PageFish
Data
search index
SHOW HN2018
How HN took it
0 positive, 4 neutral or asking, 0 mixed, 4 critical, out of 8 comments about the project. Thread score -0.35.
Idea: 0 for, 2 againstPerformance: 0 for, 1 againstExecution: 0 for, 1 against
The thread, comment by comment
Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.
positive neutral or mixed critical about another commenter or something else the maker
What HN said
What people liked
Interesting idea of using unicode characters as a high-density encoding
From 1 comment
Instead of addition, why not use an actual hashing algorithm, truncated to a certain number of bits? That'll give you better binning and give you flexibility about filesize /…
sonofgod on HN
Objections
Produces false negatives due to case-insensitive URLs and invalid hashes
From 2 comments
Using some database backed full text search will beat this by a huge margin. Especially when it comes to actually working correctly. As the author states where are some caveats…
LordHeini on HNSome queries (e.g. San Francisco) work on localhost but not Github, apparently because of case-insensitive URLs. That's probably what's up with Egypt and B-W. A better…
peterburkimsher on HN
Too limited compared to database full text search such as SQLite FTS
From 2 comments
Obviously it is faster than grep... which is "linear search"! Compare it instead to some "hashing search" to see how good or fast it really is. For example, take SQLite's Full…
gregoriol on HNUsing some database backed full text search will beat this by a huge margin. Especially when it comes to actually working correctly. As the author states where are some caveats…
LordHeini on HN
Linear grep comparison is not a meaningful benchmark
From 1 comment
Obviously it is faster than grep... which is "linear search"! Compare it instead to some "hashing search" to see how good or fast it really is. For example, take SQLite's Full…
gregoriol on HN
Requests
Use a better hashing or binning algorithm
From 2 comments
Instead of addition, why not use an actual hashing algorithm, truncated to a certain number of bits? That'll give you better binning and give you flexibility about filesize /…
sonofgod on HNI proposed to use a simple polynomial hash https://github.com/peterburk/pagefish/issues/1
netgusto on HN
Sort the data for binary search
From 1 comment
Making a longer hash than 2 bytes would help selectivity, and filesystems are usually very good at finding files by name. Also, apparently the files are not sorted by name (or I…
nine_k on HN
Questions
Whether unicode characters outperform ASCII of the same byte size
From 2 comments
Instead of addition, why not use an actual hashing algorithm, truncated to a certain number of bits? That'll give you better binning and give you flexibility about filesize /…
sonofgod on HNWhilst there's a few hashing algorithms explicitly designed to be memory and time-intensive to make password retrieval impractical, most are really fast, even in Javascript.…
sonofgod on HN
Alternatives named
SQLite Full Text Search
From 1 comment
Obviously it is faster than grep... which is "linear search"! Compare it instead to some "hashing search" to see how good or fast it really is. For example, take SQLite's Full…
gregoriol on HN
PostgreSQL
From 1 comment
Using some database backed full text search will beat this by a huge margin. Especially when it comes to actually working correctly. As the author states where are some caveats…
LordHeini on HN
ElasticSearch, Lucene, Sphinx
From 1 comment
Binning! That's the word, not "hash distribution" like I wrote. Yes, I'd like to experiment more with different hashing algorithms. It needs to run fast on the query side.…
peterburkimsher on HN