User:Larry Gilbert/EXI

The current choice for compression of the OSM database export is PBF. But why? PBF solves a problem that already had a better solution: Efficient XML Interchange (EXI).

EXI has a streaming API that mirrors XML's streaming API, so software that works with XML streams can be adapted to EXI streams with very little modification. It uses a unique coding/compression scheme designed specifically for XML's structure, providing superior compression that only gets better when used in conjunction with XML Schema. (I have personally seen it do as well as 100:1—yes, one hundred to one!)

I would like to provide a proof-of-concept recoding of Planet.osm in EXI format. This will not be a trivial undertaking due to the size of the OSM data export, requiring a large amount of storage and bandwidth. I have a modest amount of experience using Amazon Web Services, and I think that is how I will be tackling it—funding permitting.

—Larry Gilbert (talk) 19:17, 2 February 2013 (UTC)