3 ms·
Are there similar algorithms useful for compressing stuff like "a list of the top 100 scores in a game?" We already store these as diffs of the list, but we cou
by anonymoushn 4y ago
Are there similar algorithms useful for compressing stuff like "a list of the top 100 scores in a game?" We already store these as diffs of the list, but we could probably do better than zstd-ing the list-diffs.
- pizza 4y agoYou could look into multiset sequence compression methods, where the multiset elements are i sequences max_score(i, t) for at time t for each rank i https://arxiv.org/abs/1401.6410 https://arxiv.org/abs/1401.6410
- tyingq 4y agoThis discussion about a Barclay's Bank json list of 74,000 numbers might have some ideas: https://news.ycombinator.com/item?id=28326036 https://news.ycombinator.com/item?id=28326036 I was able to get it from a 1.3mb json file (uncompressed) to a 151k uncompressed but 4k compressed file using mostly deltas. https://news.ycombinator.com/item?id=28348965 https://news.ycombinator.com/item?id=28348965
- anonymoushn 4y agoThis is great! I'll have to think about how to combine "storing deltas within the list" and "storing deltas of the list," but there should be some significant gains available