Is there nostr:npub1mkde3807tcyyp2f98re7n7z0va8979atqkfja7avknvwdjg97vpq6ef0jp some filter reducing what is visible in my disjunctive query? For example with the query from the settings:
"Search: #maths OR #math OR #Mathematics OR #CategoryTheory OR #Logic OR #RDF OR #SemWeb OR #scala OR #Agda OR #FP OR #SemanticWeb OR #topos OR #ScalaJS OR #Java OR #www OR #W3C OR #WebScience OR #hyperApp OR #Cognitive OR #CognitiveScience OR #Psychology"
The following page does not appear:
nostr:note17gy2v0qcwj45y6qmlm7jyhn6j2tj2tdt05qe5wrakyej52ydhmjs92ddq6 (not my preferred choice, but as I am trying to debug this functionality...)
i see that note in results when i search for those terms
It would help if, on nostr:npub12vkcxr0luzwp8e673v29eqjhrr7p9vqq8asav85swaepclllj09sylpugg and other #nostrclients, one could have a feed of a disjunction of one's favorite search terms or #tags. I have quite a wide variety of topics that are moving quietly herel, with perhaps 1 post a day for e.g. #scala or #math, but if I could put them all together, I would, I think have a pretty interesting home page.
our search engine supports disjunction (OR) operator in query string. you can use something like this to search for multiple terms or tags:
#scala OR #math
OR has to be upper-case. here is the link to search example:
Artificial General Intelligence — A gentle introduction
well, it's fun experiment for sure 🙂 but i can already see what would be next step, probably personal caching service (for notes from users i personally follow) and some sort of nostr client, running on phone locally. to be able to browse my feed and other stuff.
OTOH lets imagine the future where everybody has local personal relay. i would connect to relays of people in my contact list, people i follow. and that's great for smaller social circles. but not so much for someone who has 100K+ followers (or even less than that). because his phone would melt to service queries from his followers. i think we need a solution where we get notes from highly popular users via our social graph neighbourhood. if my app knows that you are also follower of that highly popular user, i can query your self-hosted caching service, running on your phone, integrated into app or more modular like on android, instead of going directly to his relay. his relay will need to limit how much and to whom his notes are initially seeded. starts to look like bittorrent..
great idea. in the meanwhile i'm running my personal nostr_rs_relay instance accessible as #tor hidden service on my #android phone in #nix-on-droid. nostr:npub12vkcxr0luzwp8e673v29eqjhrr7p9vqq8asav85swaepclllj09sylpugg webapp works in tor browser and with small about:config tweak can connect to it.
we are contemplating support for all major geographical regions.
number of cores depends on how many users you want to service at the same time and if you want to do media caching. try it with at least 4 cores. as for memory, with 16GB RAM it should work fine, 32GB is what i usually suggest as it will have more opportunity to use disk I/O cache and since you probably have some other processes running on the server. for storage, after 6 months and 125mil events, our database is around 330GB. but i could prune older events and reduce storage requirements if you dont want to archive everything forever.
can you try it now? it should be fixed.
can you send me the link that you got after uploading the file? maybe i can figure out what's happening.
perfect opportunity for blues.. "walking hard, to fiat mines.."
🎶
notes and other stuff will be transmitted by various means by then.
this list of relays is the most interesting lately:
https://github.com/PrimalHQ/primal-caching-service/blob/main/relays-mined-from-contact-lists.txt
periodically i run a script to mine relays from all collected contact list events.
https://github.com/PrimalHQ/primal-server/blob/main/collect-relays-from-contact-lists.jl
you probably want to search for notes where either of these two hashtags is mentioned.
to do that you can use search query like this:
#nostrrelay OR #nostrdev
here is the link to it:
nostr:npub1j7uc37l5lzyqfyley4c3ux7cqeshk5y060fgxy3gs5r7gtu2xd5qyc8lvw just told me that upload when writing posts is not supported yet. but it's planned feature for sure. stay tuned!
hi, is it still not working? if so, i would like to reproduce the issue to fix it. can you send me the file you are trying to upload? or link to it if you already uploaded it to some other web host. maybe it's something about the format of the file.
p2p like communication over nostr could be implemented using ephemeral events (NIP-16, kinds 20000-29999). for file sharing it would be actually more like peer to multiple other peers, and since it's going through relay, pRp (peer-relay-peers). bandwith cost could be covered by zapping relay operator. single file transfer could be executed over multiple relays.
there is a reason why internet became centralized. the main cause was inability to assign IP address to every user which prevented direct communication between users and created need for big tech intermediaries. #nostr is a step in the right direction, but it's not complete solution.
if relay operator doesn't want to allow file sharing (s)he can block certain event kinds. although it would be excellent test of survivability of the protocol and nostr network in the long run.
CC: nostr:npub1qqqqqqyz0la2jjl752yv8h7wgs3v098mh9nztd4nr6gynaef6uqqt0n47m
very cool proposal. structured approach to code management is the future. like what we can see with unison language.
there are some worries about sandboxing and isolation but that's something that can be improved later on after initial experiments.
i would also prefer capability to distribute wasm module which is interpreter for other higher level language that will be able to execute code from events containing that higher level language. but the composition is the key even if you use multiple different languages for different aspects the system.
expressive power of javascript/typescript type of language is very limiting when compared to, for example, prolog where the same code can be re-used in many more ways which is valuable if you want far and wide reuse. meta-interpreters and program synthesis are easy in languages like prolog.
also as it looks rn it's very web tech specific (DOM elements etc) but what's beautiful is that rendering is optional. execution environment doesn't have to render anything just execute code-carrying events enough for various other purposes.
if all tools to develop this core-carrying events are also published and distributed via nostr in the same form (after initial boostrapping process) this will enable more axiomatic and broad self-improvement process to emerge with ports to different execution environments and rapid exploration of new application domains.
can't wait to see where this goes..