[)roi(]
Executive Member
- Joined
- Apr 15, 2005
- Messages
- 6,282
Tagging onto our breakaway into serialisation standards; here's some interesting research done by uber:
https://eng.uber.com/trip-data-sque...al&utm_source=twitter.com&utm_campaign=buffer
Strange: Cap 'n Proto didn't make a showing, guess it's either too new or too small.
Messagepack does fairly ok, near to to the top in terms of encoding / decoding and compression; but there are certainly faster ones, yet in this case Uber chose Messagepack. Overall it's was a good compromise; I don't like strict schema standards + MessagePack is a lot more visually like JSON, except that you can't easily text modify numeric values like JSON (note: well only if your numeric packing skills aren't up to scratch).
https://eng.uber.com/trip-data-sque...al&utm_source=twitter.com&utm_campaign=buffer
Strange: Cap 'n Proto didn't make a showing, guess it's either too new or too small.
Messagepack does fairly ok, near to to the top in terms of encoding / decoding and compression; but there are certainly faster ones, yet in this case Uber chose Messagepack. Overall it's was a good compromise; I don't like strict schema standards + MessagePack is a lot more visually like JSON, except that you can't easily text modify numeric values like JSON (note: well only if your numeric packing skills aren't up to scratch).