I don't understand your rating request. What's the use case?
If you rate the item after reading it, you won't want to see it in your incoming stream of new items, so I don't see how you can sort based on ratings.
If you rate the item before reading it, what are you rating it on?
As I said, I don't understand what your use case is.
Sorry about not explaining correctly. I meant: read the article, give it a spam/ham status (or 1 to 5 start rating or ...). Then on the next article(s) (in the near future) based on the words in the already classified article's words (which can be links to specific domains too (e.g. subscribed to some link aggregator, but it sometimes sends spams)) the unread article should be classified as probably ham/spam. And I don't want to see spam below a threshold... In short: Bayesian like learning...
And yeah I can use something like rss2mail and a MUA with spam filter... But I'd prefer a dedicated application (which lives on the desktop but syncs with the cloud, and multiplatform).
- never show me duplicate articles (duplicate is: points to the (almost - yeah, clever) same link)
- I can rate the items (if I want to), than it sorts based on that rating
- then allows filtering of items based on everything (especially rating)...
I'd donate...