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

How do they know if someone was going to tweet a longer sentence, like for example:

"Donald Trump is the worst president ever this country has ever had. He stole the election with help from the Russians and is running this country into the ground."

But knowing that they only get 140 chars, they just decided to tweet "Trump sucks! :-( And he's in bed with the Russians"?

You don't have to hit that 140 char limit too many times until you start self-censoring yourself and write shorter messages.




The network tab in Chrome does not seem to indicate they are collecting data for this. I would expect some sort of network activity as I type/reach the character limit, but I spot nothing.


they could store that info locally and send it when you submit the tweet. you would lose information on any non-submitted tweets though


That was my theory, maybe they don't after all..

(maybe Child is right, it's not real time data harvest but slower gathering)


So editing is now called "self-censoring"? What a time to be alive. /s


> How do they know if someone was going to tweet a longer sentence

Analytics


That works for users using an official Twitter client who actually typed in the tweet and then edited it. It doesn't account for people who decided to enter something shorter because they knew what they wanted to say would hit the limit.


True, but to an extent you know how people type, start something then stop, hesitate.. an awful lot can be retrieved from an input field.


I was referring more to people who mentally edit what they want to say before they even start typing because they know it would be too long.


Sure on that point nobody can, but between no knowledge and mind reading there's a huge amount of stuff website can gather.


Analytics != mind-reading. Unless they send every single keystroke and character deletion, they can't track that.




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

Search: