Combine multiple RSS feeds into a single feed, as a service
github.com
github.com
This could also allow users to develop/fork the aggregated feeds and share them with a community.
Tbh, I run specialized vertical services aggregating 100's of sources (rss as well as other), and the combination towards a single feed is trivial.
It's a pet peeve of mine seeing software projects that describe what they do, but without providing even a single use case. Or even what motivated the developer.
Why does this thing exist? There is probably a reason and then I could learn something.
Instead of following a bunch of them and having to sync them I guess you could just follow the combined one and be done with it. If you have multiple feed readers, each with their own way to input feeds, I think something like rsscombine might be an easy solution.
Being able to subscribe to "Debian blogs" in one step is helpful versus discovering each blogger individually.
There's a big list here [2], I think two use the linked code.
[1] https://indieweb.org/planet
[2] https://github.com/ArchiveTeam/urls-sources/blob/master/4320...
I use FreshRSS and you just put all the feeds in one category and then you can click on that category to show all the feeds together.
Frankly I’d go a bit further and spit out events, so people can code “give me the next entry” without having to mess with polling lots of rss endpoints. Just tell me when there’s a new entry.
Or turn it into a web hook so you just code the parsing, and delivery is outsourced.
Polling for changes is a terrible design. The source has to serve data on repeat to every reader that is set up to check once every week/day/hour/minute.
At the very least have a distributed cache that is set to return only the entries created after generally agreed timestamps. (Please tell me that exists now, or why it’s a bad idea.)
Regarding the only-after semantics:
- HTTP has "Delta encoding in HTTP": https://www.rfc-editor.org/rfc/rfc3229 (2002),
- usage for feeds proposed in 2004: https://bobwyman.typepad.com/main/2004/09/using_rfc3229_w.ht...,
- some (not so) recent discussion (2017): https://www.ctrl.blog/entry/feed-delta-updates.html + https://news.ycombinator.com/item?id=26091489 (zero comments).
Adoption (or lack there of) is a different story. I guess the main issue is that it puts more work on the publisher side, in particular requires dynamic aspect when generating the response.
This was much faster than clicking on 20 different feeds to see which ones are oldest or most recent.
How did you solve this problem?
I don't split an existing RSS feed, but I scrape a website, and store the posts in a database, with a column for the user who posted it.
Then I generate an OPML file with URLs like http://localhost:1234?user=... per user, and spin up a web server that fetches only that user's feed given the query parameter. Import the OPML into the news reader, and bingo bango.
https://www.listennotes.com/blog/why-podcasts-are-my-new-wik...