Re:It's time to head back to RSS?
by cyberpip on Saturday March 31, 2018 @02:01AM
(#16721981)
The major feature here would be deduplication, in my opinion. That would be hard to do though as it would have to parse the feed contents (which vary from links to articles to full blow articles sometimes) and use some kind of percentage to determine whether or not it's a dupe? Curious if there is a smarter solution to that feature possibility.
[ Parent
| on HN ]
Re:It's time to head back to RSS?
by forapurpose on Saturday March 31, 2018 @03:33PM
(#16724756)
> use some kind of percentage to determine whether or not it's a dupe
That's the problem I intended to solve by 'grouping', as I called it: Don't delete the similar items, but group them together (see my GP comment regarding the functionality).
Also, it might not be as hard as it seems with this data. There would be maybe thousands of records (though you'd want it to scale higher); how many would have the same sets of proper nouns, for example, and not be about the same topic?
FYI Inoreader does some deduplication by default. As you scroll you see a message like 'Filtered one similar article'. I haven't noticed any duplicates. That doesn't help when everyone's talking about the same thing, of course.