#💀 Cortex Engine
1 messages · Page 7 of 1
the dream would be shared tooling, it is a bit silly (but understandable) that ceramic and lime have completely different CLIs
The difficult part is that it's already so hard to make tools right
cortex contains many cool libs you people will steal
Yeah I get that, maybe the time can be justified by making the ease of use for users coming from other frameworks easier
So you try to keep thing as simple as possible
And if you need to add to that mental load a whole stack you don't know, it can make it harder
but honestly, I liked the snowkit approach
No surprise people are reinventing the wheel over and over in some ways
dump cool libs together and maintain them together
Yeah, I did try to keep that aspect of things with Ceramic a bit
snowkit did seem to have the right idea
Yes though I would suspect that the CLI part would be easier to have interopability then the rendering layer.
Ammer would have been one of the best projects in the Haxe ecosystem, I am convinced of that.
That's the thing, as far as I'm concerned, most of the dependencies I use are things I made or could modify myself, because if you hit an issue with something you can't fix yourself, it can be a serious problem
I thought you were using a lot of nodejs depedencies for the CLI?
rust kidnapped our aurel
Like, I don't use lix because I don't want to add another layer of issues. The risk of things breaking is higher than what it brings (from my perspective, it can vary depending on the context or the user)
i want something lix-like to officially replace haxelib
That would be the best option yeah
i think haxelib is close enough to being "that thing"
haxelib + haxemanager is a pretty good combo
lix has a pretty great design
Not anymore, Ceramic CLI has been rewritten in Haxe/C++
I agree, I try not to expose it user facing, because I don't want to fragment Haxe more so then the current situation.
(not released yet)
Oh nice!
i have been using the hxcpp version for awhile, works great
I had a small hope that go2hx could be used to bring in Go libraries that were similar to the nodejs deps for Ceramic but oh well, you are too fast
not sure how hard it would be to convince haxe devs to ship it by default. it is made by people with a higher profile than me and it hasn't happened yet, so the chances might be low
but just playing around with it for a bit, it feels so right
I think the best person to talk to would be @shut oracle about it, iirc Simn said he is looking to be the new haxelib maintainer, so he may have sway in bringing in something like lix for haxelib
i think a lot of it might be the node dependency
Yeah true
(might be go2hx's time to shine there, depending on what it is for)
it does make integration with node projects nice tho
you can just add it to your package.json
oh there is crobes, hi
I'm not a fan of lix
(It might be your time to shine too, because you can figure that out probably better then me, and cordinate what needs fixing for XYZ to work in go2hx to replace lix's usage of X)
i think it's a bit convoluted and also very buggy
From a quick glance, it seems that basically every class uses some form of js
(when i used it)
Yeah, I was working on some haxelib improvements, the most complex bit is haxelib git which I was improving but was blocked due to some neko crashes on windows, but those are beginning to clear up now so hopefully it will be possible to return to that!
Amazing! So you really are the savior of haxelib :)
A neko-free haxelib would be so nice lol
daz go2hx ceramic reflaxe
on second thought, maybe these keywords shouldn't be made public, hihi
... for sure that's a +4
From a (slightly less) quick glance it seems like child processes, http, and zip/tar extraction are a big chunk of it.
👀
c-ramic h-xeui?
reflaxe testing
Which keyword are you tracking? 👀
typically none, but just testing things that people have issues with
That's the goal too, hoping to move to hxcpp or hashlink
I see, http is gonna be a while until it works, the other 2 are likely to be passing by the end of the month (assuming no change of number of contributors)
Why is eval off the table?
haxelib needs to be a binary
child process is indeed one of the things nodejs was bringing to ceramic tools. Replaced it with linc_process
That would be great, 100% with this
standalone binary
So if it is hl then a hl executable needs to packaged with hlboot.dat to make it into an executable?
Probably hlc, but we were also looking into this for a faster compiling alternative: https://github.com/HaxeFoundation/hashlink/pull/744
Something that produces a exe that doesn't need any companion dll/data file would remove a lot of friction imo
So yeah hlc seems like a good idea
no arm64 😢
they'll be a few sad peeps around here 😄
yeah, it would mainly be as a way to speed up compilation during development, with hlc for a full build
That's one of the drawbacks tho i guess
Sounds like a good approach yeah
Gotta use reflaxe cpp instead
To be fair, now I made two CLI tools with Haxe + HXCPP (ceramic, loreline). It works great, definitely a valid option
No reason haxelib couldn't be done this way, but I guess HL is more "haxe core" than hxcpp
only for the brave 😉
Looks like a trojan name
totally installing a backdoor in your brain
it will crash I guess since I just realized an absolute shaderlook path is in there
ah, no it wont
does this mean im lacking a stupid amount of skill?
also audio seems to be broken
game design is not finished. jumping will cost you
i dont hear anything atleast
audio is disabled
ahh
it can/will
press ^ to open the console
or whatever it is on international keyboards
there is also a project.cortex file, you can mess with, like switching renderers
what keycode is that?
^ doesnt seem to work
I’m only testing web builds for safety purposes 😉
you know what they say
fun #1
safety #2
this game is giving me enough dopamine to pull through this evening
ScanCode.GRAVE
uh might not be chromebook ready™️ yet but seems like it works with wine lol
change the renderer to vulkan and see what it does
spirv?
no
it instantly crashes
2 quick thingies:
- input needs some work
- clear() doesnt seem to scroll up :(
it just shows the score before crashing
- input SUCKS! MVP 😄
- Scrolling is a hack!
100 % hackjob
oh ok
OpenGL almost works lol
then indeed, wine + d3d11 == best bet
yeah, didnt optimize the shaders for ogl. So that can be made to work
might be the fact that i am running it inside a container inside a vm on a chromebook
so i'll test it on real hardware later lol
encouraging
probably will get >1fps too
u prolly will
so hot reloading is active, if you mess with the css you can change the UI visuals
or mayyyybe not
interesting, it throws a null function pointer
wtf
vulkan?
no, filewatch crapping out
ah well, you cant win them all
at least this adds a few more testing considerations
ah, still had some paths hardcoded
that's actually the same behavior I see in emscripten + gles
so if I fix the shaders emscripten builds should also run fine (aside from compiling)
I quote
Tests are overrated
-- dazKind 2025
From my perspective, it's just that it's easier to quickly make an app out of hashlink bytecode. The build times for hxcpp are quite slow, especially compared to neko which is what haxelib uses now and cppia doesn't make it as convenient to bundle your cppia script as an app (though, this could be improved).
has anybody ever looked at using hxcpp cache in github CI?
incremental builds using the cache should be super fast
As long as the resulting executable can just work as is without trying to look for libneko and so on, that’s fine I guess
Using hxcpp cache in CI would make me nervous because that would mean using previous builds side effects for the next ones?
I would be scared of that too 😅 But I suppose the only way to solve those issues is to start testing it out in cases like this
I've discovered issues before where using cache can introduce an error into a build that would've worked fine otherwise: https://github.com/HaxeFoundation/hxcpp/issues/1180
There was also: https://github.com/HaxeFoundation/hxcpp/pull/1129
it depends on how the apps/tests are structured. if all tests are different executables that are built in sequence you can clear the cache when you start a build and have them share common objs during the current CI run
@void condor hah a bit better on my desktop
the recording slows it down a bit, but you can't even read the numbers
do most applications do some form of average for their fps?
is it expected that the crabs randomly disappear?
yeah, usually a slowed down average
they disappear based on the distance from the center atm
haxejam logo actually works surprisingly well as a playable character lol
haxe bcs they despawn right before they can catch me
you moved past their range 😄
the playfield is not constraint yet
it's just super basic stuff atm
but a good sample of different patterns on how to model a game in cortex
statecharts, ecs etc
does bgfx not have a shader cross compiler?
the shaders are generated & compiled from common sources to all platforms by the asset pipeline
all the resources in the game are cooked/backed versions, not the source ones
ah k
Vibe-Coding: How it started. How it's going. lolol
That’s both worrying and hilarious
lol
The replies of the first post being like "we don't need developers now, just marketing" 🤦♂️
it gets better
i'm shutting down my app 😑
Cursor just keeps breaking other parts of the code
you guys were right, I shouldn’t have deployed unsecured code to production
I'll just rebuild it with Bubble, a more user friendly and secure platform for non techies like me
I appreciate everyone
The guy keeps saying "I can't code", when we never had as many ways to learn it in human history as nowadays
I'm just amazed by the amount of energy that can be deployed to skip learning
another interesting way to look at this, is this guy made income with basically no technical skills
Oh sure that’s possible, but it’s nonetheless a fragile bet to push into production something you don’t understand at all. It’s even more meh when LLMs are incredibly powerful tools to learn faster and folks decide they won’t learn at all and instead lock themselves up with a strong dependency to a tech provided by a third party that could change the rules at any time
Anything you learn for real is invaluable compared to any fancy tool that makes it for you, and people are going crazy avoiding that
it's more of an interesting view, able to actually get users but doesn't have the technical ability to develop the thing
I understand the excitement over things that were impossible to do before by many, and now become possible
But what I have a hard time to cope with, is the amount of detour people take using these tools to avoid learning, even in situations where actually learning the craft is not necessarily harder or longer
You can really learn basics of programming in a few days now
Just open a LLM and ask where to start
When I started programming, I only had a TI89 and a paper manual with cryptic examples with close to no explanation, that was WAY HARDER to start
(Of course programming manuals did exist, but I didn’t have a desktop computer at that time 😅)
You can also search up anything, the issue with starting is asking the right things is actually very challenging if you don't know about the concept.
That’s the thing, LLMs are very good at pointing you to a concept you didn’t hear about, which is much harder to do with, say, Google search
I remember when I was learning PHP, I didn’t know the idea of concatenation, so I was searching « how to merge variables » and so on
I didn't know about for loops when I first started so I thought you had to use the timeline to do iteration in flash.
Yeah, it took me ages to just get the actual « concatenation » word
Now, here’s what Claude says for the same question:
I honestly wish I had that when I was learning the basics
omg
Maybe I am stubborn but I feel like if everything was spoon fed, I wouldn't have learned how to figure out all of the really difficult problems
the . " " .
that brings me back and hurts me at the same time
very tempted to just make something random in php for the fun of it...
I'd like to try haxe php some day
I still have trust that people who really want to learn new things will use LLMs to their advantage in that task, you can put the limit of what is doable higher than before, it doesn’t mean you have to stay passive and stop to learn
But the opposite is definitely possible too. People who won’t try to go further than what a LLM can do is definitely happening. The LLM is the limit
Me too lol
I currently use GPT-4o + search + reasoning, should i switch to Claude?
I have a summer project that will use it 😛
For programming, I find Claude better, but ChatGPT is alright depending on what you need
By the way just to play devil's advocate I searched the same thing up with duckduckgo, and clicked the first link: https://stackoverflow.com/questions/8336858/how-can-i-combine-two-strings-together-in-php
Yeah, when I searched that duckduckgo didn’t exist 😛
xD
back then it was all about forums
I basically exclusively use it for programming and explaining concepts (like batching or instancing)
I'll check it out, ty
this week I reviewed a systemd service file that didnt work right in staging and it turned out it was generated by AI and included a subtle setting that was causing shit to not work right
so much for saving time by using LLMs
have an appointment with the optician tmrrw in the morning 😬
LLMs can’t stop people to make trash lol
might have to start wearing glasses o7
One of us
i mean i feel like that is kindof the biggest problem with LLMs
it gets the broad "context" but doesnt pick up on very specific things
I also tried vibe coding for a simple prototype (webrtc, websockets)
possibly because the sample size of the option used might be too small
uh oh
It’s definitely not a silver bullet solution, but I would lie if I said it didn’t help me get Loreline done, for instance
after 30 mins I gave up, cause the output was such a fucked up mess
Imo it can be very complementary to a skilled developer
And it’s definitely better at some tasks and worse at others
i think LLMs can be useful if used correctly (specifically mock data, data transformation, docs and maybe stupid regexes in non-critical enviroments)
things that are very trivial but take time and cant be messed up (generally) is good use-case
yep
dip or no dip, there is no part dip
shit true
I think of it as an assistant, you give it very specific tasks that save you time, but you’re the one knowing the big picture and working on the interesting bits
Im also testing different stuff with my teams to discover where we can apply shit
too much dopamine to be found in this conversation
my small young brain needs that dopamine man
so far it has been a sobering experience
what do i do without the dopamine!!1
(just joking lol, i hate most social media)
anyhow bai
Sometime I just want it to fill-in a method, with proper context? Output can be near perfect
#1234939525931728906 is becoming the hangout spot now for this discord (or just us).. haha
But ask it stuff for Unity, you are out for a wild ride
I enjoy that lol, sorry not sorry daz 😂
every case where we gained an initial big win, it then deteriorated into review & rewrite loops resulting in no time being saved at all
Yeah that’s also very common. I guess some guetting used to is needed, so you know better what is worth asking and what isn’t
that's perfectly fine. Everyone is defined by the company he or she keeps around. Cortex will be fine 😄
(There are many things I do I don’t ask Claude obviously, that’s also why I don’t ise cursor/copilot suggestions as-you-type)
Cortex Congregation when
The hangout of pure programming nerdiness
The other day I asked my deepseek to convert some fucked c++ types for an extern, it did a marvelous job and I learnt something. So yeah, you gotta use it right. But the low barrier entry and the insane code generation will cause a shift in software development in general
Staying critical and ask questions no matter your starting out or veteran is key imo
LLMs are already training on their own data, so given the stochastic nature we are in for some mind breaking clusterfucks people gonna create in order to generate a quick buck
But I think anybody smart enough with a learning mindset can get a lot of good things from a LLM
true, as long as they never completely trust the generated answers they will be fine
Yeah, deepseek trained on chatgpt apparently lol
You can always dig more into a subject and the llm can often point you to an interesting direction if you don’t stop there, that’s at least my experience with it
And it doesn’t mean other ways of learning have to be avoided
Thankfully big LLM are all fed with libGen 🙃🙃🙃
sure. People that have an intrinsic motivation to understand/control something will be super-charged by an LLM under the right circumstances, the rest will just use it as shortcut to get whereever.
It's like an amplifier for people's core motivations. Interesting times
Yeah I agree with that
(for the better and the worse 💀)
brace yourself 😄
I'm just going to ignore people and keep doing my stuff at some point, that's it lol
Can't stop the world bulshiting itself
it's all hypeshit. the other day when looking at vibe coding shit I noticed a lot of people assumed that you can easily make money, bc look the dude that created a shitty plane game made 100k by adding ingame transactions... but somehow everybody forgot that this is just a zeitgeist / hypecycle thing, not a "omg, awesome product & value" one. Sad that people dont remember http://www.milliondollarhomepage.com which was the prototype for that shit
The website of Alex Tew, a 21-year-old entrepreneur, who hopes to pay his way through university by selling 1 million pixels of internet ad space for $1 each.
Just the same cycle as people going all euphoric about: NFTs, crypto and so on...
This is how I operate
Im not triggered by FOMO in my areas of expertise
Same, I realized I will be much more content doing my own things rather then chasing to get ahead of some hype wave.
Same here, I'm saying I "will" just do that, but actually I have been doing it for a long time. Never made Ceramic for getting into the hype train heh
And using Haxe, lol, just the worst choice for that
curious, u guys ever were ahead of some hype wave before?
Also hype trains are for lottery ticket buyers, it's not worth the effort, hard skills and focusing on the process are simply better strategies.
I was using XMLHttpRequest before Web 2.0 euphoria, does that count
DHTML baby
thank god noone calls it ajax anymore
I'm hoping writing interactive stories with Loreline script is going to be the next hype bubble 🤪
And on Internet Explorer you had to use window.ActiveXObject, I remember it like it's yesterday
dont you miss those simple times nowadays? 😄
Not super hype trainy but more I am satisfied that the fundementals for me wanting to do go2hx were double confirmed by Typescript moving to Golaang and the lead dev talking at length on why they chose Go which is almost 1 to 1 why I thought it was a good pick for Haaxe.
oh! now that's something
I miss the mood, not the tech
prolly the same for me
People where doing things because they found it's cool
Now they are mostly doing things because they build a business
in my case I was ahead of the whole low budget autonomous drones stuff
(at least on the internet that was much less corporate than nowadays)
Too early is still bad timing I guess
oh for sure
Oh that's pretty neat
im lucky im not motivated by money 😄
Same here, lucky enough to earn enough with my 4-days/week job and being able to do whatever cool thing I want next to it
it is! but at the some time thinking about teaching people on how to make linux-machines fly in coordinated swarms BVLOS and they use it to kill eachother was not so cool anymore
Btw, today I reached a milestone at work: Make a native library with HXCPP and plug it to Unity, run it fine on Android. The thing is going through Haxe/C++ <-> C# bindings and Haxe/C++ <-> JNI/Java bindings without issues 😛
that shit going on in ukraine was like my worst nightmare
lovely!
That's what they say, just like LLMs, you can do the best and the worst with a tech
Yeah that's brutal to see what you worked on be used in a war...
it depends on your personal investment. i'd prolly be fine if it had a benefit for the security of "my people"
It's not documented yet, but I improved the bind library (https://github.com/jeremyfa/bind) to generate the whole stack of bindings from a C# class to the actual methods callable from Haxe automatically
So you run haxelib run bind cs SomeClass.cs and then the haxe equivalent becomes available
hmmm, this could be interesting if I think about reviving my augmented reality stuff
(that's very different from a C# extern when targeting cs with haxe, as here the Haxe is becoming C++)
cortex is pretty much where my old engine was that I used on ios for ARKIT things
Right now I'm mostly handling the case of bindings within Unity, but it wouldn't require much change to support other C# environments
The C# part is there, I just merged it to master today
But it's not documented at the moment
ah ok!
bind can plug Java/JNI, Objc/Swift
And now... C#
All to the Haxe/C++ target code
I think the next thing I might try is supporting plugging simple C/C++ headers directly
you bind a header and tadam you can call it from haxe
(automatic externs)
But C++ is wild, maybe C is more reasonable
COUGH COUGH I am looking for someone who would be smart enough to do automatic C externs for CGO
Yeah I don't know, probably a good idea to not just do plain parsing without context, didn't dig through a lot, just keeping the idea around and implementing things as I go
CGO = Go's way to call bind into C code.
Technically, it would also be possible to improve bind library to generate bindings for HL, but as I'm not using HL myself...
The way it works is two step. First step is to "parse" the input and create an interface object describing methods etc... and then the "bind" step that generate the glue code. So, plugging HL to Objc or Java/JNI would mean making a new "bind" step. But the "parse" step would stay identical
If bind could do HL it would be a great replacement for Ammer I would find, especially if it was well tested and/or stable.
The tradeof with bind is that it doesn't try to be exhaustive
I'd prolly be fine with a DSL describing the interface objects
It "officially" support some ways of describing interfaces, like not all Objective-C types or Java types are supported
But it support a subset that is good enough so that you can make a "bind-facing" api in Java, C# or ObjC that bind will understand
So the tradeof is: you don't need some specific DSL. The DSL is the handled subset of the target language you handle
A fully working subset is better then an exhaustive not well supported version in my view.
So, plugging a super complicated C header is likely not going to work
But you can make a thin wrapper that can be plugged
(and of course bind could be improved over time to handle more cases out of the box)
but there always will be special cases
esp when you use it for a lot of different things
Hence you define a subset where those special cases that break it, are not included.
It's a challenge but a really useful software paradigmn, when corectness matters
lemme rephrase: required special cases 😄
the plugin methods are directly bound to Haxe with a single command using bind library
Ah now I fully agree with you
So it parses this signature:
public static void sendHttpRequest(final Map<String,Object> params, final Func1<Map<String,Object>,Void> done) {
ok, i see
This will provide the same haxe method in an Http class, when the java map becomes a key value object, and Func1 and actual haxe function. All the JNI mess is handled
From my perspective, handling C headers is harder on the parsing part, but easier on the plugging part (might be wrong thoug)
And maybe an optional DSL to make things clearer could still be useful, even if that's not needed by default. Just thinking out loud
I gotta look into this at some point. for the moment my brain is fucking drained.. @clear dome btw, what's your dayjob?
holy shit
As mentioned, I used to have a game studio for 6 years with a famous youtuber (the first in subscriber count in France for many years), then in 2022 we decided to stop (financial failure of last game being the main reason, living from games is hard anyway, even with your youtuber friend). Some former colleagues I used to work with before (for about 7 years, in digital publishing: making apps to read books, magazines and newspaper on the web (even in flash!), native android, ios etc...). Now those colleagues created a new company which is specialized in in-game advertising and I'm working with them again. I gotta say I'm not particularly fan of advertising, (wouldn't be sad if that economy collapses), but at least the idea is to have ads integrated within 3D scenes so that they don't block the play session and feel less "intrusive" compared to full-screen interstial ads. The technical stack of this is probably the most interesting part for me, and it's not that easy to find a job that ticks all the boxes regarding what you do everyday, the salary and this being in 4 days, allowing you to do other things next to it, even sell a game, and not being in a toxic environment in general. Among all the things I did, I'm definitely not complaining about my current situation
I might get into VR again
look at the size of these googles: https://www.youtube.com/watch?v=gbFU6KoEASU
How did they make this better.. and lighter?
Need a new wallpaper? https://optimum.store
Bigscreen 2: https://store.bigscreenvr.com/en-au/products/bigscreen-beyond-2
Video gear
Camera: https://geni.us/5YfMuy
Primary Lens: https://geni.us/pWnoPBr
https://www.instagram.com/only_optimum/
https://www.twitch.tv/optimum
https://twitter.com/Optimum...
The question is, does it prevent the fog on the glasses lol
super interesting! Im not judging or anything, just super curious. Some similiarities there as well, also worked in digital publishing at some point 🙂
sounds like a good place to be
The ironic part is that our tech is integrated in horrible mobile games that I would never dare to actually play, but those same games are order of magnitude more profitable thananything I did within my game studio 😆
it's so light, you can add an extra fan and you wouldnt notice 😄
yeah, mainstream markets suck, they are soo optimized
They only care about the metrics, I can tell you
yeah
But recently more high profile games are being cooked, that's interesting (and that partly runs with haxe, as mentioned before)
Cool thing is being able to see how other studios are working a bit, what kind of engine they use, as the tech I'm making has to work in a wide variety of environment and engines
Can't wait to play more with UE5
Less cool thing is having to ship things built with vs2012 because some games are old codebase lol
lol, yeah
Had to use an alternative version of lib pcre2 to make hxcpp build in that situation
market adoption of a product has 2 sides, like a coin 😄
That's also cool to be able to use Haxe. Wouldn't have used it if irrelevant
(Rust was a candidate too)
you should look at python
ono
😄
First idea was plain C++
But then using Haxe/HXCPP did really speed things up
User-facing API is in C++
Internals use haxe, but worst case scenario : if we want to get rid of Haxe, that only internal business, the user-facing API wouldn't change
turns out the whole thing is in production and runs fine
(had to patch hxcpp here and there, but that's not worse than having to make your whole stack in plain C++)
hahaha true
What I'm very curious about is running in consoles
Providing a third party lib for console developers
What could go wrong
But I'm not worried lol, I'll figure it out when that happens
on modern consoles you should be just fine
I just hope I can ship a dynamic library
if not, supply static ones
Yeah but static = risk of symbol collision
especially with HXCPP
Like, top level String type without any namespace 😦
oh, I remember
But no matter what happens, I'm sure we'll find a way
The whole stack is modifiable source code
sky-is-the-limit lol
yeah, hxcpp is not that complicated. there are people that replaced the whole GC to make things work for their use-case
I'm fine with default GC
me too
anyway, on ios, android, mac, windows
I ship dynamic library
so it avoids conflicts
On console, we'll see when I have dev kits
Oh and of course if I have dev kits I'll try to run CERAMIC on it
FOR THE FIRST TIME EVER
ofc u will 😄
exciting times lol
i remember my console dev days fondly, but I dont miss them
I heard you should use one computer for each dev kit
because each dev kit requires specific config on the machine and it's a mess to try to fit all on the same one
But maybe more recent kits are better
For now, Android stays the champion in most terrible dev stack (for me at least)
(old windows stuff close behind)
in one studio we had devkits per dev & workstation, very rarely they had to be moved. In another studio there was a single wii-devkit on wheels rotating between 6 devs
I guess it depends on what platform and version you end up with
😄
ps4/ps5/xbox likely
I checked out of profesional gamedev on the height of the seventh generation, so dunno a thing about those. But I assume they come with way less caveats/constraints these days
Probably. Also it's unlikely I'm going to fight with 5 other devs on the same dev kit as I'm basically only myself needing it
(very small tech team, but know each others for a long time and very efficient as a team, that's also a very good side of my current job)
yeah
makes all the difference
Talked so much that I didn't finish my multiline text lexing of Loreline
Gotta head to bed lol
Enjoy 👋
Actually, regarding static linking on consoles as mentioned earlier, given that I don’t expose any haxe/hxcpp type to user-facing API header, simply setting symbols visibility to private/hidden for all hxcpp stuff should work
thank you, claude is peak
Claude just helped me design an architecture for PyHLGen
maybe I need to ask claude re-architect my scene system
« You are absolutely right! »
hmmm, animating yoga layouts is actually pretty easy
yeah, it's ok. I cleanup my multi-screen obs setup. was thinking about streaming again
Never got to the idea of streaming when programming
I feel like what I'm doing is not engaging at all
Migh be disturbed by IRL events too
But at the same time, even we close to zero audience
Maybe that's a good way to get to work
yeah same, but the fact that what im doing is streamed keeps me focused on the task at hand.
Yeah that's what I just meant
Do you stream just here on discord?
twitch
I see
i should prob stream again soon
How do you set this up, you have 2 screens and you stream one?
i have 3 screens, i just stream my IDE though.
i put on some music
have a chat and music overlay set up
and it just kinda works
I see, so you explicitly stream the window of the ide, not a whole screen
I use a scene switcher plugin to follow my mouse around
yeah
i dont want to use incognito mode and be logged out everywhere
then toggle elements based on the screen
probably has the same effect like working in a library or at a coffee shop
yet i dont want to leak info about myself
Yeah that's a concern for me also
You can always switch to another software window if you, for instance, move to make some graphics
Maybe I could try streaming when making music
I move sensitive stuff on 2 of my other screens and only stream the 3 middle ones
what more info could we get from you? we already know you're a cat named jeremy
jeremy is a cat?
What
didn't know this was jeremy
profile pic joke
Ah yes, with my human pet, right
xD
XDDD
thread on casual no?
Yeah I guess
your cat got insta? if yes, connect her to my dog https://www.instagram.com/just_pauli_paul/
Yes https://www.instagram.com/jimjam_cats And they are 2
don't sound so defeated, there was a really populer wordle thread there for a while lol
You caught me 😄
maybe we need a "gatekept" topics forum in the offtopic section
A thread in casual like linux nerds would work though
yeah, just if its gonna become a thing it would be nice if they were easier to find
discord hides threads a bit
Yep, you'll figure it out 🐱
who will create the petition?
shall we make a topic forum in casual?
bet
probably won't allow anyone to make a topic because i think it would get a bit "busy"
but adding on request seems fine imo
#casual-chat message
allow politics in it with the requirement every post has to have a pet picture
lol, i think with politics it would be some opt in channel, hidden behind a command with a disclaimer 🤣
An actual trojan horse
only horse pictures then
xD
Nah we should probably stick with current rules, even me understand that
If people want to hang out and speak politics let's just do that elsewhere
There's enough places to talk politics in the world for now
Exactly
Plus there's actually a haxe politics server
REALLY
There is????!
???????
yeah, somewhere in the pins of #casual-chat
Haxe lore is deeper than you'd think
Very small server lol
rainbow haxe logo, of course
i think zeta said something like most of the people there have similar beliefs so it became a bit quiet
haha
I guess if you truly want to go talk to anyone from any political border, you go to public social networks, but do we want that. X is a mess now
i dunno, discord is the only social media i use now lol
I like it, reminds me of forums in the 2000s
Too bad it's all locked down into discord as an entity
but i only used twitter for educational material anyway so when i was on it, i never had issues with any thing
i have ways to deal with this, its just mostly the incentives aren't there
i'm not going put time into solutions when the problem isn't large enough
let join and fuck them up! There can only be one true Haxe Server, and that Server has a no-politics rule!
lmao
THIS WAS A JOKE! (for the server logs!)
How is anyone afraid of AI taking jobs...
jeez that's so crazy man
top left is tooooo real I feel like
i really have not messed around with AI
Is Haxe a white people thing? That's what you'd see from these pics too
it's biased by seeing good old UK hooligans mixed with a "dont show violence" counter-prompt
Definitely biased
i didn't know sam altman was a haxe hooligan (bottom right)
i just realised we somehow moved from #1297978372529586207 to #1234939525931728906
It's like two rooms in the same flat
yep 😄
people feel the fear when the person who decided some terrible HR software was a good choice is in the position to choose whether the AI is good enough replacement for your work
At the end, the financial ratio will decide this for us
It'll be fine. AI is not a living system that adapts. But we are https://nesslabs.com/chaos-surfing
that's bc you think there was control in the first place. instead it was just high barrier of entry
which has been obliterated
will check it out later, gotta clean this mess
turns out i also collect headset, never knew
can somebody type Ceramic or ceramic somewhere
ceramic
Oh it works
its case insensitive
lol
tyyy
danggg
just upgraded my laptop and reinstalled windows
i find it funny that you can NOT resize the taskbar by default
ok, finally got most of my shit working again. funny enough my gpus are way less busy after the reinstall
gaaah
im stuck in reinstall land
finally updated all firmware and turns out I can run my iris gpu in combo with the nvidia cards and the load is even better distributed
the rtx only kicks in when there are 3D loads while the iris and the titan composite for the screens. finally
#gamedev message
damn, it soo sad that Konami is such a fucked up shitshow.
If I had to rate game-engines, I would prolly give S-Tier to the FOX engine. That thing is still unreal
Hmm
Does konami use Fox engine for everything (atleast modern games)?
I'm interested in knowning what is running on their arcade cabs
Specifically beatmania, jubeat, sound voltex, gitadora and nostalgia
Can't find that much online about it
What are you refering to specifically?
the made some questionable business moves that caused some very cool technology being lost in the wind
You mean, moving to unreal engine?
fox was created by the studio that was run by kojima, once it got shuttered and kojima moved on all the work they did was locked down and now wasted
I hate it when cool stuff just gets lost
well yeah always happens with konami
@vapid mortar Hey, dumb hmecs question: What's the point of having pooled/generational entities given they are just integers?
only thing I see would be overflow-protection
but then it could just be int64 and we could accept an overflow if it happens
im thinking about persistence for save/load and multiplayer
@void condor given that a lot of the storage is in arrays, the idea is to keep them as compact as possible to minimize memory costs and traversal costs
you can use unique keys, but then you have to always use maps or similar data structures
traversal can get more expensive unless you do some clever things
I wanted to keep indexing as cheap as possible for components that were accessed all the time, and ideally, keep them sequential
Eventually, I want to figure out how to also sort them without breaking the references, but it hasn't been a high priority.
I guess just sorting the views is just as good
btw, if I ever don't answer a questions it's because I couldn't find the @ . Feel free to @ me a couple times if I don't respond.
okay yeah
im separating it
beacon will be the tooling + externs
lumina will be the framework
Yep, that makes sense. I think im just overthinking stuff
A few metrics about this particular scene: 2331 draw calls, 623 textures and 73 render targets were used.
73 render targets...
😄
the part about the shadow maps is interesting. the fox engine doesnt mess with complex shadow mapping algos:
Here we have 2 spot-lights in the hallway ceiling casting light downward, a 4k x 4k shadow map is generated for each.
this might just be for this particular scene, I wonder what they used for open terrain
In total there were 1331 draw calls, 132 textures and 50 render targets used.
2331 draw calls, why do I even care about the number of draw calls in Ceramic then
(but I know why: web & mobile are reacting very badly against draw calls. MGS not a mobile game anyway)
I guess drawcalls have way less implicit costs/statemangling in dx11+ & vulkan compared to GLES/WebGL
Likely
That said it's funny to target 1 draw call to render a whole 2D game frame 😄
you pay with cycles either way. The question is where you burn them 😛
https://apps.apple.com/fr/app/ta-mère-en-slip/id1327819012 (sure that's not the most outstanding graphics out there)
That game used to need about 10 draw calls to render
because it used multiple texture atlases and additive blending mixed with "normal" blending
But after I had to optimize the default render path of Ceramic when working on Make More Views (which is much more HD graphics heavy)
And rebuilt that game with newer Ceramic
and it ended up being 1 draw call
Because additive blending has become "free" when rendering premultiplied-alpha textures
And because I added multi-texture batching
These are of course mostly useful to have by default for a 2D render path
And because Ceramic is mostly "old school single-threaded rendering using GLES", reducing the draw calls saves a lot of CPU time
(whereas the logic to implement that free additive blending and multi-texture batching is not particularly expensive, just more complexity in the implementation)
not sure i understand that "free" additive blending when it comes to rendering UI. you still need to maintain the geometry buffers and can run into fillrate issues concerning alpha-blended shapes
from those screenshots i assume instanced quads + geometry from svgs, both low reasonable counts
Ceramic renders elements in the order of their depth. The "free" additive blending comes with the fact that you don't need to change the render blend mode to draw the element. When using blendFunc(one, one_minus_src_alpha, one, one_minus_src_alpha), you just need to change the alpha value of your color to make the blending look additive
so maintaining the buffers is likely not a bottleneck
No they are actual textures, this is not because that's the best way to render them, but because all the graphics were made by somebody else with Spine (it's all animated)
(The reference I used to implement that: https://amindforeverprogramming.blogspot.com/2013/07/why-alpha-premultiplied-colour-blending.html)
It's a rather specific use case, but not that uncommon in 2D, so it's nice to have it by default in Ceramic. Of course if you use a different blend mode it will need a different draw call
chatgpt literally just used the phrase "A client of mine (anecdotal, not data) had similar issue" in response to my question 🤣
cool! The comment section is even better xD
Ah right, very insightful comments indeed xD
xD
I prefer to use chatgpt for this
Ha ha
tend to hit up grok
xD
I am both amazed by the quality of the output, and horrified by the consequences of such tech in the society
We automate away the things that made life worth living, while I spend 4 hours writing emails because the AI can't be trusted with automating anything that would give more time to do the things that make life worth living.
I've just read today somewhere on the internet that you should be shitty at a craft in order to become good at it eventually. I agree with that and super good-looking AI-generated stuff shouldn't stop us
if ai existing makes you not interested in doing those kinds of things then maybe you were never actually interested in those kinds of things
I'm speaking of time and community, not interest
yeah but the automation of a thing doesn't negate the "life fulfilling" aspects to it
Making something leisurely like painting to express yourself available in 5 seconds so that we have more time to spend at the office
only if you see things like that, sure
is the opposite direction that we should have gone as a species
there's nothing stopping anyone from spending 10hrs on a painting
I worked 12 hours today and had 30 minutes and counting of leisure so it's hard for me to see it other ways 😄
if you wanna do it, you'll do it
30mins a day is a lot of time
you can get decent at an instrument on just 30mins a day over the period of a month or two.
it is seriously a good amount of time
That's a very quick conclusion. I'm not worried about myself not doing those things. That being said, AI is not the only thing that can stop you from learning a craft. Paradoxically, having "good/better taste"/"being more critical about craft" in general can be a big obstacle too. Like, you know what you are doing is shitty, according to your own perspective, which can discourage you to continue. This is usually getting worse as you grow older, as you likely care less about being bad at something when everything is new to you when younger
you're not engaging my point. It's not about what can be done in half an hour. it's that we fed all of human knnowledge and creativity into a blender along with zillions of jigawatts of power and we still only have half an hour.
it is, for sure, but, complaining about what we can't do is not useful or insightful. At every step in history, there has been a reason to not learn a craft, and at the current step of history, it has never been easier to learn a craft despite
AI should definitely do more our laundry and less our art, that's for sure
people were probably too scared of terminator robots to prioritise that route of development
I think it's also a matter of cost. 100% digital AI scales much better than robots
Robots need a lot of hardware, it's very expensive
yeah but art generators are basically free and worth little
They are definitely not free, there's just a lot of VC money being thrown at it
I see this comeback a lot but it always assumes that the person is saying "now I can't learn __ because somebody automated it". I really want to stress that that was nowhere in my statement.
you can run it on a personal computer now
what other point is there?
I want the plaguarism machine to, if it must exist, give me more free time. Not find ways to approximate what I'd do with more free time if I had it. Having an approximation of your free time waved in front of you while you spend most of your life trading labor for capital to survive feels like being kicked in the genitals
That's kind of the end point of AI tho
The whole point of AI, is to basically get rid of jobs
Unfortunately, only a fraction of the society is going to benefit from that
I think it will be about the same 😅
i think out of the most likely scenarios you're most likely to be right there shadow, nothing will change
If I had a guess of what the most stable outcome would be, it would be to have jobs become disconnected from any type of production
Advanced economies are already heavily service based, it seems that will just take on news forms of humans servicing each other, and the new backbone of (non service based) labour being changed out with machines
i don't really understand this at all, i trade labor for capital. I don't feel like getting kicked in the genitals
Free time cannot be "generated", its a function of skill
Daz can get probably 20x more work done than me in 30mins
It isn't because daz is more gifted, has more resources or whatever. Its just he's more skilled
AI can be used for bad (to continue being lazy) or for good (to accelerate your skill development)
Nows the time I think, it will almost certainly get paywalled after enough time has passed
I'm thinking that also on this point also we will struggle to find common ground, since you are talking in terms of efficiency as a skill set ("you don't need more free time, just use it better") while I'm thinking in terms of where it matters to be efficient, with the end goal of more of life being not a game of productivity and efficiency.
It's hard for me to explain, but a whimsical "what if I painted some cats playing with the haxe logo", and the AI responds "it'd look something like this. get back in the mines", is the way especially the generated images speak to me.
Well without clearly conveying your point I obviously can't respond appropriately 😄
That's part of my point when I'm saying AI will benefit only a fraction of the society. It will increase concentration of wealth (once again) on the people who possess the means of production, which will include all the AI tech as well, and push away everybody else from that. There is no guarantee that a new layer of production-less services can be built on top of that, that will fulfil all the jobs lost from the production-based job market
u flatter me but I aint no wunderkind
"it'd look something like this. get back in the mines"
If that's how you see it that's exactly what it is
What's the quote
skills are like knifes, you gotta sharpen them
If you think you can't do it, you're right. If you think you can, you're also right
We could have hoped that reduction of labour to produce the same wealth would mean less work for everyone. But that's definitely not going to happen this way imo
anybody noticed how all these AI success stories are never random normies using AI for the first time? It's always a person with at least 10k hours experience in some area that leverages the new tools
And to add more to @neat smelt about being concerned with AI taking over creative skills, as an individual, if I do need to make a living wage, I'd rather have that by making something creative, than making something physically or mentally exhausting and repetitive/uninteresting
in the creative domain, there's no limits
What I mean is that people are paying their rent and their food by making art/music/etc... and AI is litterally taking this away from them
Of course if you don't need to earn a salary, sky is the limit
I still don't know where I stand on that side of things
like, what would be different if another human was taking the work. It would be the same outcome but it wouldn't be "bad"
In other words, I think as a society we should value and reward this creative work
And embracing AI is doing the opposite of that
like here is one thing I dont get: People use AI for emails... why though? The most time I spend with writing emails is just as long as I would prompt some AI for an answer. And in cases where I spent like an hour on an email it's usually I have to prepare a lot of data and review my wording vs. my intention. No way an AI could help me there since i would have to prompt it along my heavy thoughts
lack of communication skill so they offload the task to something they consider more competent
"uses AI to write an email" -> SENT -> "uses AI to make a summary of email"
If the AI can be trusted with it it's probably not a very important email 😄
it messes up an invoice request or work order and you have to tell the supervisor that Grok is to blame
The difference then would be that the net number of creative jobs available to humans remains the same.
the other day I generated some documentation in our internal wiki. Afterwards reading all of it felt like shit bc the notes I used for the prompts where really the essence and much more userfriendly to follow than to read miles of prose the AI spit out
AI is very wordy isn't it
Unless you tell it to shut up
I also had some great experiences
Like, most responses are 80% more words than I need
Copilot must be told every few turns "concise. data only. no anecdotes or greetings." when I trialled it
you can add a copilot-instructions file somewhere in your project
I don't use it in an IDE
Well, I don't use it at all, but I was not trialling it in an ide.
Just the web client
like I had to deal with a lot of filebased data on a server the other day and I used AI to generate a python script to sort and isolate specific chunks for further processing. would have taken me like 20 mins to iterate and test this, with AI it was literally 1 min and 2-3 mins tweaking&testing before i was able to extract that data successfully
Yes there are a ton of uses that can help save time
yeah 😅 It does save me time here and there, but I'm evidently not got enough responsibilities to actually use it to save time with
I've tried to think of how I could use it around here but I haven't had too many majorly useful ideas
I saved A TON of time when making Loreline thanks to Claude, yet didn't make any compromise to the quality of the code
It really is powerful when used for the right task
currently my main use of ai is if I get stuck on something, I'll ask it and see if it can help lower the hammering time
I also think the more a user is skilled, the more they can take advantage of such tool
definitely
Being a junior developer in current job market isn't going to be easy
I can see workplaces indirectly encouraging skill stagnation by prompting reliance on the ai
now you're a junior developer for 2 years instead of 6 months because ai solved all your learning problems
they will develop new skills. they will learn to guide more AIs at the same time
they will become AI engineering managers
who knows, I think there's already skill stagnation prior to ai anyway. People have just gotten a bit lazy and probably created the gap for something to come along
If you manage to get the job, you can learn quickly thanks to AI tools too and become "less junior", but apparently nowadays a lot of Junior devs are struggling to find their first job at all
(at least in France)
well, that's probably true but I wonder if that's just company over relying on ai
not only
There have been a lot of new learning programs to become developer in France
even as a job reconversion
This increase the pool of junior devs, compared to a decade ago
But at the same time, companies seem to struggle finding more senior profiles
ah, so it's a perfect storm kind of thing
I guess, and AI is probably making the problem worse, even if it was there already
hmm, synchronous eventhandling can cascade quite a bit. this is a frame where the player dies and the game cleans up the scene and switches to the gameOver screen
damn that's cool
also just realized that I render and then update the gamelogic, making stuff lag a frame behind
Ah right, I guess something could be done to make this more « flat » while staying in the same frame
Actually I did something like that in Loreline recently, because the way it works is:
- Loreline runtime execute a script AST
- It calls user-defined methods to display text and choices and waits for "end of text display" or "choice input" until a callback is called.
- All is fine if the callbacks are called asynchronously, but if called synchronously, it would have increased the call stack after each text or choice to display (until stack overflow)
- To overcome this, the runtime has a system that flattens the calls when it detects the callbacks have been called synchronously. That way you can have 10000 message handlers that call the callback synchronous without increasing the stack
Kinda messy, just like it should be!
My main rationale is that the current nwjs version doesn't run on my Chromebook, plus it is just overall clumsy. Electron gives us some neat stuff that ...
thingie
oh don't fret the diff count too much, i just moved some stuff around
pretty nice! will check it out on the weekend 😄
CastleDB on electron seems like a nice upgrade
awwwwwww
dazDog 😮
dazLogo did a dazReaction??
dazCute
dazIst very true
dazNotFunnyAnymore
dazNotUpForYouToDecide
dazConversationIsVeryFunnyYa
breaks the chain
dazIsntVeryCamelCaseOfYou
cobalt dropped the ball 😔
dazillion possibilities of daz*
one in a dazillion!
just installed quake3 and played some deathmatch rounds with bots... holy crap that shit is still good!!
loading into a game is literally a few ms
and it runs with a dazillions frames per sec
I remember I was playing it on dreamcast with a gamepad because I didn't have a PC
tf
might be a regression in haxe5
there isn't much different with hx5 yet
there are breaking changes but pretty minor quick fixes
waaaahhh, first time looking at hxWidgets, amazing shit!
man every time I see this kind of UI it throws me off
it was fine in win10
but in win11 they threw a skin over some components
changed the colours a bit
and called it a day
especially input
yep, but we dont care about the theme or colors in this case. I need something quick to prototype
I dont want to build the first version of the editor using cortex' ui, since Im constantly distracted by missing bolts and blobs. Instead im gonna use wxwidgets and throw everything together exploratively till I have my workflows and then I can look at what I actually really need build to UI wise (if anything)
I wanna prototype the games
Haxeui just adds more layers I dont need. wxwidgets is enough
or what about Qt? (iirc there are bindings for it)
I guess yeah
Yeah same, « native desktop UI » sometimes just feels old
removing design as a parameter helps narrow the focus down to the task
defo agree with that, wxWidgets just seems like an interesting choice to me
personally i'd use haxeui hxwidgets for such a job, simply for the easy xml but if you're proficient or there's good tooling for the native version then I could understand it
using wxwidgets directly is a shortcut since I can generate code & stuff pretty quickly bc most LLMs learnt more about wxwidgets than haxeui 😛
HAaaaaaa
fair point
wonder if cppia + hxwidgets is a thing
what about imgui though?
there you have the reduction of "the extra layer"
and i think it could potentially be even easier to prototype with
nono, I want os level controls & functionality for the ui
I mean I guess if that is a requirement hxWidgets makes sense
why that requirement out of curiousity?
llm knows more about native wxwidgets than haxe widgets
true, but it maps almost 1:1
I dont wanna mess with anything related to the UI except bolting shit together for the workflows
doesnt work out of the box
basically prototype stage
isnt imgui ideal for this though?
you could just have an editor class and call functions in your engine to get stuff and display it
a list of entities could be as simple as
(pseudo-code)
for (entity in CortexGame.getEntities()) {
ImGui.text(entity.name, 0);
}
to me that seems easier than having to do work to construct layouts or keep updating the UI in an acceptable way
the fact its immediate-mode makes it very hackable and easy to get stuff to just "function" with little work
with some nice features like docking too
yeah, I gotta start moving faster again. the past weeks I have been stuck bc I keep losing my focus with shit like "oh, I should prolly improve the textfield, oh the sliders need a css overhaul, oh I should prolly add embedded pngs to the css for the checkboxes, oh I need to write a menu for dropdown, oh.... " STOP THAT SHIT; my game doesnt even need all that complex UI and the game has a higher prio than the editor.
LOL
Im testing something, it should be native, no special rendering and as stable as a fucking rock
We got another nerdsniper in the haxe discord xD
I literally have been ignoring my textfields and sliders in my project
for the last few days
every now and then, my eyes will just pan to the horribly styled textbox
it only bothers when its getting in the way
true
but i've been in that place you're at for a while, and i'm in the place you're at now
but I get constantly reminded
Wouldn't less work be better for testing though? ImGui is very easy to embed and is also very well battle-tested
so my solution is using a completely different tech.
found my ways to just say fuck it, we ignoring backend stuff 🤣
Idk I'm just a GenZ programmer what can I say aye
I really don't see the appeal in hxWidgets though especially not when testing and quick and dirty code is the goal
+1 on thus
@grizzled laurel from the bottom of my heart, you know I love you, but seriously, fuck imgui, im not gonna touch a gpu renderer UI for what im doing, so please stop before I start screaming at you
Well I love dazDog and wouldn't wish for it to be jolted awake in the case it's sleeping, so I'll "shut the fuck up" xD
dazDog approves
I mean dazDog, is it a he or she
I said "he's" but I genuinely don't know lol
It does look very peaceful though
dazTwoCats when using retained UI for quick editor stuff
Also have you guys tried out the new AI image tech?
That stuff is genuinely insane
I asked it to draw me as a cartoon, it genuinely feels like there is intent behind stuff like shading now
His name is Paul, 11 years old, best buddy you can wish for
Awww
Yeah I saw that
Same with the insta for Jeremy's cats
What kind of personality does he have?
Is he the playful kind
Or the kind that's wants to cuddle you all day
wait you aren't a hamster???
i request a refund
A refund on what
uhh
I should be getting a refund here
he is very alpha, opinionated and stubborn. he either likes you or not, also you need to keep him in check
I donated to the shrine of logo
but once he respects you, he'll do anything
lol forgot about that
He should be the manager of a tech company
😛
I also fed the AI my quite outdated linked profile pfp
In reality I was a recruiter on linkedin all along
You've been fooled!!
xD
You know I believe more companies should introduce a department pet
he is my VP of engineering
Perhaps call it the debugging pet (kindof like debugging ducks) but then like talk and pet a dog or cat to debug your code
Instead of talking to a lame rubber duck
Talk to something that may or may not talk back aye (even if it's incomprehensible)
tamagochi all over again
Trueee
But like real
And it helps you debug
Wait debugging tamagochi...
That would sell really well
i wonder how "in" mobile pets are these days
I know they sprung up at some point
but maybe there's a lil gap there
Prob not much (defo not in Western countries)
Because shite platforms like tiktok gives too much dopamine meaning stuff like that can't compete
They'd have to add rgb and brain numbing content to tamagochi to make them sell
nah man, if they've been "ignored" for a while, its probably a good time to pounce on the idea
Would that even be a tamagochi any more?
Maybe
But it'd need to have modern-is features
Ideas like Nintendo Street pass could be implemented to make it cooler
Oled screen ofc
and can't be uncanny valley
anyone remember eyepet?
Never known about it and judging by the image I don't wanna know
LOL
wtf
Logo you're American right?
yeah
Because I'm assuming that is an American toy
it was a game for the ps3
Look only Americans can cook something up that looks that cursed
by a british studio called... London Studio
No hate intended
But it is how it is
Ps3
Oh man
I was an xbox 360 kid
Loved that console
it was a free game for the ps move iirc
featured is the worst controller design of all time
there's also likely a gap for "just good games"
The square button layout arrgh
Yeah
this did really well
https://store.steampowered.com/app/1424330/Wobbledogs/
Wobbledogs is a 3D pet simulation where you raise your own personal hive of rapidly mutating dogs, physically simulated all the way down to their guts! It’s a casual and chill sandbox experience for players of all ages who want to care for their very own virtual pets in a surprising, unique, and stress-free environment. The dogs start out re…
$19.99
10601
79
they can get really messed up
baha
C++ Library to use Particle Effects from the New version of TimelineFX Editor - peterigz/timelinefxlib
Cant wait to give this a try, timelinefx is such an awesome effect editor
I have it bookmarked from years ago, just never had a game to warrant looking into it further
I used it for several flashgames back in the day and have been begging for a 3D version for ages. Super stoked that it's happening
well dazCoolAndAll but how does one not go insane from attempting to learn graphics programming
You see, don't attempt to do it all at once. Allow yourself to fail.
I've tried 3 times before I actually gained a deeper understanding how things worked
And that is a really important foundation.
The sheer scale of it
There is so much to graphics programming
And idk where to look for rendering optimization techniques
I know abt stuff like batching, instancing, and not-drawing-triangles-that-are-off-screen (i forgot what it was called)
I still lack the knowledge and experience to know all techniques there are for doing many things.
But I do at this point know some basic techniques and I have a decent knowledge about how things work on a foundational level.
well that's the problem then, why are you looking at optimisations when you don't even know how the api stuff works
you're jumping ahead
Don't do premature optimization
see what works and you can work on other techniques later to optimize certain parts of your renderer (if it becomes a problem)



