Hacker News Reader

Comment by mattzito | Hacker News Reader

No, alas - the Twitter terms of service are very clear. If you are storing a tweet, when someone deletes that tweet you have to delete all of your copies of it. Full stop.

View on HN

This feels very "right to be forgotten" which was almost universally reviled here. I'm actually surprised there isn't more discussion about the similarities (and similar possible negative outcomes) in light of Twitter's TOS and actions.

It'd be very hard for Twitter to determine which API account is not honouring deletes.

What? How would that be hard to determine? You need an API key to use the API.

Here, have some api keys: https://github.com/search?l=php&p=4&q=access_token+access_to...

And how would Twitter know which API key somesitewithtweets.com is using? There are ways, but if the API key user is careful and doesn't allow arbitrary searches (so Twitter can't try searching #someguid and then kill that account) it'd be pretty hard for them to figure it out. They could try adding fake deleted tweets on a per-account basis, then seeing who publishes those. A simple two-account double check or manual intervention would prevent that from leaking info.

So what? How do you connect that API key to some anonymous web site that saves deleted tweets?

The only way I can think of is for Twitter to publish different deleted tweets to different API keys and then see which one the web site surfaces. Not very hard but not trivial either.

Then just eschew the Twitter API and scrape the website, then tell Twitter to go pound sand. Doesn't do anything about getting hit with a frivolous lawsuit but that's a different topic.

That would violate the Twitter ToS, too: https://twitter.com/tos

If their TOS prevents someone from taking a photograph of their website on a screen, it probably wouldn't hold up in court.

Then will Politwoops take Twitter to court for fair use of the API?

Probably not Politwoops, because they were using the API. But perhaps a successor that did not use the API, but scraped and screenshoted only might have a good chance.

publish it somewhere it is magically picked up by a 3rd party who isn't bound by the TOS and republishes somewhere.

problem solved.

Then retweet it.

Why not just store them in the blockchain, where there's no possibility of erasing them?

Do the TOS require you to only store them in mutable data storage?

Reply on news.ycombinator.com