Getting mature
Now that BSDF is up and running pretty well for Python, Matlab/Octave and JavaScript, there are just a few things left to bring it to a level where its ready for a large audience:
-
Get feedback and settle the spec (e.g. #9 (closed)). -
Custom extensions for JS. -
Improve extensionAPI for Python (e.g. allow defining as class or dict). -
Improve blob API for Python + docs and tests. -
Improve stream API for Python + docs and tests. -
#5 (closed): Standard extensionfor nd arrays, plus implementations. -
#6 (closed): More docs and examples, including for custom extension. -
Test with random data should include bytes (and or array data). -
Tweak python impl for performance (e.g. encoder has closures). -
Decide on policy wrt closed streams and implement that. -
A policy wrt hashes, and implementation of validating hashes in more languages (not critical). #15 -
Mention better in docs that BSDF supports zero-copy reads (for uncompressed data) due to byte alignment. -
Verify that version of all implementation is the same for first two numbers.
Edited by Almar Klein