Good day. I have read several times in Discord that ByteCode will be removed in TP 4.
Taking into account our intention to support WASM compilation of Gremlin, could you provide reasoning for the ByteCode removal ? I am concerned that it will have a negative impact on the users with small databases whose data is completely cached in memory and who want to see low latencies during processing of their requests, as generated parsers are notorious for low throughput and massive heap pollution. We are developing LDBC benchmarks, and I have the intention to conduct experiments that prove or disprove my concerns about this use case. Meanwhile, understanding the reasoning behind such removal will be quite useful.
