Thread

Article header

Why it's not Annaxandria

Why we're putting books on Nostr, and not where they already are.

We've had npubs regularly suggest that #Alexandria is superfluous because of file torrents. Just put up a website with a search engine for Pirate Bay, and you're done.

Anyone can vibe-code that in 2 minutes, so why spend 2 years torturing ourselves ?

There are a few reasons, why we didn't do this:

  1. Someone already did it. It's called Anna's Archive. Most of what they list is open-access, but they also knowingly list copyrighted material, so they are stuck in a perpetual cat-and-mouse game with the domain registry and the NSA. No, thanks.
  2. Everything that isn't under copyright — and everything that is — can already be viewed and downloaded from a website, as a file. The world didn't need another one of those. (Although we also provide a download utility, we are working to make it obsolete). The world needed someone to solve the problems those download-sites haven't solved.
  3. Torrents are (since v2.0) hashed, but not signed. That means you can confirm that one torrent has the same content as another, but you can't determine the publisher. That means that you are reliant on centralized index repositories (such as Pirate Bay 🏴‍☠), to know if you have the correct, complete file, or just some garbage.
  4. Files are a form of document archival, unlike highly-accessible events. Saving a book to a file essentially takes it out of active use. It's the equivalent of putting it in a cardboard box and storing it in the attic verses leaving it on the nightstand. The book on the nightstand is the one you talk about, scribble in, screenshot and share on social media. Everything else is just collecting dust. Keep all books you care about on your e-nightstand.
  5. Everything should be atomized and just stay that way. There is no benefit to splitting something large up, according to an index, transporting it in pieces, and then reassembling it according to the index, to use it. Just leave the books in pieces. Store it in pieces. Comment on it, in pieces. Highlight parts of pieces. Share pieces. Analyze and search for pieces. Rearrange pieces within and remix pieces between indexes.
  6. If you want something highly-decentralized and censorship-resistent, utilize a protocol people use to communicate about mundane things, and that isn't bound to any particular channel. Because Nostr events are signed and therefore verifiable in isolation, you can transport them over any medium and in any constellation: http, udp, cd-rom, tor, meshastic, radio, Bluetooth, e-mail, SMS, carrier pigeons... or torrents. And you can determine what the precise content of that transport is.
  7. Unlike BitTorrent, which has been around a very long time, Nostr was designed by people who understood how Bitcoin works. They baked crytography, atomization, latent/passive distribution, decentralization, push-communication, and censorship-resistence right into NIP-01. It's all in there and doesn't have to be messily glued-on to the protocol, after the fact.

https://media.tenor.com/lcO2YMmnwEIAAAAC/all-i-have-to-say-thats-all.gif

Replies (7)

🛡️
Are you only doing books or are there plans to do video too.? This is what nostr needs to make a decentralised YouTube, (or really any video files because nostr servers usually lag with video). Clients could pull the little chunks like torrents do from whatever relays they can get a fast enough connection to.
🛡️
I get that Alexandria isn’t working on video, but the model is still highly relevant if nostr wants to improve on YouTube rather than just replicate it. Today video on Nostr via Blossom is effectively “download a big file and play it”, which works but has real performance and UX limits compared to modern streaming. An Alexandria-style indexing layer could describe verifiable sections, chapters, and highlights, or even reference streaming segments, without touching transport. That enables things YouTube doesn’t do well: signed, shareable clips without re-uploading, first class chapters, and protocol level integration with streaming/torrenting so clients fetch only what they need. It also turns highlights into durable objects that can be linked, quoted, and embedded across apps.
We actually have video-books, from @liminal 🦠 , already. That doesn't actually need any new spec, just a "type" tag with "video" or "audio" in it. We were planning on using this for audio books, so we kept the content flexible. I haven't developed a module to seemlessly merge the audiobooks, yet. The files are on media servers. I only embed them when exporting, as I don't want to turn relays into de facto media servers.
When you're dealing in plain text, you're only limited by how you render the text. We use asciidoc in combination with other libraries to show everything from from text to PlantUML diagrams to music notation. Here's an article that combines an embedded youtube video with a readable transformation of the transcript. https://alexandria.gitcitadel.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 For videos, there's nothing wrong with literally slicing up videos and placing them in sections. With how these books can be organized, you can get both books and movies that branch out and connect to other ideas, quotable, highlightable, etc. And this is just what we have now - we have a lot more planned in the future. There are many projects that deal with structuring videos and how to navigate them. We're inventing something that works in parallel, rather than reinventing things experts are already working on.
🛡️
My excitement jumped to thinking of other applications. I should say, great work. I'm really excited about what you are doing at project Alexandria.