1. 28

This is sort of a Request for Comments. I’m creating a news portal that aims to be light, open, versioned via github and readable in html, xml, json and markdown using a free api. It aims to declutter current news and put them in relation to previous events - each news item has a list of related stories. Data is based on wikipedia current events. Backend static generator is written in Perl.

  1.  

  2. 12

    Comments:

    • Cool idea.
    • Thank you for not loading a bunch of unrelated crap and for respecting our privacy. :)
    • You don’t show the backend code anywhere on your Github it seems. Like, the Perl scripts you use to munge data don’t seem to be in your public repos.
    • You posted a project here without previously engaging in our community and without showing the code that makes it work (as opposed to the markup that code generates), and your Twitter is mostly just reposts of reposts. So, please stay around a while and do more than just drive-by your project! Comment on other stuff too. :)
    1. 8

      Thanks for the comments:

      • I wanted the Context to be crap-free, cool you noticed :)
      • So far, the Perl scripts are on my private repo, I want to see how this project evolves. I first focus on releasing the crunched data publicly.
      • I was a daily reader for a long while now, but got the invite only today. I plan to engage more from now on. My Twitter is more like my bookmark list, but I do reply quickly.
    2. 8

      The general idea seems to be like Google News, but without the tracking and the Google and images. I like that and I made a similar site in german.

      1. 2

        This looks cool! How do you aggregate news links from different sources? A combined feed?

        1. 5

          It is essentially a feed reader with manual weights for ordering.

          1. 2

            Nice idea with weighting by tag and source quality, I didn’t think of it like that.

      2. 2

        At least WikiNews could be an additional source.

        In general, you only provide links with the headlines as text. That should be fine for all the other sources as well.

        1. 2

          Yes, it should. But I had a problem with correct filtering of crappy news or clickbait headlines. With your weighting approach it would be easier. Additionally, for context cross-reference, I’m relying on wikipedia keywording for now, but with additional sources I would have to implement keywording based on, for example, parts-of-speech dissection (verbs).