| > 1) The in-memory DB .exe was around 500 KB. Imagine that It still is, but the hot path is much smaller than that. > 2) The Q language syntax, while consistent, is fairly arcane and throwback to decades past. I can't comment about this. I don't mind the syntax. I prefer k syntax though. > 3) The documentation and driver support is abysmal. This is getting a lot better. The fusion API[1] goes a long way towards better "drivers", and the new documentation site[2] shows a lot of energy being put into organization. There's also Q for mortals[3] which is linearized for people who like that. [1]: http://code.kx.com/q/interfaces/fusion/ [2]: http://code.kx.com/q/ [3]: http://code.kx.com/q4m3/preface/ > 4) It's supposedly extremely fast, but I can't help but wonder if this is a lot of successful PR and hype There are a lot of good benchmarks of kdb[4] unlike Oracle :) [4]: https://stacresearch.com/m3 |