Spreadsheets covering every problem in this repository. Click a file on GitHub and it appears as a table with a search box. The five family files with gaps are sorted smallest first. Download the raw file for additional sorting or filtering.
One row per problem. The columns tell you where the problem came from and, for some families, how close the published bounds are.
A script writes these files from the problem records. Values are copied from the records, and each gap is calculated by subtracting one published bound from another.
Rebuild them with python3 scripts/build_browse_tables.py. Do not edit them by hand,
because the script overwrites them.
The combined table uses three record kinds.
A conjecture is an individually curated conjecture in
problems/conjectures/. A family row is one instance of a sourced mathematical
family. A catalog row is a sourced entry in problems/extended-catalog/.
all.csv lists all 14,865 records of all three kinds. It only has
columns that mean the same thing for every kind, so it has no gap column. It is
over a megabyte, so GitHub may decline to show it as a table. The smaller files
below are the ones to browse.
kind-conjecture.csv lists the conjecture records.
The catalog-<source>.csv files list the extended catalog, one file per
source.
The remaining files each cover one family. Five include a gap column, with the
unit in its name, such as gap_blocks. The difference-set file has no gap
column.
The combined, conjecture, and catalog tables use the first six columns below.
Family tables use id, source, best_known, and, where applicable, a gap
column.
| Column | Meaning |
|---|---|
id |
name of the problem, unique across the release |
kind |
conjecture, family, or catalog |
title |
the problem's title |
source |
which published collection it came from |
family |
the mathematical family a record belongs to, empty otherwise |
status |
the status recorded from the source |
best_known |
the current record to beat, family files only |
gap_* |
how far apart the best answer and the best proved limit are |
A gap_* column is the difference between the best published construction and
the best proved bound. Its unit appears in the column name. Compare gaps only
within the same file.