Introduction to WebAssembly Text

Previous   Next
Using a Language Runtime as a WebAssembly Host Environment Up Local Variables

4: WAT Datatypes

At the moment, there are only four WebAssembly numeric datatypes:

  32-bit 64-bit
Integer i32 i64
Floating point f32 f64

That’s it — just numbers…

No string datatype; no character datatype.

In fact, there isn’t even a Boolean type! 1

Other data types exist such as the Vector type v128, reference types and function types, but these are beyond the scope of this introduction and will not be discussed here.

Interpreting Integers

One very important point here concerns how you interpret integers.

A floating point number always carries a sign value, but when examining an integer, you are free to choose whether the value is interpreted as an unsigned sequence of bits, or as a two’s complement integer.

This means that when applied to integers, certain comparison instructions such as gt or lt must additionally state whether or not the most significant bit should be treated as the sign bit.


  1. Don’t be concerned at the lack of a specific Boolean datatype because this is, in fact, just syntactic sugar.
    In WAT, the outcome of a condition is stored simply as an i32 where zero means false, and any non-zero value means true. It’s that simple…