Currently I have to write some code for every new source, partly for the extraction of articles but also to add it to the list of customizable ones etc. Unless they were limited to a standard format then it's hard to add new sources without getting stuck your hands wet. I am trying to add new sources, I added Computerphile this morning, I'll note down those two.
As far as duplication is concerned - yes, I've even had a go at implementing it. I was using the jaccard index to compare stories within the last 2 days. In the end I found it was hard to be confident enough about the match. I don't want to miss stories and so personally just put up with the noise.
Comments
I love this setup, especially how you used #tweets as a unified ranking system and can sync read pages across devices.
Is there a way to add custom sources? For example: http://www.datatau.com/, https://soylentnews.org/ or any number of other aggregation sites
Also, sometimes Ars and HN will cover the same story or HN will link Ars; have you considered a duplicate detection system?
Currently I have to write some code for every new source, partly for the extraction of articles but also to add it to the list of customizable ones etc. Unless they were limited to a standard format then it's hard to add new sources without getting stuck your hands wet. I am trying to add new sources, I added Computerphile this morning, I'll note down those two.
As far as duplication is concerned - yes, I've even had a go at implementing it. I was using the jaccard index to compare stories within the last 2 days. In the end I found it was hard to be confident enough about the match. I don't want to miss stories and so personally just put up with the noise.