MINIMALDATA is a different and much larger ruleset. It also has quite low test coverage (since it is only a standardness rule), so from what I can tell, many more tests would have to be added to make it consensus ready.

MINIMALDATA does the following, very little of which is tested:

• All pushes when executing scriptSig must use minimal push form.
• All pushes when executing scriptPubKey must use minimal push form. (this does nothing to reduce malleability, and only makes scripts unspendable)
• All pushes when executing redeemScript must use minimal push form. (this does nothing to reduce malleability, and only makes scripts unspendable)
• The operand of OP_CHECKLOCKTIMEVERIFY and OP_CHECKSEQUENCEVERIFY must be a number encoded as minimal byte array.
• The operand of OP_PICK and OP_ROLL must be a number encoded as minimal byte array.
• The operand of OP_1ADD, OP_1SUB, OP_NEGATE, OP_ABS, OP_NOT, OP_0NOTEQUAL must be a number encoded as minimal byte array.
• Both operands to the following opcodes must be numbers encoded as minimal byte arrays: OP_ADD, OP_SUB, OP_DIV, OP_MOD, OP_BOOLAND, OP_BOOLOR, OP_NUMEQUAL, OP_NUMEQUALVERIFY, OP_NUMNOTEQUAL, OP_LESSTHAN, OP_GREATERTHAN, OP_LESSTHANOREQUAL, OP_GREATERTHANOREQUAL, OP_MIN, OP_MAX
• All three operands to OP_WITHIN must be numbers encoded as minimal byte arrays
• The nPubKeys and nKeysCount arguments to OP_CHECKMULTISIG and OP_CHECKMULTISIGVERIFY must be numbers encoded as minimal byte arrays.
• The second argument to OP_SPLIT must be a number encoded as minimal byte array.
• The first argument to OP_NUM2BIN must be a number encoded as minimal byte array.
• Maybe some other things, I didn't look in full detail and the lack of tests does not give any help here.

The proposed rule does the following:

• The scriptSig must have canonical form (only pushes, using minimal push forms) for producing a given stack.

See T667