But in addition you could also allow for 5% of the block space to be free-use (similar to the old space for mining txs by priority) after having filled up 95% of the block space deterministically.
What is the motivation for doing this? Every tx included this way needs to be transmitted and validated with every share-chain block, so possibly as fast as every 250ms.
The entire purpose of the deterministic block template idea is to avoid having to transmit or validate any txs at all (because it’s already been done in the DAG parents).
As regards censorship resistance, graffiti blocks, parasitic assets and the like, Braidpool will need to have its own tx acceptance rules (yet to be determined), which will fall somewhere between Bitcoin’s P2P rules and Bitcoin’s block rules, avoiding the well-known pitfalls. Personally I don’t see any point in being very opinionated about acceptable txs and I don’t plan to be as restrictive as OCEAN and Knots, nor do I plan to implement tools in Braidpool that make censorship “easy”, whether the user wants to omit graffiti/runes/BRC-20 or prevent Russians from donating to Ukraine. We will be liberal in what we accept. Those who want to omit certain txs will have to fork the code and run a modified client (like Knots). Those who want to be even more liberal than what we choose can either fork the code to make a new instance of Braidpool or solo mine.