have mechanics? It's all game mechanics, he just wants to use some different word for it

[@booates]

@@VolcanicPenguin well i am guessing/assuming here, that he is also friends with other indie devs who make stuff like walking simulators, maybe im wrong about that. its a pretty silly thing to dislike so you'd have to ask him for the real reasoning assuming there is any.

[@bolicob]

mhm

[@salysPRO]

mhm : )

[@narnbrez]

i see why he disables comments :p

[@adruvitpandit5816]

Jonathan speaks so much i just loose my chain of thought what he was saying. Really need to focus, or please Jonathan could you compress your thoughts to few sentences.

[@crimsonhawk52]

train* of thought is the common expression, but I like chain more

[@RicardoSuarezdelValle]

I think the problem is not following the rules of the game, if the water behaves like water it should behave like water not like weird conveyer belts, and the design should bend to the fact of water physics, if you want rivers you need real physical world rivers, is you want ocean currents you need the mechanics that produce ocean currents etc, if you want a magical world were water behaves weird then you must declare the rules of this world and make the "water" behave according to that

[@VladyslavPavliuk]

Does anybody know what the headphones do they use?

[@Cloxygen]

mhmm

[@johnsmith-jw1do]

Covid has taken its toll on Casey's hair. The jew fro is real.

[@MrHimbeertoni]

I think Casey has Italian background, not jewish

[@truthteller4689]

Seems a bit like Baba Is You

[@VolcanicPenguin]

It's basically a clone of Baba is You and Minecraft

[@XDinky]

Both baba is you and this game are based on the same class of box-moving puzzle games. I believe the first one was Sokoban

[@ddstar]

rofl.. casey's answer was to implement some overly complicated water system.. typical

[@dimarichmain]

Mh.

[@chillagma]

Your Programming langauge NEEDS MORE capability. Jai is basicallally a downgraded c++. C++ can program several things. Jai can only program games about the same level other game engines.. Do more amazing things!

[@theitatit]

C++ is awful in many ways.... header files are garbage and slow compile time is just horrible... Jai on the other hand is attempting to increase productivity with fast compile time and being able to do more by typing less.

[@chillagma]

I AGREE with you but I BELIEVE Johnathan can make his engine more powerful.

[@TheSandvichTrials]

Jai is much more powerful than C++ in my eyes, at this point. Nothing about it limits it to games. Do you have any concrete suggestions you want?

[@chillagma]

@@TheSandvichTrials can it do things like panaromical and miegakure very easily? It seems like jai's language does genre specific games in my opinion. I want to see more "outwordly" things like those examples. Also, it would be REALLY COOL it could do gpu engineering too.

[@TheSandvichTrials]

Anonymous you seem to be conflating game engines and programming languages. To suggest Jai is somehow videogame genre-bound is bizarre. Nothing stops you from using it to write a 4D game like miegakure, the same way C++ can... and I don’t know what you mean by doing “gpu engineering”, anything that can call into C (which Jai obviously can) can use any of the currently available graphics apis. If you mean Jai should be able to target shader code, that is something Jon has thought about but isn’t hugely interested in, he thinks it could be a user library instead.

[@truthteller4689]

Trying to reinvent MineCraft????

[@VolcanicPenguin]

Sees a 3d grid based game, must be a Minecraft clone

[@flamendless]

Every 3d and has "blocks" mean a Minecraft clone right away?