The tokenURI() spec from ERC721Metadata and the uri() spec from ERC1155Metadata_URI are practically identical. What was the specific reason to this versus still at least using the name “tokenURI” so that apps can keep it limited to one spec if supports interface signals either or is supported? Since ERC-721 and ERC-1155 can’t be used together, there would always be a unique contract address to differentiate between the two of them.
ERC-1155 has a unique JSON schema that includes optional formatting to allow for ID substitution by clients but I haven’t seen this used often, and supports interface signaling ERC-1155 could indicate to clients how to handle the token URI.