# Why Not RDF in the AT Protocol? by Paul Frazee ## Metadata - Author: [[Paul Frazee]] - Full Title: Why Not RDF in the AT Protocol? - Category: #articles - URL: https://www.pfrazee.com/blog/why-not-rdf ## Highlights added January 31, 2024 at 11:02 AM To solve the semantic element, we introduced the notion of a namespaced identifier (NSID) which is simply a form of reverse-DNS. We chose reverse-DNS to strongly indicate that a data type was being identified and not a resource.3 ([View Highlight](https://read.readwise.io/read/01hnepfd5n7m9vk7vbex9j6c5r)) - Note: What? How does reverse DNS indicate a datatype? We wanted it to translate cleanly into static type systems for generated code, and we wanted runtime validation to be reliable enough that applications would not break due to bad data. ([View Highlight](https://read.readwise.io/read/01hnepjsdzkv6h2ntcqtkznkv1)) Lastly, we wanted to address evolvability of the data as it relates to forwards/backwards compatibility as well as the introduction of new behaviors from outside of the original schema authors. ([View Highlight](https://read.readwise.io/read/01hnepkfhywtbcfbj2q5tsa2aw)) When data is transmitted, it uses the `$type` field to identify its schema. This resolves the union to one of the specified types. ([View Highlight](https://read.readwise.io/read/01hnepqjre8q6wap6skj2bm0f9))