Rainer Simon
LightNews LightNews
aboutgeo.bsky.social
Rainer Simon
@aboutgeo.bsky.social
420 followers 150 following 230 posts
A collector of things worth knowing and things not worth knowing. Freelance software developer. Web/UI/Maps/Visualization/Image Annotation/Open Source. https://rainersimon.io
Posts Replies Media Videos
Reposted by Rainer Simon
dhawards.bsky.social
Nominations are open for DH Awards 2025.
Anyone can nominate things that are #DigitalHumanities, available to voters, and updated in 2025. #DH

Nominations close 2026-02-26 and the only way something gets on the ballot. You only need to nominate something once.
dhawards.org/dhawards2025...
DH Awards 2025 – Call For Nominations | Digital Humanities Awards
dhawards.org
December 31, 2025 at 5:53 PM
Reposted by Rainer Simon
kgeographer.bsky.social
Oops, left out the url to my web site that explains me more fully...https://kgeographer.org
kgeographer.bsky.social kgeographer 🇺🇦 🇬🇪 🇩🇰 @kgeographer.bsky.social · 15d
I'm now seeking part-time development contract(s) after a semi-retirement phase of personal research. Projects or experiments in spatial humanities, geo-inflected NLP, LLM stuff, software design, data modeling & munging all possible. DM me if you think of something and RT if inclined. Thx.
December 21, 2025 at 3:03 AM
Reposted by Rainer Simon
danielvanstrien.bsky.social
Built a 2.5MB image classifier that runs in the browser in an evening with Claude Code.

I used a dataset I labelled in 2022 and left on @hf.co for 3 years 😬.

It finds illustrated pages in historical books. No server. No GPU.
December 19, 2025 at 12:08 PM
aboutgeo.bsky.social
🎄 December = website refresh season!

I’ve just updated my site with more info on how I help Digital Humanities researchers, DH labs, museums, libraries, and archives with solutions for data exploration, annotation, and analysis — plus a selection of past projects.

Take a peek: 👉 rainersimon.io
rainersimon.io
Digital Humanities / IIIF / UX / Open Source
rainersimon.io
December 16, 2025 at 1:09 PM
Reposted by Rainer Simon
hildedw.bsky.social
First steps for pilot work on the automated transcription, translation, and entity/event extraction of Qing archival dossiers and their linking across other datasets and other archival and historical sources.
Impressive results even on cursive script.
December 13, 2025 at 10:37 AM
Reposted by Rainer Simon
iiif.bsky.social
🔊The Call for Proposals for the 2026 #IIIF Annual Conference is now open!

Read the full CfP and find a link to submit: iiif.io/event/2026/n...
IIIF 2026 Annual Conference, The Netherlands, June 1-4, 2026. Amsterdam, Leiden, The Hague In collaboration with dutch digital heritage network
December 10, 2025 at 6:52 PM
aboutgeo.bsky.social
Clarification on what our "VLM pipeline" is (thx @miaout.bsky.social): code and utilities that chain together workflows across AI models – incl. the newer vision-capable language models (VLMs). Goal is to extract structured metadata from images.

Have material you'd like me to test? Ping me!
aboutgeo.bsky.social Rainer Simon @aboutgeo.bsky.social · Dec 3
Working with a small museum/archive/project with digitized images but little metadata?

I'm looking for testers for a VLM pipeline for auto-enrichment (transcription, captions, tags, IIIF). If you share a few sample images, I'll run them through + share results. Would love to hear your feedback!
December 5, 2025 at 11:28 AM
aboutgeo.bsky.social
Working with a small museum/archive/project with digitized images but little metadata?

I'm looking for testers for a VLM pipeline for auto-enrichment (transcription, captions, tags, IIIF). If you share a few sample images, I'll run them through + share results. Would love to hear your feedback!
December 3, 2025 at 6:05 PM
aboutgeo.bsky.social
Q to the hivemind: I'm wondering if anyone else around here's been building RAG systems for cultural collections/museums?

I'm curious about what/how you're indexing. Metadata, transcripts/chunks? Images full or segmented? LLM-generated descriptions? What's your queries and response workflow?
November 27, 2025 at 6:02 PM
aboutgeo.bsky.social
Odd find of the day from #MachinaEmblematica: the Simivulpem, a wolf/monkey hybrid.

Story sounded plausible. But I wonder how much is just AI hallucination. Google only gave two hits, both on the Google Books edition of Symbola et Emblemata.

Any early modern creature lore experts who know more? 🙂
The Simivulpem, also known as the "half-fox" or "fox-monkey," is a fascinating creature described in the Symbola et Emblemata. This unique animal, found in regions of India, is a hybrid-like being with distinct features: the front part resembling a fox and the back part resembling a monkey. It has a tail similar to that of a cercopithecus, bat-like ears, and a pouch under its belly, akin to a marsupial's crumena, where it carries its young. This emblem symbolizes the protective instincts of parents and the care they provide to their offspring. The Simivulpem's ability to carry its young in a pouch highlights the theme of parental love and protection, ensuring the safety of its young as it moves about. This is further emphasized by other similar creatures mentioned, such as the animal in the Moluccas islands, which also carries its young in pouches, and the marine dog, which can retract its young into its body during danger. The Simivulpem serves as a metaphor for the nurturing and safeguarding role of parents, urging us to appreciate and cherish the care we receive from those who love us. It also reflects the idea of carrying our burdens and responsibilities with us, much like the creature carries its young, reminding us to be mindful of our duties and the support we have from others. An image (supposedly) of the Simivulpem
November 26, 2025 at 11:01 AM
aboutgeo.bsky.social
🎉 IMMARKUS 1.0.5 is out! The latest release of our open-source image annotation tool for humanities, heritage & museum professionals is now available – with bug fixes, productivity enhancements, and improved #IIIF annotation import.

Try it here: immarkus.xmarkus.org
Docs: github.com/rsimon/immar...
Screenshot of the IMMARKUS annotation interface with three historical images—an Oxford bookplate, a medieval-style world map, and an early printed nautical map—displayed side by side, with metadata fields visible on the right.
November 25, 2025 at 6:58 AM
aboutgeo.bsky.social
The #MachinaEmblematica – our mysterious chatbot guide into the 16th century encyclopedia "Symbola et Emblemata" is now reasonably mobile-friendly.

Check it out if you haven't already!

machina.rainersimon.io
An emblem depicting frogs on a pond. Chat message from the Machina Emblematica explaining how frogs are used as a symbol for arrogance and deceit, that cannot withstand the light of truth.
November 24, 2025 at 2:31 PM
Reposted by Rainer Simon
hildedw.bsky.social
Cold and rainy but good fun at the annual Science Day in Leuven where we demonstrated how anyone can easily collect and work creatively with high-quality images streamed from cultural organizations globally. #dvdw2025 #IMMARKUS
November 23, 2025 at 3:04 PM
aboutgeo.bsky.social
We built #MachinaEmblematica as an outreach/engagement tool rather than for research. But as someone with zero knowledge of 16C emblem books (or Latin!) I find it an amazing way to poke around & discover odd new things—far richer than randomly browsing #IIIF images alone.

machina.rainersimon.io
November 21, 2025 at 5:49 PM
aboutgeo.bsky.social
@rockettara.bsky.social and I've been teaching the #MachinaEmblematica some new tricks: better image interpretation and conversational context, smarter source retrieval from the Symbola et Emblemata. It’s a bit slower now… we’ll work on that next. 😉

Try it here: machina.rainersimon.io
The Machina Emblematic user interface explaining an image depicting a dragon. Image on the left, text on the right.
November 20, 2025 at 5:15 PM
aboutgeo.bsky.social
text-annotator-js – our JavaScript library for text and PDF annotation – is looking for a new name.

Something shorter, nicer and more memorable. (Current favourite: "leuchtmarker" – German for "highlighter pen".) Got ideas? Drop them below!

github.com/recogito/tex...
Better Project Name · Issue #238 · recogito/text-annotator-js
The name "Recogito Text Annotator" was actually meant as a temporary working title. As always with working titles, it stuck... IMO it's both land and a bit too verbose. Also: it's currently under t...
github.com
November 20, 2025 at 11:00 AM
aboutgeo.bsky.social
Our #MachinaEmblematica has a few quirks. But if you ask for specifics of a particular page or emblem, I feel that the multimodal model generates pretty useful responses!

machina.rainersimon.io
A screenshot of the Machina Emblematica, giving an explanation of the depicted emblem (a beaver), the accompanying text, and the possible meaning and author's intention.
November 18, 2025 at 2:26 PM
aboutgeo.bsky.social
Side-project announcement! @rockettara.bsky.social and I built a machine that never existed… to read a book from 1590. Meet the #MachinaEmblematica, a chatbot for exploring the Symbola et Emblemata. machina.rainersimon.io
Rendered image of the "Machina Emblematica" a fictional arcane mechanism with cogs wheels and pistons. The chat interface of the Machina Emblematic website, dark background, an open page of the Symbola on the left side, the conversational chat interface on the right. An artists impression of the Machina's brain, looking like a layered microchip with an organic nervous system. The image is used as a decoration on the splash page.
November 14, 2025 at 9:35 AM
aboutgeo.bsky.social
Trying a new approach to my overlapping annotations problem:

• No more stacked underlines
• Instead, overlaps get small start markers
• Exact overlaps show a tiny count number
• Color density still hints at overlap

Try it live: dense-annotations.netlify.app

One caveat… see next post 👇
Example text with multiple orange overlapping highlights, illustrating the new design ideas.
November 5, 2025 at 12:22 PM
aboutgeo.bsky.social
Here's my main dilemma with overlapping text annotations: Underlines become a mess once you stack more than three. At the same time, there's really no way around them if you want to show where overlap starts and ends.

I guess the Q is how/when to drop underlines without losing too much information?
An image illustrating the problem: a text snippet which has two overlapping annotations across the text "that ingenious hero who travelled far and wide". Without using underlines, it's impossible to differentiate the following two cases: 1. The first annotations goes across the whole text, the second annotates only "ingenious here". 2. The first annotation annotates "that ingenious hero", the second annotates "ingenious here who travelled far and wide". In both cases, the region of overlap is the same, which means both cases look the same when using color density.
November 4, 2025 at 2:13 PM
aboutgeo.bsky.social
Task for today: brainstorming better ways to visualize heavily annotated text. The more I think about it, the harder it gets...
• Partial overlaps → clutter & a11y issues
• Exact overlaps → even worse
• Stacked underlines → work great... until they don't.

Seen elegant solutions? I’d love to chat!
A heavily annotated text snippet (the opening paragraph of Homer's Odyssey) with many overlapping orange highlights and underlines, and numeric markers that indicate places that have multiple annotations on the same text range.
November 4, 2025 at 12:09 PM
Reposted by Rainer Simon
performant.bsky.social
Check out this amazing project so many of us at Performant Software have been working on for the past couple of years. We can't wait for the launch and for everyone to see it!
natboundunbound.bsky.social Native Bound Unbound @natboundunbound.bsky.social · Oct 29
The world premiere of Native Bound Unbound: Archive of Indigenous Slavery will be streamed live on YouTube this weekend.

Watch live:
Day 1 (NHCC - Albuquerque) www.youtube.com/live/xiA5Mp_...
Day 2 (MIAC - Santa Fe)
www.youtube.com/live/5r-vWjG...

Follow @natboundunbound for updates and links.
October 31, 2025 at 4:46 PM
Reposted by Rainer Simon
ciaota.bsky.social
Had the chance to join a workshop on image annotation in IMMARKUS led by @hildedw.bsky.social, Rainer Simon, Sunkyu Lee, and Dawn Zhuang today. This was great fun and definitely a tool I will use when working with visual media and texts in the future! #digitalhumanities
Participants looking at the screen that shows a slide with maps.
October 23, 2025 at 2:05 PM
aboutgeo.bsky.social
We’re improving accessibility in the Recogito Text Annotator, our open source text annotation library.

Keyboard support is solid, but highlights aren’t directly tied to the text. They're a floating layer above it, breaking screen reader access. If you’ve faced similar a11y issues, I’d love to chat!
October 22, 2025 at 1:14 PM
aboutgeo.bsky.social
Arrows between annotations - a recurring feature request I’ve revisited a few times. Building yet another version in a current project:
• Hand-drawn (no auto layout)
• Can snap to annotations - or not
• Works as both a visual tool & a semantic/data-model link

This feels like the nicest version yet!
October 10, 2025 at 1:10 PM