In 2018, the expected runtime FPS of a game was 60. In 2024, the expected runtime FPS of a game is 144. On my high-end gaming PC, the default playerscripts use 0.5 milliseconds of frame time. For some context, 144 FPS requires your game to run within 7 milliseconds.
The CoreScripts are indeed written in React, however what’s interesting is that when you open up the microprofiler, you will never see CoreScripts being a bottleneck.
I’m sick of this conversation which gets solved with one look at the goddamn microprofiler.
Hi, it’s still on the roadmap., targeted for later this year. You can find it under Avatar Creation and Marketplace. In the DevForum post, we just call out features that have been launched/added/or delayed. This feature is still on track.
Adding on to not leave it unsaid, the new Explorer is being written with absolutely brutal efficiency as a top priority.
The current Qt UI today renders out an item for every single instance in its view, plus every single instance you’ve ever expanded. The stack cost of a single item alone is 272 bytes, ignoring the cost of heap allocations for Qt event handling, Roblox event handling (a .Changed connection is made for every single instance ever viewed, individually), icons, etc.
The new Luau explorer, by contrast, uses a virtualized list, meaning that it only renders what is actually in view. It will also, for the most part, only connect to relevant signals on instances currently in your view. Who cares if the name of an instance that you’re not looking at changes? Who cares if something that you’re not viewed is selected?
Through tons of optimizations like this, we’ve measured that the Luau Explorer is significantly faster for huge data models compared to the current Qt one. We’ve even measured Explorer’s efficiency on horrific machines worse than what basically anyone in reality is using, and it’s great!
Could this have all been done in Qt? Yes, but it’s unclear whether or not that would actually have made a noticeable difference (we understand the precise performance implications of Qt much less than we do our own engine), and yet it’s demonstrably true that development efficiency is sky high with Luau compared to Qt and C++. It even allows us to open up the Explorer for use in experience, whether it be part of the console or as a dev module (not in 2024, but it’s been a core part of the implementation strategy). It’s simply the unambiguous right tool for this job.
Cannot find anything related to group community roles. I am looking forward the ability to set role permissions per-place but I do not see these in the roadmap. Is it intentional?
While this is a cool feature please make sure it can be opted out of: some people may not wish to have physicaly simulated characters in their games and prefer traditional animation of characters.
This will be an option in starter player when it releases, currently there’s a fast flag for a beta feature but it doesn’t change anything and animation constraint isn’t documented
Rest assured, we are indeed still planning on SA! We will share more when ready, and we will get on the Creator Roadmap when we’ve a bit more confidence in dates.
It’s impossible to Ctrl-F without first expanding each category (the names of which are not conducive, IMO, to the content inside them), which I’ve found particularly annoying since this page was introduced.
A more eye catching way of marking features as live would be nice. Maybe push them above or below all the pending changes and place the (Live) link at the start.
Also, the images in the dropdowns push the entire list off screen if the category you want to open is in the middle of the screen.
