Same can be accomplished by using xs_vec_equal and xs_vec_set. This functions are a bit easier to use than xs_* but I'm not sure the difference is significant enough.

XS is a file that has lots of data inside.
My functions are strictly Vector related and they support both float and numeric parameters.

As you can see @ https://github.com/ValveSoftware/source-sdk-2013/blob/master/sp/src/public/mathlib/vector.h#L564
Zero is a basic Vector function and in my opinion it should be inside vector.inc file.
XS file should contain larger functions, operations and stuff like that.

Also, XS restricts all functions to use float parameters.

Does pawn have any difference between set 0 or 0.0 to Float?

warning 213: tag mismatch via http://www.amxmodx.org/webcompiler.cgi

I mean in memory. Warning we can bypass and if there are no difference for what we need 2 funcs for int and float?

I don't understand what you mean. Show me an example so I can understand.

We know pawn has only cell type.

new Float:a = 0;
new Float:b = 0.0;

If in pawn VM there are no difference(example, all bytes are zero for first and second variant) we can use one func for int and float.

@Mistrick there is no difference, we can write function that accepts both _ and Float.

Basically we can just add ZERO_VECTOR and then use:

vec = ZERO_VECTOR;
xs_vec_equal(vec, ZERO_VECTOR);

If I add public stock const Float: ZERO_VECTOR [ 3 ] inside AmxConst.INC, it will act as Float only.
It's impossible to declare it as public stock const { _, Float }: ZERO_VECTOR [ 3 ].

Furthermore, I still don't want to use xs_* functions to check whether a Vector equals to zero.

So, do you still want me to add public stock const Float: ZERO_VECTOR [ 3 ] inside AmxConst.INC?

@ClaudiuHKS I think integer vectors should be deprecated.