Startup and operation of Ultra Recall is generally very fast. Ultra Recall is a native Windows application (it's not dependent on the .NET Framework), resulting in quick startup times. Data is loaded on demand when it needs to be displayed, rather than loading the entire database into memory at once. Ultra Recall uses an embedded database engine to store its data (items, attributes, icons, keywords, and all related information). The database engine provides atomic updates (all data changes are all or nothing), rock-solid stability, and high performance.
Ultra Recall is also quite efficient in disk space and memory usage. Documents and rich text data are stored in a compressed format within the database.
An index of item text (gleaned from all the text in item attributes and contents of known document types) is also stored within the database file, allowing fast searching of all data, and since indexing is performed as data changes are made, it is always up-to-date and never needs to be rebuilt.
The bottom line is that Ultra Recall is very fast and highly efficient in both disk and memory usage. Performance is very good even on an older system such as a Pentium II-class computer with 128MB of RAM. Even with the all the metadata that it maintains about each item, an Info Database file is typically the same size or smaller than the total size of all documents stored in the database.
This post also provides some tips to maximize performance in large databases:
http://www.kinook.com/Forum/showthre...?threadid=3627