Erik Osheim

Currently we do linear splitting. It would be nice to make this pluggable, so we could experiment with Linear, Quadratic, and other splitting strategies. This change should be benchmarked to...

It would be nice to let users use their own Point and Box implementations, and potentially use 3D structures as well. If we had a geometry type class this might...

Speed up ByteBuffer parsing

1

This is relatively slow for some reason. Figure out why!

These are currently missing

There are some smart strategies here: https://github.com/sirthias/benchmarks/blob/master/src/main/scala/benchmarks/json/ArrayBasedAstParser.scala We should steal some of them.

Make JObject and JArray immutable

8

Currently (and historically) `JObject` and `JArray` are mutable. They are case classes wrapping mutable data structures. @ijuma wisely points out that this is maybe not the best for folks who...

Faster whitespace detection

6

At nescala Matthias suggested there was a faster way to check for possible whitespace characters using bitwise logic. We should look into that.

The async parser was built to handle streams of many small JSON values. It does poorly on very large json files containing one value (for instance, an array). This could...

need to run "clean" before "test-coverage" in sbt

1

If I run test-coverage without running "clean" first, I usually get very incomplete test coverage reports.

I think it would be good to collect common questions that people might have about the need for this project. It would help frame the issue, clear up misconceptions, and...