first comment from me already is why does LengthSquared need a new type? requiring INumber is enough
first comment from me already is why does LengthSquared need a new type? requiring INumber is enough
KId rename SquareRoot for consistency with existing symbols in the bcl
MSystem.Numerics uses Sqrt
Mwait
Mno system.numerics uses SquareRoot. huh
Mwhat do you mean by new type
M(x*x)+(y*y)+(z*z) is very likely to overflow with integer types
Mso yeah i could jsut leave that burden onto the user, but that is 90% likely a mistake
Msame as DistanceSquared
P- Given that we have both Point and Vector, surely the latter is assumed to be euclidean and therefore should have length stuff ergo being float only, not just any INumber?
- readonly structs all the way, especially now we have withers
- definitely prefer static methods on non-generic types, this was an explicit choice in Silk.NET.Maths so that you wouldn't need to repeat the generic unless you absolutely had to
is much cleaner thanthing = Vector2.Abs(thing)
. in most cases extension methods just work also, but you do raise a good point that sometimes you will have to repeat the generic but i'm okay with this.thing = Vector2<T>.Abs(thing) - not related to the proposal, but do BigInteger and Complex have the generic maths interfaces implemented? Silk.NET.Maths supports these today.
this is assumed if we're targeting 3.0. 3.0 will target the latest even-numbered SDK at time of release.#if NET8_0- happy with
instead ofSqrt
. generally prefer the names that we've used inSquareRoot
because we've already bikeshedded these.Silk.NET.Maths
I don't think Silk.NET.Maths ever did support Vector-as-Scalar as we had once thought of doing (i.e. allowingBecause the type of T has to be narrowed to IFloatingPointIeee754<T>, we can't actually implement the interface on Vector3<T>.
) so we can live with this. it's cursed anyway.Scalar.Add<Vector3<float>>(l, r)
no more work than we were prepared to do for 2.0Ideally, this should be vectorized. In practice, that's a lot of work.
yeah I guess?Extension 3: INumber<Vector3<T>> Implementation
if we're doing something like this it should be an extension method that allows the specification of the coordinate system, not a simple property.Extension 4: Directional static properties
yep fineExtension 5: Convenience methods
if someone's willing to do it then yeahExtension 6: IUtf8SpanFormattable (.NET 8)
should be explicitImplicit cast to Vector2<T>, drop the Z component- generally make sure you have API parity with Silk.NET.Maths. Silk.NET.Maths is the one area where we don't actually need to do any surface API changes (generic maths notwithstanding) because it's pretty solid, and changing anything here would do weird things to user expectations.
documentation/proposals- The motivation with all of this was to allow Vector3<T> to use int for storage while still returning float for methods like Length(), which is a rare but useful use-case. The extension methods for IFloatingPoint<T> allow you to not have to specify the return type
- ok
- so from options 1-4 which is your personal preference?
- yes for BigInteger, Complex only implements INumberBase, i can broaden the scope of T if desired
- noted
- noted
- noted
- i don't understand what you meant by this
- noted
- noted
- noted
- i already implemented it
- noted
- noted
K
I'll look once home.Only thing I want to point out is that Silk maths is the only lib were we didn't really pay attention in the working group tbh. We've revised and been unhappy with many things so do also take everything there with a grain of salt. Overall it's solid of course.
M"no more work than we were prepared to do for 2.0" does that mean you're open to implementing it, or don't want to do it because it's not worth it?
Pyes open to implementing, vectorization is actually called out in the approved 3.0 SDP
Mwe can copy tanner's existing code for some operators, others will have to be done from scratch
Mmaybe could coordinate to share code with them?
KOnly from perkseys comments - do consider thinking about separating maths and geometry. Graphic programmers often put those together, but a vector and a point are very different.
Also, my stance has always been that we should allow the user to do things even if we think it's not such a good idea, ie VectorN<sbyte>
Also, my stance has always been that we should allow the user to do things even if we think it's not such a good idea, ie VectorN<sbyte>
M
Pit's worth noting @tannergooding often attends working group meetings, and will probably want any discussions we have here to feed into whatever system.numerics does down the line
Pso code sharing can probably be discussed there as well. i mean it's all .NET Foundation code really, so reuse is trivial.
but it's nice if there's prior art to inform what questions need to be asked and answered for the .NET API Review
but it's nice if there's prior art to inform what questions need to be asked and answered for the .NET API Review
Mno one said i couldn't do this
M(for real, why isn't this public???)
Pi mean, isn't that trivial to implement
Pit's public in NumberFormatInfo, just not TChar
Mthat would require encoding, which allocates in a supposedly allocation-less API, so we wanna cache, so we wanna use a weak table...
Kif doing utf8spanformattable remember utf8.trywrite for efficiency and interpolation and also consider implementing the other formattable-related interfaces
PMbut the group separator format can be reset, so we need to make sure to check if it's up to date when we pull from the weak table, and also it needs to be thread-safe...
Mdid you read my code 
Mi am using Utf8.TryWrite
Khaven't had time but good
but good
KKinda sniped by 1. here. I always wanted to have the ability to have ops have the ability to have a storage precision and each op have a specified precision.
Iirc there's lots of gains to have.
Iirc there's lots of gains to have.
KI thought my being busy as hell was over and then nope heck me I guess
Mworst part is having to manually call TryWriteInterpolatedStringHandler because interpolated strings can't natively do what the API needs them to
Kwait really?
Moh yeah: and i have to allocate a
new string(format)Ie (wrong types for clarity)
VectorN.Dot<i8, f8>(a, b)
Pthis stance does fundamentally change the proposal, and as much as i have agreed in the past, given that we have strict rules codified in the BCL now in the form of the generic maths interfaces, it makes sense to follow them given they're the stewards of the ecosystem
M(could rewrite the interpolated string handler, but...)
KWhere a/b are VectorN<i8> and the output is f8
Kalso where is this code i dont see it in the proposal and dont see any other recent links to check
Mbottom of proposal
MGitHub
Contribute to uwx/SilkGenericVectorPoC development by creating an account on GitHub.
KIdk, where does the BCL say Length of an sbyte vector makes no sense?
Kthe method declaration is in the proposal not implementation - ah ok
Mthe proposal already provides that
Mi'll leave it up there, rest is up to the WG
Mdo i have to write the API specification for all like 8 types?
KSorry I haven't looked yet, but from what's said here I gather that it returns float always?
Mno, it returns whatever type you want