docs(mutators): add mutant operator name proposal

[nicojs]

A proposal for mutant operator names. They are loosely based on the naming scheme found in infection PHP and pitest.

The goal of each mutant operator name is to have a conclusive name that describes the mutation. I used the following naming scheme:

*NodeIdentification*[To*Replacement*]*Mutation*

• NodeIdentification: Name the AST node being mutated.
• Mutation: Name the mutation being done:
  1. "Removal" when the node was removed.
  2. "Negation" when a node is replaced by its opposite ¹
  3. "Boundary" when a node is replaced by a similar node to test a boundary condition (like < to <=)
  4. "Replacement" when a node is replaced by something other than the opposite or boundary.
• Replacement (only when the mutation was "Replacement"), name the node being replaced to ²

¹ I choose to interpret the "opposite" loosely here. When in doubt, I decided to use it. For example: "AdditionAssignmentNegation" for += to -= and "AllMethodCallNegation" for All to Any.
² When naming the replacement node, I removed "expression", "operator" and "method call" for brevity. For example: "SingleOrDefaultMethodCallToFirstOrDefaultReplacement" and "NullishCoalescingOperatorToLogicalAndReplacement".

When identifying AST nodes, I did lean towards the JavaScript node types. Please verify if this is OK for C# and Scala, if not, please correct me. I used the following sources:

• https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference
• https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators
• https://sharplab.io/
• AST-explorer.net

We also have mutations that replace a part of a node, instead of the node itself. I had to invent logical names for them. They are summed up here:

• "FilledStringLiteralToEmptyReplacement" and friends
• "ObjectLiteralPropertiesRemoval"
• "ArrayConstructorItemsRemoval"
• "OptionalMemberExpressionOptionalRemoval"
• "OptionalComputedMemberExpressionOptionalRemoval"
• "OptionalCallExpressionOptionalRemoval"

I didn't do all of them, because I first want feedback on these names.

[nicojs]

Hmm I probably went a bit too far here. Since:

forall is not the opposite of exists
drop is not the opposite of take

🤷‍♀️

[psfinaki]

"Nullish" is not a concept in the .NET world, and I wouldn't know what to expect from this. Although I don't have a better idea, maybe only something even more verbose like "NullNullishCoalescingOperatorToLogicalAndReplacement".

(just passing by here - actually I think all this PR is a good idea)

[rouke-broersma]

I don't think Nullish makes sense anyway. The end result is that the value is treated in such a way that it is the same as Null, so NullCoalescing is good enough imo.

[rouke-broersma]

I would propose to remove the syntax type and other classifiers (negation, operator, expression etc) where possible.

For example:

AdditionOperatorNegation -> Addition
AdditionAssignmentNegation -> AdditionAssignment

In the spirit of KISS, I think we should simplify where possible.

[dvcopae]

Proposal has been extended in #2995.

[nicojs]

I've changed the PR based on a conversation we just had.

Attendees: @hugo-vrijswijk @rouke-broersma @richardwerkman @xandervedder

• Remove the technical node names from the node identifier. Like MethodCall, Expression and Literal.
  • This improves readability and makes it easier for non-technical people to understand (mind you, easier, not easy 😉).
• Remove Replacement postfix. The AbcToXyz should be clear enough
• Change Nullish in favor of Null for example: NullCoalescingAssignmentToAndAssignment instead of NullishCoalescingAssignmentToAndAssignment
• LogicalAnd And. LogicalOr Or. LogicalNot Not.
• OptionalMemberToRequired instead of OptionalMemberRemoval (my idea ☝)

[hugo-vrijswijk]

These lines break the website build: https://github.com/stryker-mutator/stryker-mutator.github.io/actions/runs/8295994425/job/22704167287#step:7:39

[nicojs]

Yep, sorry! #3066 fixes that.

Btw, it's not the line breaks, but the way of escaping backticks.