swiiifz, #IIIF from a single zip file, served by a Service Worker
https://atomotic.github.io/swiiifz/
Little updates: the zip is portable; it doesn't matter what URL is used to generate it. Multiple viewers are available for testing. Tify and UV have problems loading thumbnails, but this is an issue with `mkiiif`, and I'll look into it.
RE: https://ebooks.social/@matt/117475191321085720
This analysis is so great.
Good to see P396¹ (SBN) is the second most popular² identifier among libraries catalogs
¹ https://www.wikidata.org/wiki/Property:P396
² https://thisismattmiller.github.io/wikidata-naf-identifier-inventory/#P396
I have high expectations for this new search engine. It is from the creator of Solr, is not backed by venture capital, does not mention AI or agents, offers a clear and simple HTTP API (JSON and Protobuf), requires no client, has everything clearly documented at https://luxir.org/docs/latest/, and can run on very low resources.
I had similar expectations for other tools that, over time, changed their focus to AI, analytics, cloud, and so on. Search tools should remain simple.
DuckDB as a small, shareable catalog https://duckdb.org/2026/10/07/view-only-mode
"A DuckDB database file does not have to contain any data. It can store nothing but view definitions that point at files stored somewhere else, such as Parquet on object storage. The file stays a few hundred kilobytes no matter how large the data it describes, and anyone who attaches it sees a set of named relations that are ready to query."
The fake Umberto Eco created with AI: a performance that fooled intellectuals https://www.finestresullarte.info/en/contemporary-art/the-fake-umberto-eco-created-with-ai-a-performance-that-fooled-intellectuals
I've had this bad idea for years, but Service Workers have always been black magic to me. So I asked Claude for help and we built swiiifz¹: a static #IIIF served from a single zip file by a Service Worker, no server needed, and viewed in Triiiceratops².
Just a proof of concept, nothing to take too seriously.
¹ https://atomotic.github.io/swiiifz/ (code: https://github.com/atomotic/swiiifz)
² https://triiiceratops.org/
Sometimes I need to convert mail archives (mbox or maildir) to HTML, and I usually struggle with old tools like mhonarc or HyperMail. HyperMail-rs is a new Rust rewrite. It has bugs, but the authors fix them soon.
This + Pagefind is safe to adopt to create static and searchable email archives.
Text editors need to be fast. Even if there is no reason to abandon Vim never ever, I am pleased to see that https://gram-editor.com/ is really fast (it's a fork of the Zed editor with no AI features). It has replaced Sublime Text as my generic text editor.
As if the experience of the web wasn't already declining these days: YouTube announces a feature for A/B testing of videos. It means different people will see different content. And for those involved in web archiving, it's a disaster.
https://techcrunch.com/2026/09/23/youtube-adds-new-creator-tools-like-video-a-b-testing-dynamic-thumbnails-and-live-dubbing/
I hear a lot of talk about this new Jev¹ AI model. Has anyone in the library or archives world already written about it? I could imagine its use for subject classification or metadata quality assurance. Links or references for further reading are welcome
oh Firefox why, why?
about:config → browser.nova.enabled=false
good design pattern for archives 🤩 https://artbase.rhizome.org/artists
big wall of hyperlinks, is often better than a search input or a facets explorer
RE: https://mastodon.online/@nikitonsky/117285681651599706
YES! I love this kind of AI. This shows respect to people using the terminal. Please, avoid also capital letters
"Apple Reference Image builds on Apple's unique foundation of capabilities in hardware and software, including sensor identity certification at the factory, silicon security, and Private Cloud Compute, giving photographers a new way to provide a verifiable photograph"
https://security.apple.com/blog/apple-reference-image/
"When a user decides to "develop" the negative for verification, it is sent unprocessed to Private Cloud Compute."
https://www.macrumors.com/2026/09/15/apple-reference-image-info/
C2PA is already dead?
L’Internet Society Italia esprime forte preoccupazione per la disattivazione del dominio autistici.org da parte del Public Interest Registry (PIR), conseguente alla designazione di Autistici/Inventati da parte del governo statunitense come Specially Designated Global Terrorist (SDGT).
https://www.isoc.it/la-disattivazione-di-autisticiorg-minaccia-la-natura-globale-e-libera-di-internet
Static ActivityPub Publishing
https://socialwebfoundation.org/2026/09/04/static-activitypub-publishing/static-activitypub-publishing/
L’auto-biblioteca di Ettore Fabietti https://mastodon.bida.im/@abo/117148976997783915
I think he is fully retired now, but a decade ago in Basilicata, Southern Italy, schoolteacher Antonio La Cava (not a librarian) distributed books using his three-wheeled vehicle ↓
Bibliomotocarro, is this Italy's smallest library?
https://www.bbc.com/reel/video/p06l4h3h/watch
"The idea is simple: Mozilla builds the browser people trust, and partners like Exa.ai bring the specialized pieces that make it better."
https://blog.mozilla.org/en/firefox/firefox-exa-partnership/
But perhaps people will no longer trust you at some point.
RE: https://nfdi.social/@toolregistry/117120571271056863
Yes, unfortunately we have heard about it, and perhaps it does not belong in the domain of digital humanities.
A thread discussing the need for Mastodon archival services ↓
https://cosocial.ca/@evan/117113173208161488
my highlight:
"A lot of people don't want this, though. They want to exercise their right to be forgotten. Or they want to enjoy the ephemerality of temporary posts. They don't want their data stored in a system they don't have direct control of."
Mastodon offers robust APIs for exporting content
I am using this simple script, daily, to keep an archive in DuckDB
https://github.com/olithissen/masto.sh