RFC: Normalized form for standard TSDoc tags

Consider this example:

/**
 * Adds two numbers together.
 *
 * @remarks
 * This method is part of the {@link core-libary/Math | Math subsystem}.
 *
 * @param x - The first number to add
 * @param y - The second number to add
 * @returns The sum of `x` and `y`
 *
 * @beta
 */
function add(x: number, y: number): number;

Is the ordering of tags important? Would this be legal:

/**
 * @beta
 * @returns The sum of `x` and `y`
 * @param y - The second number to add
 * @param x - The first number to add
 * Adds two numbers together.
 *
 * @remarks
 * This method is part of the {@link core-libary/Math | Math subsystem}.
 */
function add(x: number, y: number): number;

What about this?

/**
 * @beta @returns The sum of `x` and `y`
 * @param y - The second number to add @param x - The first number to add
 * Adds two numbers together. @remarks This method is part of 
 * the {@link core-libary/Math | Math subsystem}.
 */
function add(x: number, y: number): number;

Some questions:

1. If it's not legal, what's the best way to specify which orders are allowable?
2. What should the parser do with ambiguous orderings?
3. Should the TSDoc library include an operation for normalizing an AST?

Start with the three TSDoc examples and the questions about tag ordering, ambiguous parsing, and AST normalization. Review the parser and AST behavior relevant to these examples, then define the allowable ordering and the expected handling of ambiguous input. Done means the RFC has a settled specification and a clear decision on whether normalization belongs in the library.