Yeah, parsing & cleaning up HTML will always be a heuristic that fails sometimes.
I'd highlight your point nr. 5. A very large percentage of articles use Wordpress or other website builders. Once you support a few large sites your coverage drastically increases.
With Unclutter I also found it helpful to have an automated fallback -- if the text content of the page reduces too much, it disables some of the content block methods one by one.
I'd highlight your point nr. 5. A very large percentage of articles use Wordpress or other website builders. Once you support a few large sites your coverage drastically increases.
With Unclutter I also found it helpful to have an automated fallback -- if the text content of the page reduces too much, it disables some of the content block methods one by one.