#The further you get from the void the placement gets more of centered
36 messages · Page 1 of 1 (latest)
floating point precision
i can already tell this will be a wont fix
They've said they are looking at possible fixes for it, it's more a question of how feasible it is to implement without negatively impacting performance and breaking saves from my understanding
I could be wrong though
ok thanks
They could look at how Minecraft solved this issue
Also, point towards the case for Fixed Point types
Minecraft solved it with a world border
wouldn't floating point issues be noticiable near the current world border?
literally just open up minecraft bedrock
Can't you use a long instead of a double ? (or even a int 128 ?)
long is an integer; it doesn't have decimal points
But does the game need decimal points ?
yes, rendering reasons
decimals are used a lot more than you'd think
I know
I just hoped that it was a choice that they made and not a limit of their engine
tho
can't they divide the size of everything by 10 ? while it doesn't fix it, it should at least help a bit
I guess, though there might be a lot more... technical reasons why it'd be harder than you'd think to do that
The rendering only needs to take small(ish) region into account though. Could use a big int of some sort, and the double as an offset from that.
well yeah
The rendering falls apart pretty completely if you go out even 100x100 large squares anyway
but the current implementation uses floats
Sure, but that could be fixed
No they fixed it with chunk generation and de-generation
Yeah, there is effectively no limit in Minecraft anymore.
Like (effectively) limitless numbers aren't hard
Eventually you become bounded by the storage for the size. Once you exceed like 2^(2^64) *8 things get complicated (and things get slow long before that).
There are number systems that are truly limitless, but most are complicated and computationally expensive.