11–20 of 154 posts
Hello HN! Post author here. I’m happy to answer questions & fix typos once morning rolls around here in Australia
Earlier quoted context omitted.
Have you used CRDTs to solve any practical problems? If so, how does the CRDT solution compare to a non-CRDT solution? If a non-CRDT solution is feasible at all?
Article mentions at the beginning that the author used CRDT in Google Wave/ShareJS.
FWIW, I am myself also curious about this (the question of comparing CRDT to non-CRDT solutions): I found OT beautiful, but never really felt CRDT had the same feeling of elegance; and so I am downright fascinated to see the person I have always seen as a "god of OT" deciding to forsake it and move to CRDTs going forward. Are they really that great? Is there something simple I am missing for the explanation for why they are so much better? (Is it maybe that they can support arbitrary merging without a sequencer to linearize the edits? Honestly, my question probably sucks a bit as I haven't spent any time thinking about OT or CRDT in at least three years--and even then only once since a few years before that, as I have had other things to spend most of my mental energy on recently--and so I am failing to remember the details of my use cases or the tradeoffs I saw or the implementation issues I felt were easy/hard.)
Hello HN! Post author here. I’m happy to answer questions & fix typos once morning rolls around here in Australia
Obviously this might move later (maybe it can only increase?), but usually not by much, so I would guess it would make an efficient starting point / be immediately correct 99% of the time?
Looks like you already have 2 good solutions to this though (start from index of recent edits and range tree).
On a meta-level, does anyone else think that the whole idea of writing a peer reviewed paper that is just a benchmark of different algorithms should be really rigorously reviewed before being accepted? Writing good benchmarks is hard, and so highly contextual that writing fair comparisons beteen algorithms (or data structures) is almost impossible unless you're an expert in all of the algorithms involved.
Best example is discrete optimization research (traveling salesman, vehicle routing and its variants, schedule rostering etc.). Stuff you find in the papers there achieves state-of-the-art results very slowly (using integer linear programming or some rarely optimized heuristics) making you believe these instances of a general NP-hard problem can't be solved quickly.
When you start tinkering, you either find that data structures can be added that reduce the complexity significantly or that there are regularities in instances that, when exploited, support massive speedups.
I would say that TSP research is an exception but most of stuff coming out that has a lot citations is way too slow and is never as brilliantly implemented as Lin Kernighan heuristic or other stuff from the age of insanely slow computers.
The difference between diamond native and diamond WASM demonstrates how, even with WASM, native implementations beat browsers hard, and native implementations performance-wise are still very worth, specially for lower powered devices, and, perhaps, reducing battery usage (as consequence of less CPU use) in mobile devices.
On a meta-level, does anyone else think that the whole idea of writing a peer reviewed paper that is just a benchmark of different algorithms should be really rigorously reviewed before being accepted? Writing good benchmarks is hard, and so highly contextual that writing fair comparisons beteen algorithms (or data structures) is almost impossible unless you're an expert in all of the algorithms involved.
On a meta-level, does anyone else think that the whole idea of writing a peer reviewed paper that is just a benchmark of different algorithms should be really rigorously reviewed before being accepted? Writing good benchmarks is hard, and so highly contextual that writing fair comparisons beteen algorithms (or data structures) is almost impossible unless you're an expert in all of the algorithms involved.
Problem is that academics are rarely experts at programming or have knowledge of computer architectures as much as someone in the industry. There are various tricks that are never taught at college, therefore academics have no idea some stuff even exists. Best example is discrete optimization research (traveling salesman, vehicle routing and its variants, schedule rostering etc.). Stuff you find in the papers there a…
Yes, or looking for things like mathematical purity, linear problem statement, etc
In practice: you don't need the best solution and you can get a great solution in a multitude of ways.
Hello HN! Post author here. I’m happy to answer questions & fix typos once morning rolls around here in Australia
In general, when I see C code and I’m asked to speed it up, I always use “100x” as my target baseline.
Correct me I'm mistaken The difference between diamond native and diamond WASM demonstrates how, even with WASM, native implementations beat browsers hard , and native implementations performance-wise are still very worth, specially for lower powered devices, and, perhaps, reducing battery usage (as consequence of less CPU use) in mobile devices.
On a meta-level, does anyone else think that the whole idea of writing a peer reviewed paper that is just a benchmark of different algorithms should be really rigorously reviewed before being accepted? Writing good benchmarks is hard, and so highly contextual that writing fair comparisons beteen algorithms (or data structures) is almost impossible unless you're an expert in all of the algorithms involved.