Greetings
There’s been some discussion of late about the use of Scrivener as a document repository and research database as well as a writing tool. I hope developers will keep this usage case in mind, with consideration of the efficiency and scalability of algorithms.
With 8300 documents, Scrivener’s load time for me is roughly 220 seconds these days, or 3:40. This is off a 6gb/sec Sata-III SSD.
And memory footprint is 584 mb. This seems like a lot for an application that keeps only a few docs open at a time…but then there is Search All and those cascading Go To and Move To menus. Performance is fine once it loads; perhaps it’s eggs and omelets. But there may also be some correctable inefficiency in an inner loop.
The Binder is built on a Tree View, which is known to be a slow control. Developers, may I request you check that you’re allocating the Tree initially to its eventual size, rather than reallocating it as it loads. Just a hunch. And, more generally, that you make a project with ten thousand or more imported documents, to test performance and scalability for document repository customers. All will benefit.
Thanks for considering – Jerome