tony9402/baekjoon: 23 problem lists for a judge that ended

Twenty three tag folders, every badge still reading Doing

The table near the top of the README indexes 23 algorithm tags, numbered 00 through 22, running from Data Structure and Tree to Trie and Dynamic Programming On Trees. Every row carries a folder path under ./algorithms/, a count of recommended problems, a count of total problems, and a status badge. All 23 badges read Doing, so no tag is marked finished, and status.md is linked separately as the running progress file.

Graph Traversal is the widest folder at 96 problems with 31 of them recommended, and Brute Force sits just behind at 94 with 33 recommended. At the narrow end, Dynamic Programming On Trees holds 8 problems and 4 recommended, Trie holds 9 and 5, and Topological Sorting holds 14 and 5. The distance between the two columns is where the editorial judgment lives: for Graph Traversal the recommended set is under a third of the folder, while Backtracking keeps 28 of 75.

The recommended column is a filter, not a syllabus

Selection followed one rule: take as many problems per tag as could turn up in a coding test, then mark the ones worth finishing before moving on as recommended. The README states that the numbering has nothing to do with study order, so folder 14, Divide and conquer with 18 problems, is not a step that comes after folder 13, Backtracking with 75.

The reader this is written for is applying somewhere specific. Algorithm types differ by employer, and the instruction is to work the tags matching the company you are targeting rather than the whole table, though picking from the recommended column inside those tags is also allowed. That advice carries a caveat the author attaches to the company level file as well: most of it was gathered from people who had taken the tests, so entries can be missing or wrong. Read CodingTest.md as a lead list.

The filtered website is off, and the judge with it

Two services this repository grew around are no longer running. The first is its website. The README describes an extension of the Baekjoon problem list with filtering and search that would be more convenient than browsing on the site itself, and that whole line is struck through in the file, followed by a plain statement that the web page is halted because the home network is unstable.

The second is the judge. The README states the Baekjoon service has ended, that some of it survives in web archives, and that a return is expected at some point, which is the stated reason the repository is kept. So the problem numbers here point at a closed destination. For browsing convenience the README points elsewhere, to a static page built from the solutions repository at https://tony9402.github.io/algorithm-solutions/.

Solutions moved out, and the open pull requests with them

Answer text used to live here and mostly no longer does. Instead of uploading solutions to this repository, contributors are asked to open a pull request against tony9402/algorithm-solutions, and the README notes that everything already merged into the main branch has been moved over. People holding a pull request that is still open here are asked to submit it again there.

A solution/ directory remains at the top level, and two similarly named files sit next to each other. picked.md is today's problem, the single entry that replaced the older mock set where problems were drawn regardless of algorithm category, and picked_legacy.md sits beside it. guideline_for_contribute.md is the file for the contribution side.

Three languages planned, C++ first, nothing released

The missing piece is worked solutions, and the plan is written down: once the problem collection is far enough along, solutions will be posted in Python, Java and C++, with C++ going first. C++ is also the language recorded as primary for the repository, which fits a set built around data structures and graph work rather than scripting.

None of that is installable yet. There are no GitHub releases, so there is no tag, archive or build to download, and the last push was on 2026-11-01. For most tags the workflow therefore stops at the list: this repository tells you which problems to solve, and the answers sit in the sibling repository's static page. Reading a solution in place is not a step you can count on today.

No algorithm explanation lives in this repository

Open a tag folder and you get a list of problems, not an account of the technique behind them. The README says algorithm concepts are not described here and points to link_for_study.md, a collection of links for studying algorithms, with an extra warning aimed at people who are starting out. algorithm_list.md sits beside it at the top level.

That is the project's firmest boundary. It can tell you Greedy has 62 problems and 27 of them are recommended, which is real filtering work, and it cannot tell you why a greedy choice collapses on the case you get wrong. Anyone who wants a study sequence brings the explanations from outside, and the README asks beginners to do that reading before touching a problem.

Cloning it and reading picked.md

The README gives no installation command, which for a repository of Markdown files and folders is not really a gap. What it does give is where things live: problem folders under ./algorithms/, today's problem in picked.md, progress in status.md, company level categories in CodingTest.md, and the solution browsing page at https://tony9402.github.io/algorithm-solutions/. The default branch is main and the licence field says MIT.

Fetch it with a clone and there is no build step afterwards:

git clone https://github.com/tony9402/baekjoon
cd baekjoon

What you should see at the top level is working directories and Markdown files:

algorithms/
baekjoon_utils/
solution/
scripts/
markdown/
assets/
picked.md
picked_legacy.md
status.md
update_log.md
algorithm_list.md
CodingTest.md
link_for_study.md

Also present are .github/, .gitignore, LICENSE, README.md and guideline_for_contribute.md. Opening status.md is the honest first move, because that is the file the README links as current progress while every tag row still reads Doing. Each tag row then resolves to a folder path, so a first session is three reads. Start with the tag your target company asks for:

./algorithms/graph_traversal
./algorithms/brute_force
./algorithms/dynamic_programming_on_trees

Open a tag folder to see its layout. Solve locally, and remember the judge is gone, which leaves no online submission to run code against.

One workflow badge is live, two sit inside HTML comments

Automation exists in the layout but is thinly advertised. The README links a badge for the .github/workflows/auto_update.yml workflow. Two other badges, for auto_solution_update.yml and auto_pick.yml, sit in the file wrapped in HTML comments, so they are not announced as running jobs and nothing in the README says whether they ever were.

For upkeep, read the facts narrowly. The repository is not archived and the last push landed on 2026-11-01, so the Markdown is still being touched. There is no release history, which makes sense for a repository whose products are lists and links. The MIT licence covers redistribution of those files without imposing anything on a reader's own solution code.

Editorial conclusion

Adopt tony9402/baekjoon as a filtered index when your target employer names algorithm types and you want a recommended subset per tag. Skip it if you need explanations, answer text, or a working judge. Check first that the problem numbers in the tags you plan to use still resolve somewhere, because the README states the Baekjoon service itself has ended and says the repository is kept on the expectation that it returns.

Frequently asked questions

Official sources

1. Issues
2. License: MIT
3. README
4. tony9402/baekjoon on GitHub