>And the table would very likely be mis-representative (or downright wrong) for every combination of OS, filesystem, physical storage type, sqlite3 version, and data workloads other than the precise ones it was tested on. I guess I am not explaining the idea well. We're not defining deep storage, OS, FS, versions, etc., here. If it is presented right, a nice vanilla table should suffice. Of course, in the end, there should be a final statement with that table explaining the exercise and provide what it is trying to do: show a difference on how WAL speed is better in some scenarios than other..., etc. Please test your scenario before jumping into conclusion, etc., etc." It does not have to go deep into any specific area, but just plain vanilla SQLite WAL processing. However, it will help point #1 carry weight.