>>If you have a just one thread of one process of one program accessing the database, and a ton of memory and filespace free, WAL is significantly faster >That's actually something I'd like to see proven somehow. This is the reason why I asked. If the first point of the explanation of the WAL site states that it is faster in most scenarios, it would be nice to have a little side by side WAL vs Normal scenarios. Where scenarios may be: * 1 writer, 1 reader, 100K records, * 1 writer, 5 readers, 100K records * 2 writers, 10 readers, 100K records then, * simple write, search, and replace * semi-complex write, search, and replace * complex write, search and replace with their appropriate response speed. A nice table like that would make people love it or super love it. :-) Just thinking out-loud... josé