dear community. please consider also joining our discord room!
... if we could only make this a sticky post ... ha only discord can do that !!!!
@dsyme is happily getting into aardvark after seeing a talk at FableConf
great to hear 😀
I'm excited to see that talk. I'm new as of like a month or two, but I'm enjoying it so far!
I apologize for the newbie question, I’m sure this is in the docs... but can aardvark run on android. The lack of samples would suggest no.
hey, we did expirments with Android some years ago, but we never really made it happen.
So the answer to your question is no, but due to the linux support and the OpenTK guys (having android support) it shouldn't be too hard to implement. the main problem is that the rendering backend has lots of native depenencies (DevIL, etc.) that would need to be built for android. last but not least we would have to implement some kind of UI Integration...
and btw: i don't think it's in the docs 😀
@krauthaufen thanks for the fast response
are any of the experiments still around?
sadly i don't think so, they were based on a old backend implementation
Hello, Does anyone know how the performance of aardvark would compare to a game engine such as Stride or Unity?
it really depends on what you do and what workload you are expecting, how the scene/data is structured and how much of the scene is dynamic. for quite some time aardvark rendering focuses on semi-static scenes - i.e. large scenes which cannot be optimized easily but individual objects do not change that frequently. whenever hardware there is graphics driver support for a particular scene "structure"/kind (e.g. many repeated objects) of course this is different. there is a now old yet mostly still valid paper on comparing also to other engines: https://www.vrvis.at/publications/pdfs/PB-VRVis-2015-015.pdf
Thank you for the info
Hi, why Adaptive is not async and maps are not parallel by default?
Async would help if you wire cvalue directly to the UI input, and parallel map calculation done right way would increase performance.
I view Data.Adaptive same way as Ocaml Incremental library, correct me if that wrong perspective.
fsharp.data.adaptive is similar to ocaml incremental but has different semantics. on-demand vs strict change propagatiaion afaik. adaptive, async and parallel are different things - yes stuff could be combined but this makes semantics much more complex imo.
I didn't spot that difference, thanks.
There's no way to Adaptive value propagate change to consumer right now?
Other words consumer have to pool for values?
You can install callbacks, either a MarkingCallback that notifies you when the value got out-of-date or a normal Callback that would then act like a push notification with the new value
The way how I connect this adaptive system to an oldschool UI (WPF or Winforms) is to use marking callbacks to trigger an invalidation and only pull the values when the actual re-draw happens, so there is no overhead if parts are not visible and you always have the benefit that if there are a lot of simultaneous changes that this lazy evaluation avoid any redundant updates
Thanks for the hit! This was exactly what I assumed Data.Adaptive was doing.