RSS Feed

[danrobi11]

Can seedit make uses of a RSS Feed or does it already? If so, whats the RSS format?

[estebanabaroa]

the way a community publishes new posts is via IPNS / IPFS, for example this is the IPNS for the community "business-and-finance"

https://k51qzi5uqu5dktu25c6qau4i6tfftjeopmean34ca1db8t6gorp09xbaql0mjc.ipns.ipfsgateway.xyz

inside in has the community metadata like the title, rules, etc

it also has the first page of posts, and the CIDs for the next pages of posts.

everything is as JSON, not in RSS format, so I don't think it's possible at the moment for RSS readers to read it.

the community owner could technically publish their posts as RSS, but where? to some server? to IPFS?

and then RSS readers would point to an IPFS gateway? What would be the most censorship resistant way to do this (a single gateway could censor, it's not p2p)? what RSS readers are popular for people to use nowadays and what functionality do they support?

I don't think anybody on the team uses RSS that much so we would need someone to help us think of the best way to do it.

edit: another issue with gateways is that if you just read data raw from a gateway, it doesn't validate the protocol, it doesn't validate signatures and invalid fields, so you could read and present fake data to your users. So I don't think gateways could be used that way, it would be very insecure, you wouldn't know if any of the posts are signed by the actual author of the post.

[kuchta]

I don't know how publishing in Plebbit works, but from my limited understanding of what you just said, simplest way how it could be implemented is to generate RSS in the same way as any other frontend content the user requests regardless of what frontend instance he is currently using. So if someone is using seedit.app than it should point under seedit.app.

[estebanabaroa]

Isn't community owner validating post/comment author signatures before it publishes

yes, but the reader must also protect himself from a malicious community owner. the client must protect the reader from a malicious community owner and from a malicious gateway, both can tamper with content. a full plebbit client does this, but an RSS client can't really do this, it just trusts that whatever the server endpoint sends is valid.

because there's always some form of trust

in the plebbit client there isn't, the gateways can't like to you, a community owner cant lie to you about a user having written a specific comment or not. the client downloads the content and verifies the signatures before presenting them to you.

RSS is not a plebbit client and never will be

yes since the RSS model is that you just trust the source server, it's probably better to have an indexer website like plebscan.org (which validates content and signatures before indexing) as the source, if you have an untrusted gateway as the source, the community owner or gateway can lie, whereas if you have plebscan.org as the source, it's more trustworthy in a way, because it's a web2 service with some reputation, it would be weird and unecessary for them to start tampering with posts.

user's are expected to go to the source anyway...

yes that's good, the source could be put as https://seedit.app/p//c/, so if the user visit it, it would display an error if the community owner or gateway lied.

[kuchta]

because there's always some form of trust

in the plebbit client there isn't, the gateways can't like to you, a community owner cant lie to you about a user having written a specific comment or not. the client downloads the content and verifies the signatures before presenting them to you.

Yes, but the plebbit client itself CAN lie to you if it is controlled by someone with malicious intent, or if got hacked, or if the user obtains it from an unofficial source etc...

RSS is not a plebbit client and never will be

yes since the RSS model is that you just trust the source server, it's probably better to have an indexer website like plebscan.org (which validates content and signatures before indexing) as the source, if you have an untrusted gateway as the source, the community owner or gateway can lie, whereas if you have plebscan.org as the source, it's more trustworthy in a way, because it's a web2 service with some reputation, it would be weird and unecessary for them to start tampering with posts.

I beg to disagree. Users or RSS service providers can get around the problem of malicious gateways more easily, than malicous services like plebscan.org, since they can use trusted gateways or even run their own gateways.

[estebanabaroa]

I beg to disagree. Users or RSS service providers can get around the problem of malicious gateways more easily, than malicous services like plebscan.org, since they can use trusted gateways or even run their own gateways.

I think not many people run gateways, it takes a lot of bandwidth just to run the DHT, even if you almost never use it, and "trusted gateways" are only verified by a special client that implements the verification, an RSS reader doesn't do this unless it has special code specifically for it. Do RSS readers have this implemented? If not a trusted gateway is just a regular gateway.

The plebbit client is basically a trusted gateway, it verifies data from the gateway before it lets you consume it.

[kuchta]

I think not many people in general use decentralized services either, but those who do will be much more likely to run their own ipfs node, which is the most trusted gateway you can get and if you use hosted RSS reader service, then the service operator can run their own ipfs node, which could easily became differentiating factor in the near, more decentralized future. Plebbit client may be regarded as a "trusted gateway", but unless you implement RSS reader functionality, it wouldn't be general enough for consuming all news from one place, which is a quite nice idea for a new project, IMHO...

[estebanabaroa]

I think not many people in general use decentralized services either, but those who do will be much more likely to run their own ipfs node

I think most people who use decentralized services don't run IPFS nodes, because it's too difficult, uses too much bandwidth, etc.

For example I use bittorrent all the time, and I use bitcoin / eth all the time, but the bittorrent DHT is lightweight and bitcoin / eth have public RPCs (is gateways) that are trustless (through the bitcoin / eth client wallets).

So the plan with plebbit is that the native client is light like bittorrent, and that the web client verifiy the public RPCs / gateways, so that there's no trust involved. This would allow people to use plebbit easily, as easily as bittorrent or bitcoin / eth, those products have 100s of millions of users.

I guess another scenario that we didn't talk about, for the "hardcore" user of plebbit and RSS, would be, the user runs a plebbit native client, the native client downloads the data P2P, verifies it, and then publishes a localhost RSS endpoint that the RSS reader can read. I think that scenario is more probable than a plebbit user running a full IPFS node, I think almost no plebbit user will ever run a full IPFS node, as it uses too much bandwidth. The plebbit client doesn't actually run a full IPFS node, it doesn't run the IPFS DHT, because it used too much bandwidth, so we disable it, and we use trackers instead for peer discovery (like bittorrent), eventually we plan on having fallback peer discovery mechanisms like different forms of peer exchange, and fallback to a lighter kind of DHT, more similar to the bittorrent DHT.

And maybe the localhost plebbit RPC could be mirrored by certain services, like plebscan.org

I know in theory it would be really cool if you could just run any IPFS node/gateway and then have access to plebbit, but in practice I'm not sure it's doable, plebbit is a protocol above IPFS, like HTTP is a protocol above TCP, if you tried to read HTTP but you client only implemented TCP, it wouldn't really work that well.

[danrobi11]

A few years ago, I was using galactech. There's was a way to get RSS notifications from standard RSS reader. The URL used to look like so: https://bafybeigah4nqxqyuyy74wnnaq45gf636upc7k5fnfb3mfgsm6vmoo4umk4.ipfs.dweb.link/#/feed

[estebanabaroa]

A few years ago, I was using galactech. There's was a way to get RSS notifications from standard RSS reader. The URL used to look like so: https://bafybeigah4nqxqyuyy74wnnaq45gf636upc7k5fnfb3mfgsm6vmoo4umk4.ipfs.dweb.link/#/feed

one issue I forgot to mention that is problematic about gateways is that if you just read data raw from a gateway, it doesn't validate the protocol, it doesn't validate signatures and invalid fields, so you could read and present fake data to your users. So I don't think gateways could be used that way, it would be very insecure, you wouldn't know if any of the posts are signed by the actual author of the post.

the only way to do it would be for the RSS client to have some plugin that would validate the data, verify signatures, etc, or to trust an intermediary that validates it for you, but then they can censor you. for example we have https://plebscan.org/ which is a search engine and indexer for plebbit. we could add RSS to this website, but the owner of plebscan.org lives in europe and is legally required to censor many posts.

if you think adding RSS to plebscan.org would be useful, let me know I can ask the owner to implement it. since plebscan is also a search engine, you could also subscribe to a specific keyword RSS feed.

also another interesting thing is plebscan is opensource and anyone can use it to run an indexer.

does anyone run RSS readers locally? we could potentially add RSS feeds to the CLI plebbit node client, plebbit-cli https://github.com/plebbit/plebbit-cli this way the user could run a node as a daemon, fetch the data P2P, and then he could get an RSS feed endpoint locally at http://localhost:9138/rss or something like that. this local RSS feed would be fully censorship resistant and validated.