Why typed array constructors require offset to be multiple of underlying type size? [duplicate]

Question

In typed array specification there is a constructor that allows to take an existing ArrayBuffer and treat is as another array type. It is interesting that offset parameter must be a multiple of the underlying type of the constructed array. What was the reason for this limitation?

For background - I am trying to encode a binary buffer to be sent over WebSocket. The buffer contains various variables of different size. Originally I wanted to use the following code:

var buffer = new ArrayBuffer(10);
new Uint32Array(buffer, 0, 1)[0] = 234;
new Int16Array(buffer, 4, 1)[0] = -23;
new Uint32Array(buffer, 6, 1)[0] = 6000;  // Exception is raised here, 
                                          // because 6 is not multiple of 4

To make this work I would need to rewrite last line as following:

var tempArray = new Uint32Array(1);         // Create 32-bit array
tempArray[0] = 6000;                        // Write 32-bit number
var u8Src = new Uint8Array(tempArray, 0);   // Create 8-bit view of the array
var u8Dest = new Uint8Array(buffer, 6, 4);  // Create 8-bit view of the buffer
u8Dest.set(u8Src);                          // Copy bytes one by one

This is too much code for such a simple operation.

Answer 1

Unaligned memory access doesn't work on all platforms:

Why is creating a Float32Array with an offset that isn't a multiple of the element size not allowed? http://en.wikipedia.org/wiki/Data_structure_alignment

Use DataView object to access non-aligned data:

https://developer.mozilla.org/en-US/docs/JavaScript/Typed_arrays/DataView