Ask HN: Do you still use RSS?

Yep, every day. I do notice that it’s getting worse, though. Many websites’ feeds are just broken these days. They’ll randomly refresh after months of inactivity with dozens of new articles from the intervening period; or they’ll re-list a slew of previously listed entries; or it’ll be entirely random whether a new entry shows up when a new item is posted on the website, such that I can only take an entry as a sort of “unread badge reminder” for the website rather than taking each RSS entry in my reader as an individual article I should read.

I think some of this is the fault of bad RSS feeds being generated by bad CMS software that never gets its RSS path tested; but other parts of this are the fault of the RSS feed getting fronted by a CDN like Cloudflare, or the URL pulling the feed in through some sort of PuSH hub with non-compliant retrieval semantics, or etc.

It’s gotten me to feeling that I should be writing software to synthesize my own RSS feeds from these websites using scraping, with the RSS feed itself just being an edge-triggered “no longer fresh” trigger to get the site put back into the scrape queue. Most of the sites I subscribe to have very simple linear indices available somewhere anyway, so scraping that into a feed really shouldn’t be that hard. (I remember when Dapper was trying to do this, but Dapper was trying to solve the problem in full generality for even cases where the site generates “changes” without creating new permalinked entry URLs, which is hardly ever a problem any more.)

Yes. I can't even imagine using the hundreds of different ways of "subscribing" to things every site feels the need to reinvent for itself. YouTube is a perfect example, using their subscribe method when will you actually know when a subscribed channel posts a new video? Same day? Week later? When it hits a certain view count? That's not up to the user.

After Google killed Reader, I switched to theoldreader.com and then inoreader. Very clean interface, the Android app rocks. I'm still too cheap to go on a paid plan, but even so the only limitation I see is the number of total feeds. Then again, I'm not sure that's a bad thing considering I have way too many feeds as it so having the possibility to add any more would only worsen the problem.

https://feedly.com/ works perfectly as a Google Reader replacement backend (i.e. an “RSS subscription and read-state-tracking” service.)

I don’t like Feedly’s own web-app frontend, but you don’t need to use it; there are many nice, fancy RSS-reader clients implemented as native apps on various OSes, that can use Feedly as their backing service, sort of like a native mail client using IMAP.

Personally, I use Reeder on macOS and iOS, but there are at least a dozen such clients that connect to Feedly.

I should also mention, if it’s the backend features that you have a problem with, that there’s at least one piece of RSS server software, https://feedafever.com/, that all these clients also natively support (or at least supported; the product is dead now, so some clients are sadly removing support.) I mention it because it was a host-it-yourself solution, and therefore, even though it was a closed-source product, it had a clearly-defined client-server protocol and clients supporting Fever servers also necessarily have/had a “server API endpoint URL” field in their Fever service configuration. This means that it’s possible to implement a Fever-protocol server yourself, and any client still supporting the Fever “service” will then sync their subscriptions and read-states against your backend. So you can actually have whatever backend features you can think of, if you really care to.

(Tangent: why has nobody written an RFC for a standard RSS client-to-service sync protocol? Fever’s protocol is pretty crap, honestly; it used very nonstandard names, aggregated events into messages at the wrong levels, etc. There’s no reason that all these RSS sync-backend services need their own custom APIs; they all have exactly the same data model. Though, I guess, when you start to think about it, you realize that there’s no reason that these backends shouldn’t be speaking IMAP to their clients...)

Yes. I've read about this thread in selfoss (https://github.com/SSilence/selfoss), "a multipurpose RSS reader and feed aggregation web application."