The risk with URL shorteners is commonly that people get tired running them and they disappear, making the web even more brittle than before (I wrote about goo.gl doing this recently [0]).
Centralization can be great here because before giving up the ghost, ethical operators of shorteners can donate their database mapping short links to full links to archivists. That's what efforts like 301works.org [1]. Are all about.
I can, however, see how decentralization can be helpful because it limits the blast radius of a service shutting down. ATProto is also cool because you can migrate your data to another PDS without explicit coordination of your current one since its baked into the protocol.
Despite that, I can imagine this potentially being the worst of both worlds. You have a centralized domain and resolving service but the data is distributed, so the operators could still unilaterally take down the resolver and the URL. Also, if individuals aren't paying attention, they might not migrate their data off the PDS. My ignorance of ATProto is showing here... can a PDS unilaterally move the data it hosts to another PDS (say, one ran by the archive team?) If so this can be very cool. If not, it sounds more fragile than a traditional url shortener that is prepared to donate its URL mappings.
I think a happy medium is a link shortener that uses https://perma.cc/ for resolving and generating permanent short links (Internet Archive is a supporting partner).
> You give Perma.cc the URL of the page you want to preserve and cite. Our software visits that URL, preserves what's there, deposits it into our collection, and gives you a unique URL (e.g. "perma.cc/ABCD-1234") - a "Perma Link" - that points to the record in our collection. You then can use that Perma Link in your citation to give readers access to a stable, accurate record of the source you referenced, even if the original disappears from the web.
(TLDR You're building an OpenURL resolver on social rails with sharing capability, with functionality to fall back to generating your own OpenURL when a permanent ident does not yet exist)
Super cool idea, not sure how I feel about this.
The risk with URL shorteners is commonly that people get tired running them and they disappear, making the web even more brittle than before (I wrote about goo.gl doing this recently [0]).
Centralization can be great here because before giving up the ghost, ethical operators of shorteners can donate their database mapping short links to full links to archivists. That's what efforts like 301works.org [1]. Are all about.
I can, however, see how decentralization can be helpful because it limits the blast radius of a service shutting down. ATProto is also cool because you can migrate your data to another PDS without explicit coordination of your current one since its baked into the protocol.
Despite that, I can imagine this potentially being the worst of both worlds. You have a centralized domain and resolving service but the data is distributed, so the operators could still unilaterally take down the resolver and the URL. Also, if individuals aren't paying attention, they might not migrate their data off the PDS. My ignorance of ATProto is showing here... can a PDS unilaterally move the data it hosts to another PDS (say, one ran by the archive team?) If so this can be very cool. If not, it sounds more fragile than a traditional url shortener that is prepared to donate its URL mappings.
[0] https://jszym.com/blog/archiving_googl/
[1] https://archive.org/details/301works
I think a happy medium is a link shortener that uses https://perma.cc/ for resolving and generating permanent short links (Internet Archive is a supporting partner).
> You give Perma.cc the URL of the page you want to preserve and cite. Our software visits that URL, preserves what's there, deposits it into our collection, and gives you a unique URL (e.g. "perma.cc/ABCD-1234") - a "Perma Link" - that points to the record in our collection. You then can use that Perma Link in your citation to give readers access to a stable, accurate record of the source you referenced, even if the original disappears from the web.
Related:
Cool URIs Don't Change (1998) - https://news.ycombinator.com/item?id=49231809 - August 2026 (69 comments)
https://en.wikipedia.org/wiki/OpenURL
https://arks.org/
https://doi.org/
https://specs.ipfs.tech/ipfs-uri/
(TLDR You're building an OpenURL resolver on social rails with sharing capability, with functionality to fall back to generating your own OpenURL when a permanent ident does not yet exist)