Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is everyone's problem with RSS, filtering through the hose pipe of information. Some of the feeds you want everything but a lot of the news sites or popular ones you want a cut down version that caters to your interests since few sites filter on their end and even then it could be poor cut up. It needs an easy way to do that, a basic Bayesian filter trained with down votes would be a drastic improvement over none at all.

You can definitely build a better RSS reader and a big part of that is going to be tailoring so its not just a hosepipe of everything. I use FreshRSS every day and it supports filtering but its a pain to use and its not what I am looking for really it has no intelligence. I read maybe 1/10th of the articles it presents and I have cut feeds where the signal to noise is poor for my interests. This is definitely an area worth improving.



A simple algorithm is to sort content by the signal-to-noise ratio (SNR) of the source feed.

Basically, the more you "liked" past content from a given feed - the more priority is given to new content from that feed.

A more detailed description of the algorithm I use: https://linklonk.com/item/3292763817660940288

As I understand the OP system does not collect/use your content ratings, only external upvote counts. But this is something I think they should try.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: