YouTube keeps recommending SpacetimeDB to me even though I have never made a videogame. Now I'm seeing it on HN too.
Are there any game devs here who can comment on how useful this actually is? The pitch seems pretty cool but I can't tell how much of it is just good marketing.
Not a developer affiliated with the project. But here's my take:
The problem is that the proof is in the pudding. Some of their claims feel outlandish. Imagine writing your game with SQL queries. Rather than crafting update packets that only contain relevant game info you just do `select * from players where abs(player.x - current_player.x) + abs(player.y - current_player.y) + abs(player.z - current_player.z) < 100^2` and it just updates automagically 60 times a second.
Running physics sim in SQL stored functions sounds insane but that's exactly what they're advertising. They say it's due to in-mem DB optimization but I'm not sure how they horizontally scale that (they allow you to run standalone but any scaling requires their cloud product)
It's in closed alpha so there's a limited number of players but I am curious to hear more about how BitCraft is scaling out. Is it a single, seamless world for all players? How much scaling is currently required? How is data replicated when horizontal scaling comes into the picture?
Given the name I was also expecting some clever SQL-like data structure/query for fast spatial lookup but I digress.
Game physics involves a lot of things (solvers, collision detection, numerical stability, etc.).
I am skeptical of their claims of being able to run physics in what I understand are stored procedures for their database.
I looked at their docs for physics https://spacetimedb.com/docs/unity/part-4
where they demonstrate the simplest form of collision check (sphere overlap). I fail to see how that is an improvement or speedup over existing methods. Some quotes:
"This may not be the most efficient way to do collision checking (building a quad tree or doing spatial hashing might be better), but SpacetimeDB is very fast so for this number of entities it'll be a breeze for SpacetimeDB."
>> Nothing is quantified with numbers.
"For every circle, we look at all other entities."
>> This is the most inefficient N^2 way you could do collision detection.
And not to mention networked physics is a whole additional layer of complexity, where you have to use some form of prediction techniques and very likely end up changing your core physics code to accommodate this.
All of this suggests to me much thought has not been put into the claim "you can also do physics with it!" and its implications. Perhaps it is enough for extremely simple physics, as demonstrated with their demo game. If the author is reading this, I suggest spending some time understanding what this claim implies and qualifying it better.
However I must mention that I applaud their courage to try something so outlandish. If you truly believe your claims are possible, I encourage you to keep working on it.
But I'll be convinced when extraordinary evidence backs extraordinary claims.
Source: I work on a commercial game physics engine and related netcode.
I believe they've done more complicated physics in their game bitcraft, but that code hasn't been released yet (no idea if it will be.) But there has been work in the discord recently and 2 member have individually implemented rapier[1] at this point. I can't say anything about the related netcode however as I don't know how much they've focused on it. At the very least work has been done.
The unity tutorial is just to get up and running as fast as possible in the simplest way possible, not to make the perfect decisions for a networked physics based game.
I however do think they should provide an actual reference to prove the claim,
Nice, that's very interesting and a commendable achievement!
If that is the case I would really like to see some internals of how the physics engine has been implemented in their pattern. I'm not asking for the rapier port source code, but it is hard to think in terms of a new advertised programming paradgim when there's no working examples.
For example I am very curious to see how a constraint solve is implemented in a SQL-like fashion. You would need various math operations and an efficient matrix representation off the top of my head, and I can't think of how that maps to a SQL-like interface.
On of the people who implemented it is from unity, from my understanding he can't actually release the code. For the other guy, this was his last comment on the matter
> Sure, my plan was to write a readme and share the code with my colleagues anyway. It should be ready by the weekend because I'm in the middle of a refactoring right now.
> The gist of it: each time a System adds a "movable" component to a StDb table, it also adds the rigid body and collider in Rapier. Then, a scheduled reducers updates the simulation then fetches resulting positions and velocities from Rapier and updates the "movable" components. I have no client prediction at this time.
So my understanding is rapier is just compiled into the wasm and run like normal while the results are stored into the table for clients to receive via subscription. I won't say it's a perfect solution, but efforts are being made by the small community to see how we can really push the db.
Note: while it's all very procedure like you can just store data normally in the rust side, you just have to push the information to the table for it to be query-able.
I see, that's extremely interesting. As mentioned earlier, I do respect efforts to push on something new, especially if it's against conventional wisdom and would be interested in poking around with an open mind. I'll join the Discord.
Thank you!
Then again looking at the FAQ for Bitcraft, we have this:
"What engine does BitCraft use?
The BitCraft client is developed using the Unity game engine. For the server, we have developed a sophisticated distributed system called SpacetimeDB, which hosts the entirety of BitCraft’s server architecture. "
Correct me if wrong, but this suggests the physics engine is Unity's, and not something implemented in SpacetimeDB.
I didn't truly know and I didn't want to just throw out a guess so I just asked. They mentioned they have a basic hand rolled system for their physics. So not using unity physics.
I guess making the claim that is was more complicated might have not been right for me to do, but it would be hard to believe that they did the degenerate case of O(n^2).
Speculation: I have only written a physics engine once during college (so please excuse my ignorance), but I think for a basic mmo a simple chunk and AABB approach would work. It would be easy to query for the surrounding chunks and just run the collision on that subset. I know one of the team members was working on porting a minecraft 1.7.3 server at one point, but I don't know if they got up to the point of moving collision off of the server and into the database.
Thank you for asking and forwarding the answer. I appreciate your comments.
Speculation as well: Looking at the trailer for Bitcraft, it looks like they have articulated characters and cloth physics. This raises more questions for me.
I am contributing to a project by a SpacetimeDB community member to get Rapier running within Unity. Instead of trying to get Unity PhyX running on the server, we just use a deterministic Rust physics engine on both client and server. It is a mostly complete drop in replacement as it stands and is deterministic (Unity's PhyX is not). We'll keep improving it!
We rely on the guarantees of Rapier's determinism. I haven't done extensive testing on the determinism claim yet, but it has held up so far.
https://rapier.rs/
It's a cool technology, but the hard thing about building an MMO is architecture, not just making the server faster / wider. If you want a good overview of the types of problems to be solved in scaling an MMO, check out https://youtu.be/-c4t3Y5l5jY.
There is a pretty big underlap between these problems, and the ones solved by SpacetimeDB. So much so that if you built a system using both SpacetimeDB and another, high-level tool to manage chunking, transitioning, server provisioning etc. it would obviously be the latter system that was more indispensable to scaling your MMO.
As a game dev I don't really see what problems this solves. Most of the work you would have in games is related to simulation - something this doesn't help you with at all. It can actually make simulation more difficult because all your state would be decoupled from a game engine (which is built for simulation). So you'd end up having game servers running the simulation syncing to SpacetimeDB... or of course you try to build the simulation inside SpacetimeDB with no help from it.
1. Server <> client sync is handled for you
2. Server <> database sync is eliminated
3. Deployment is super easy because you just upload your wasm module to the database, and SpacetimeDB schedules it (similar to Kubernetes honestly, but at a different level of the stack)
The first two problems are nothing, IMO. Server <> client sync is handled by your engine or, if you need to roll your own, it's not that hard. Syncing to a database is more tedious than difficult.
Getting rid of a game engine on the server side adds a lot more work. I don't think it would even balance out.
Having implemented the first two for a web game i absolutely guarantee you those are far from "nothing". It takes considerable diagraming & handling to coordinate FE + API on state changes.
If they in fact can solve that (haven't checked myself), it would be a massive time saving for development.
Maybe I'm just used to it having implemented it several times. But my point is that they require a relatively fixed amount of effort. Once it's implemented you shouldn't need to touch it again.
Are there any game devs here who can comment on how useful this actually is? The pitch seems pretty cool but I can't tell how much of it is just good marketing.