Privacy

When a page is requested, the site adds one row to a small database. A row holds:

• the hour of the request (not the minute) and the page's path;
• the country, and the name and number of the network the request came from: a university, a company, or a home or mobile provider;
• a short label for where the reader came from, such as google, linkedin, a university's domain, or this site, in which case also the page the reader was on;
• whether the device is a phone or a desktop;
• whether the request came from a person or an automated client, and for an automated one its kind: a link preview, a search or AI crawler, an AI assistant, a monitor, or a script.

When a page is hidden or closed, a small script sends one more row: the same network details, how long the page was visible, in buckets (under 10 seconds, 10 to 30 seconds, and so on), and how far it was scrolled, in quarters. Downloading the CV is recorded like a page view.

IP addresses, or anything made from one; cookies or any other identifier of a reader; user-agent strings; full referrer addresses; query strings; anything that links one visit to the next. The only cookie this site ever sets is one I set in my own browsers so that my own visits are not counted.

A row has no identifier, but on a very small network a few rows in the same hour may be one person. I treat the data that way, and I read it in a terminal, not on a public page.

Nothing is recorded for a request that carries Global Privacy Control (Sec-GPC: 1) or Do Not Track (DNT: 1). Turning either on in your browser is enough, and it works in private windows too.

Rows are deleted after 400 days.

The site is served by Cloudflare Pages. As the host, Cloudflare sees the IP address of every request, as any host does; I never receive it. Cloudflare Web Analytics, which is switched on, counts page views, referrers, countries and devices without cookies. The code that writes the rows is public: analytics/visit.js lists every field and is the authority if this page and the code ever differ.