CPython core team over time
Runs core_devs_activity.py directly in the browser via
Pyodide (CPython + matplotlib as WASM). Listed roster vs. recently-active
contributors over CPython's history.
—listed today
—active in window
—% active
—Python render ms
Sources & caveats
- Roster from the devguide team-log — 207 historical members back to 1989-12-25 (Guido).
- Activity from the GitHub Search API: PR or issue
authored in
python/cpython. - Comments are not counted: GitHub has no per-user-per-repo comment search; a member who only reviewed PRs appears inactive.
- Pre-GitHub era (before the cpython repo migrated to GitHub in Feb 2017): the active series is conservatively low.
- ~50 listed members never had a GitHub account and resigned en-masse on 2017-02-10 — they count as listed during their tenure but never as active.
How this page runs Python in your browser
- Pyodide downloads a WASM build of CPython + stdlib from jsDelivr (~10 MB compressed).
- matplotlib is loaded via
pyodide.loadPackage(~15 MB more). - The data bundle
data.jsis loaded as a normal script tag and exposed aswindow.__DATA__; the page hands those objects to Python viapyodide.globals.set. - The Python program (embedded in an inert
<script type="text/python">block) does the month-walk, day-precision membership check, and active-count via binary search — identical to the desktop script. - matplotlib renders to an in-memory PNG buffer; the bytes are
base64-encoded and stuffed into an
<img>tag, which sidesteps Pyodide's canvas-backend placement story.
Cold start is 5–20 s on most connections; everything cached on revisit. Per-render after that is 50–200 ms.