bangbing (4) [Avatar] Offline
When I read chapter 8, I can quickly get the purposes and differences of the first 3 header types based on the explanation:
1: Already in the tables
2: Not entirely in the table yet, please add it to the dynamic table.
3: Not entirely in the table, and don't add it to the dynamic table.

However, I personally feel the explanation of the 4th type "never indexed" is a bit brief and unclear, especially the first sentence mentioning "any subsequent re-encoding" seems to be quite abstract.

After read the corresponding part in RFC and some online discussion, I realize this type is specific to enforce the header encoding behavior of intermediaries. Maybe explaining the difference between type 4 and 3 using an example would benefit readers' understanding.
Barry Pollard (22) [Avatar] Offline
Thanks. Will review the text to try to make it a bit clearer.

Keep 'em coming!