Hacker News new | past | comments | ask | show | jobs | submit login




I did the bottle exercise. Very interesting.

I see the term I was looking for was LSM (language style matching). That should help me do more research.

Thanks for the link.


Here's a project idea you are free to steal: write a browser plugin that analyzes HN comments and shows up something on the screen. Like "honest", "deceptive", "gratuitously negative".

Brownie points for HN admins if they add text analysis as feedback when we press submit.

I'm tired of signing up with a brand new HN account every few months to cover my tracks after embarrassing myself with less than noble posts.


I'm less interested in analyzing the meaning of the text than the privacy implications for identification. The way I see it, textual fingerprinting is similar to facial recognition in that it's next to impossible to avoid in public spaces (walking into a subway, using Facebook's app, posting on HN, etc.).

Maybe a style obfuscator. It'd be easier than trying to recognize emotional content.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: