IRC channel logs

2024-01-09.log

back to list of logs

<mala>jfred, I remember where I've seen willowprotocol before -- Iroh, an IPFS cousin, implements it
<mala> https://iroh.computer/
<jfred>oh interesting, I hadn't heard of this
<jfred>cwebber: Propagator question for you when you get the chance. I was just checking out Brainy and my first thought was to try and build a D&D character sheet that would calculate the various dependent stats from the character's base stats. But it seems like once a pcell has taken a value, that value is fixed. Is this inherent to propagators/how would you deal with changing information?
<jfred>(And I know Brainy is very experimental, etc - if this is something that propagators can potentially handle but Brainy can't at the moment, no worries!)
<cwebber>jfred: aha! You are correct... At the moment. See Alexey Radul's dissertation, it shows how to wire up functional reactive programming to propagators
<cwebber>The trick is to treat time based inputs as a partial information type
<cwebber>Similar to uncertainty intervals
<cwebber>Newer timed values replace older information
<cwebber>Sadly I gave gotten this far yer
<cwebber>Yet
<civodul>cwebber: hey, what intro paper to propagators would you recommend?
<cwebber>I am mid TMS implementation
<cwebber>civodul: Alexey Radul's dissertation, 100%
<cwebber> https://dspace.mit.edu/handle/1721.1/54635
<cwebber>And my talk is, I think, a reasonable fast intro to watch first
<cwebber> https://m.youtube.com/watch?v=jI8gA68OXLM
<civodul>awesome, thanks!
<cwebber>Then if you want a reference
<cwebber> https://groups.csail.mit.edu/mac/users/gjs/propagators/
<civodul>(and yes, i’m aware of your talk too, though i’d like to sit down and read, behind the window, in winter, you know)
<civodul>thanks for the references!
<cwebber>Brainy uses syntax closer to the revised report over Alexey's dissertation but is not faithful in full to either
<cwebber>Np
<cwebber>It's one of my favorite reads ever
<cwebber>And I have returned to it many times
<cwebber>Please let me know what you think of it!
<civodul>sure
<civodul>is there a PDF of the revised report?…
<cwebber>No
<cwebber>There is of Art of the Propagator though
<cwebber>Which is shorter than Alexey's dissertation, but not as thoroughly illuminating
<cwebber> https://dspace.mit.edu/handle/1721.1/44215
<civodul>nice
<civodul>the dissertation is more intimidating :-)
<cwebber>it's much longer it's true
<jfred>cwebber: aha, that makes sense! will have to go back and read that dissertation, I've read through it some before but helps to have more context from trying things hands-on
<cwebber>:)
<alanz>random question: Is Veilid being considered for transport at any time?
<dthompson>alanz: nothing in the works but the idea has been brought up!
<alanz>Nice. I have a sense it is going to become a standard connection layer for a lot of things one day
<dthompson>the ocapn protocol has the notion of "netlayers"
<dthompson>veilid could be one such netlayer
<alanz>yes, I was just reading the 0.12 NEWS and that made me think of it
<cwebber>:)
<Zarutian_iPad>ACTION looks at https://veilid.com/docs/
<cwebber>the Veilid main author and I have talked and we both expressed that a Veilid netlayer would be a cool idea
<Zarutian_iPad>I am trying to recall what sort of overlay net veilid is
<Zarutian_iPad>is it like tor and i2p or more akin to fast anonymous mail/message relays
<Zarutian_iPad>remailers like Nym uses (see https://blog.nymtech.net/sphinx-tl-dr-the-data-packet-that-can-anonymize-bitcoin-and-the-internet-18d152c6e4dc )
<cwebber>the scope of Veilid is a bit broad, it handles both a content-addressed use case and a connection-oriented use case per my understanding
<cwebber>> The framework is conceptually similar to IPFS and Tor, but faster and designed from the ground-up to provide all services over a privately routed network.
<cwebber> https://veilid.com/docs/overview/