We've accomplished a tremendous amount in the time since we started RethinkDB. However, this is only the beginning, as much remains to be done, and many ideas are yet to be fleshed out. This is a rough roadmap of things to come:
One of the key missing pieces is a live disk garbage collector. A 2gb RethinkDB index file contains only 50mb of real index data. A garbage collector is essential before we can start testing RethinkDB on larger data sets.
Since we're still in the pre-alpha stage, many of the advertised RethinkDB features need more testing, support tools, and in some cases concrete implementations. Once a garbage collector is in place, our first priority will be to work towards a production-quality beta of RethinkDB.
Improved performance. RethinkDB has tremendous potential to perform better than its competitors, and we've only scratched the surface in this area. While RethinkDB is already much faster in many cases, we need to integrate performance analysis tools into the build process and ensure our performance improvements realize the full potential of append-only algorithms.
Time series extension. One side effect of append-only indexing used in RethinkDB is ability to instantaneously query the database the way it looked at any given point in its lifetime (assuming target users are willing to pay for the extra disk space consumed by index history). This is a very interesting feature, but not immediately useful to most customers, so it isn't a priority.
In addition to the above items, we have many more revolutionary features planned that we'll continue adding in parallel to core development. Subscribe to our blog to find out when these features make it to the repository.