On 27 January 2017 at 07:48, Per Bothner <[email protected]> wrote: > What is interesting about the prototype (besides the sidebar) is that > it *enables* (but does not implement) features like whole-document search. > This is difficult and/or expensive with a normal multiple-page > implementation. > The prototype implementation can load the entire document *if needed*: > Normally, it only loads pages as you visit them, but previously-loaded > pages are still in the "frameset" and accessible without network access. > To do whole-document search you do need to load all the pages, but no > page is loaded more than once, even if you do many searches, and do lots > of back-and-forth navigation. > > It's the best of both worlds: The benefits of --no-split (e.g. efficient > search of the entire document; no reload as you navigate back and forth) > with the benefits of --split (fast startup; don't download pages you don't > need).
I see. That sounds like a good idea.
