Weather Inclusion

2026-08-15

For more than ten years, I've been picking away at reviving Tamsin Bowles' Weatherpixie. A website (weatherpixie.com) and a Twitter account (@weatherpixieHQ), I was enchanted by the tiny pixel paperdolls, and when they went dark, with an apparent backup failure, I decided to revive a version as a direct homage.

See: pixiereport/about.

Today's Goal

Add a PixieReport inclusion to the blog. Maybe it'll be in the main template. Today it's a randomly chosen weather station image with generic alt text. We can do better than that. Perhaps tomorrow.

PixieReport

If no image renders above, time to add fail2ban.

Or, as literal text:

<img 
  src="https://pixiereport.com/png" 
  alt="a randomly chosen pixel doll weather report"
  title="PixieReport Weather Image"
  width="125"
  height="175"
/>

Current Directions and Challenges

• Usability: The original Weatherpixie site was easy to use, having a wizard to pick the weather report location and the doll set.
• Visibility: The WeatherpixieHQ feed was visible, periodically posting a weather report to the Twitter feed, and linked back to the site.
  • Before Twitter got X'd out and removed the free tier, there was a PixieReport feed which posted a random report every fifteen minutes from my old laptop via (MacOS) cron / scheduled runs.
  • The replacement for the Twitter feed is a Bluesky feed with very occasional manual posts and updates; it's present, but not very lively. I spoke too soon! There's another Weatherpixie fan on Bluesky asking about PixieReport. Two new doll sets added.
  • One open direction is reviving something similar as a Mastodon auto-posting service, maybe or maybe not having a responder to mentions.
• Ops: I manually run the Let's Encrypt certbot updater, since I haven't separated the services running on http:80 from those running on https:443, and certbot needs to take over http:80 to verify the DNS:IP mapping before signing and updating the cert.
  • I could script the down / run / up cycle, though it hasn't felt like an obvious win.
• Traffic: I'm hammered by scrapers, significantly originating from the Tencent and Alibaba cloud netblocks, which overload the site without regard for robots.txt or whether maybe there's not much there there worth a continuous and repeated crawl.
  • A damage reduction attempt has been via blocking netblocks from which scraper traffic originates. It hasn't been a lasting success, and blocking x.y.z.w/12 netblocks feels like a blunt instrument.
  • An obvious direction is implementing fail2ban since the scrapers have patterns which I would not expect to see from humans. Update: I blocked more netblocks using netfilter.

Blog Formatting

The mobile width (media query, max-width = 760px) was really beating up on the included image rendering. Maybe I should provide or suggest an iframe? Sigh.