Colophon

Editorial policy

Every word on this site is mine: I write it, or check and approve it before it is published. Facts about me come from my own record, and anything I have not yet confirmed is shown in square brackets until I do.

Every claim about heat and health links to its source: the original study, agency or report, not a news story about it. Every number names its source and year.

Nothing here is confidential to my employer or to the partners I work with. I do not publish anything about my friends, my family or my private life, or about political affiliation.

When I get something wrong, I correct it and record the change in the corrections log below.

Views are my own

The views expressed here are my own and do not represent those of my employer.

How this site is built

The site is made with Astro, which turns plain text files into ready-made web pages before anyone visits. The long texts are written in Markdown, a simple way of writing formatted text. The finished pages are hosted on Cloudflare Pages.

There are no cookies and no analytics. The only script from another company is Cloudflare Turnstile, the robot check on the contact page, described in the privacy notice. Photographs are resized into several smaller files, so each screen downloads only the size it needs, and no image is larger than 300 KB.

The headings are set in Fraunces by Undercase Type and the text in Source Sans 3 by Adobe. Both are free under the SIL Open Font License 1.1 and are stored on this site: read the Fraunces licence and the Source Sans 3 licence.

Each time the site is built, an automatic check reads every page and stops the build if it finds an em dash, the long dash my house style does not use.

Built to be read by machines too

More and more people ask search engines and AI assistants who someone is. I want those answers to come from me, so the site is written for machines as well as for people.

Every page carries a short, hidden description in a shared format called schema.org. It tells a machine who I am, links to my LinkedIn profile as “also me”, and describes each talk and each field test with its date. It is the same description on every page, so there is only ever one me.

Every long page has a plain Markdown copy under /docs/, linked at the foot of the page. There is a sitemap listing every page, and a short summary of the whole site for AI tools at /llms.txt.

The robots.txt file welcomes the AI crawlers of OpenAI, Anthropic, Google, Perplexity, Apple, Amazon and Mistral by name. The only part closed to crawlers is /api/, where the contact form sends its messages.

Accessibility

I aim for this site to meet the Web Content Accessibility Guidelines (WCAG) 2.2 at level AA. In practice, that means:

• text contrast of at least 4.5 to 1, and 3 to 1 for large text, icons and form borders
• lines of about 66 characters, so text is comfortable to read
• every link, button and field shows a clear outline when you reach it with the keyboard
• the whole site works with the keyboard alone, and “Skip to content” comes first
• if your device asks for reduced motion, nothing on the site moves
• every image has a description of what it shows, and headings are in order
• links are underlined, so colour never carries meaning on its own

If something on the site is hard to use for you, please tell me through the contact form and I will fix it.

Corrections log

Date	Page	What changed
25 October 2026	Whole site	Log opened. No corrections yet.

AI disclosure

Written by Michaela Schrodt; structured and built with Claude Code (Anthropic).