I find the idea quite interesting, so interesting that I implemented my own version already over a year ago.
While I also started with aggregating only the top sites and so on, I found it quite "unpleaseant" that it lists all of the news-items in that way. I tried to make another ranking of the news and also tried to get more factors into the formula.
At the end there was a custom crawler, a special API using Solr and a custom Frontend [1] in two versions [2] which aggregates everything together.
I have also open-sourced the frontend [3][4]. If you have fun, have a look into that, but unfortunately the source-code diverged from the frontend Reader [1] a bit, because I started refactoring the db-schema and had not time to update my repositories yet. The second frontend [2] just happend because I was curious about the abilities of Foundation from Zurb. Works really well, btw. Also from your smartphone.
I plugged all of that into a machine learning algorithm too and let everything classify - online and in realtime. And renamed it to babbleaway, because the name "Newscombinator" was a bit too... well... fill in a word of your choice. The HN Community was not so pleased :)
Thanks so much for all the info Tom! I'll look into everything shortly. Let me know if you want access to the sql db at any point, everything is being saved ;)
3 comments
[ 215 ms ] story [ 319 ms ] threadI would love to hear your thoughts! Do you like seeing posts from multiple websites in one feed?
While I also started with aggregating only the top sites and so on, I found it quite "unpleaseant" that it lists all of the news-items in that way. I tried to make another ranking of the news and also tried to get more factors into the formula.
At the end there was a custom crawler, a special API using Solr and a custom Frontend [1] in two versions [2] which aggregates everything together.
I have also open-sourced the frontend [3][4]. If you have fun, have a look into that, but unfortunately the source-code diverged from the frontend Reader [1] a bit, because I started refactoring the db-schema and had not time to update my repositories yet. The second frontend [2] just happend because I was curious about the abilities of Foundation from Zurb. Works really well, btw. Also from your smartphone.
I plugged all of that into a machine learning algorithm too and let everything classify - online and in realtime. And renamed it to babbleaway, because the name "Newscombinator" was a bit too... well... fill in a word of your choice. The HN Community was not so pleased :)
In case there are questions, feel free to ask.
Merry Christmas. Tom
[1] http://reader.newscombinator.com/#/ [2] http://live.newscombinator.com [3] https://github.com/tomw1808/newscombinator [4] https://github.com/tomw1808/newscombinator-live