4 ms·
Thanks for the heads up! - The rankings use karma and number of mentions, which for a week basically bring books by top karma users to the top. The actual form
by leandot 10y ago
Thanks for the heads up!
- The rankings use karma and number of mentions, which for a week basically bring books by top karma users to the top. The actual formula is:
book_entry['rank'] = (avg_karma_count * avg_karma + book_entry['karma'] * book_entry['karma_count']) / (book_entry['karma_count'] + avg_karma_count)
- I plan to make Top 10 monthly or yearly, just wanted to see if people would like the idea at all. Seems like people like it. As a side remark I also did http://hnbuzz.com/ http://hnbuzz.com/ where you can filter, segment HN posts daily, weekly, monthly etc. It is not about books though.
- Yep the signup seems too aggressive, it will get changed :) After I clicked it away the first time it didn't bother me anymore while coding the website so I got carried away and left it as is. But it is also a learning experience to see what people think about those.
The project is totally side project, I actually do that often - build something I like to have and use a new tech to do it to learn. Drop me a mail if you want to know more.
- vram22 10y agoWhat's the new tech you used for it?
- ycombinatorMan 10y agoitd be nice if mentions had a bigger weight in this
- leandot 10y agoThe tech that actually made my life much easier was Docker. Using docker-compose to manage the tech stack (redis, elastic, apache, flask), docker-machine to abstract away where docker is running, docker-haproxy to be able to run multiple services on the same instance with virtual hosts, sane deployment with docker versioning, etc. etc.
- jrlocke 10y agoI would love to use this when it runs on data from a longer time period.