Jack VanDrunen

Screenshot of Isaac's blog listed on longform.fyi

There it is! Btw, I really like the idea of borrowing @pckt.blog's content format behind the scenes. Might do that myself

I've patched the error on my end! Things are still backfilling, but Ramblings should appear at some point tonight...

Standard.site Validator A friendly tool to check if a post has valid standard.site records

My validator only handles one case (where the url + /.well-known/site.standard.publication resolves to the publication), so I assume yours is allowed

Is this the validator you used? site-validator.fly.dev

Oh good, I'll add this case to my validator

I don't see a /.well-known/site.standard.publication endpoint, either at /ramblings or at your domain root

The most promising path for Christianity to regain legitimacy will be through respecting the expertise of scientists.

https://metapol.is/intelligent-design-or-intellectual-humility/

The floodgates have opened! My @standard.site topical discovery tool longform.fyi now indexes @leaflet.pub publications. The main requirement now is having at least 5 English-language posts. Categorization itself is still all over the place... apologies to those who've been misclassified...

oh sorry, I meant that the 30 categories are pre-defined, stolen from the categories that explore.ghost.org has. Only the non-empty categories are displayed on longform.fyi itself

This is all very rough and ready, I started with gensim which I knew from my computer science days, but found that SentenceTransformers works a lot better

Cosine similarity between the mean of the 5 most recent post vectors, and the category vectors, similar to the quickstart: www.sbert.net/docs/quickst.... Category vectors are taken from descriptions that I just had AI generate: gist.github.com/jackvandrune...
(30 categories taken from Ghost Explore)

Right, of course. I see now. It's unfortunate that it works that way, especially for static sites where the deploy time is probably more likely to be delayed...

Time to fix my validator I guess.

Published records should be immediately verifiable #13 The recommended workflow in the Quickstart has publishing to PDS come before verification by deploying the static site. But if an AppView is watching the firehose, it might try to validate the public...

@stevedylan.dev I think this is a potential problem for Sequoia too. Opened a PR: tangled.org/stevedylan.d...

Embarrassing when you realize that your own @standard.site aggregators are dropping your own posts.

Records really need to be validate-able before being pushed to the PDS, or they might be thrown out by AppViews that watch the firehose.

What would it mean for theology if natural laws could explain the development of complex life?

https://metapol.is/a-sterile-universe/

I need to collect more debug stats when backfilling. Probably I'm unjustly throwing out documents somewhere along the line

It does look sparse... Criteria are at least 5 English-language posts with textContent fields attached to a publication with showInDiscovery true (or not set)

Should be fully backfilled at this point. Notably absent are @leaflet.pub publications, because they don't implement textContent (yet...)

Spent the last few days working on another proof of concept for @standard.site topic-based discovery: longform.fyi

Possible categories are taken from the Ghost discovery feed, and categorization uses SentenceTransformers. I assume that eventually this will be done in-house by most AppViews

Considering the lack of replies I’m going to assume no. Have at it!

Would love to get in touch! Shoot me an email? (contact info on my website)

I'm holding out hopes for @offprint.app as (eventually) a good drop-in replacement

This is basically my hope for the main sermon.garden domain - I'd love to build something like an AppView that functions like an ATproto SermonAudio alternative. But I'm way too busy to build and maintain a "real" app at this point.

bc standard.site validation requires the well-known URI and the metadata in documents, I threw together a script (not even a proper plugin) that scrapes my site and produces the mirror for validation. Wanted to take advantage of SS discovery asap. Not sure what my long-term solution will be...

As @standard.site adoption grows, the future will hopefully be in the direction of topical curation (and algorithmic discovery). "Global discovery feeds" only work as long as the network is small. (Also, global discovery feeds have a lot of "testing testing, my first @offprint.app post" right now 🤭)

Has anybody made an open @semble.so collection for examples of the atmosphere *working*?

Cool — I think one of the first @standard.site aggregators I've seen with niche topical curation, not just global index/search!

👀✍️ theologians, exegetes, sermonologists…check it out!

Hopefully discoverable using any feed reader: longform.sermon.garden/feed.xml

I'll add some more details to the header when I get a chance

The reason decentralisation matters for something like age verification isn’t just “run your own server,” as Hodgson acknowledges that server admins are still subject to local law. It’s that when your social life isn’t bundled into one platform, no single company’s policy change can disrupt all of your communication, community, and content at once. What we’re finding is that for decentralisation to really make an impact, it needs to happen on multiple axes at the same time. There is the decentralisation in the way it is usually understood by communities on ActivityPub and Matrix: from a single centralised server to many decentralised servers run by independent groups. This gives communities autonomy over their own spaces, but each server still replicates the same software and feature set. There is the decentralisation in the way it is done on atproto: from a single software stack to separating identity, data storage and apps. This means your identity and data aren’t locked to any one application, and different apps can offer different experiences on top of the same underlying infrastructure. And there is a third axis that is now starting to become visible: the decentralisation of features. Rather than a single app that bundles everything together, like Discord, multiple different apps each specialise in a few things and are interoperable with each other. This is the axis that the developments this week are starting to illustrate, and it may be the one that ultimately matters most for resilience against the kind of platform-wide policy changes that sparked this conversation in the first place.

new from me: FR#153 – What does a Discord replacement look like?

connectedplaces.online/reports/fr15...

Thanks! For now, just reaching out to me. The feed tracks site.standard.document records that backlink to the curated site.standard.publication records, so the publication needs to be fully @standard.site compatible. Happy to hear any suggestions. (There's an RSS feed too!)

Created a @standard.site aggregator for biblical studies publications. So far, @joelwerner.me and I are the only active authors I've found using compatible platforms. More are welcome!

longform.sermon.garden