Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A new and improved Twitter API

blog.twitter.com

Post reply on HN
Print view View original

101–110 of 116 posts

Re: A new and improved Twitter API

#101

Post by fireattack »

L3viathan wrote: post #65
benlumen wrote: post #33

I'd love access to the Twitter API for research purposes for my writing, but I'm priced out for that purpose. I can't prove it, but it feels like they only want big businesses researching their markets, and not people like me asking questions about Twitter itself - even though letting people do that would result in negligible additional traffic. This might be controversial and I know people will say that they're a pr…

This is quite new: https://github.com/lschmelzeisen/nasty/

Doesn't seem to support "from:{screen_id}" in search, a bummer.

Re: A new and improved Twitter API

#102

Post by esnard »

londons_explore wrote: post #96
esnard wrote: post #56

Earlier quoted context omitted.

You're technically right, but the Twitter API won't return the next 3600, even after deleting the first ones. The response will be empty, even if you know for sure there's more.

Meaning the API is leaking the number of deleted tweets? Could you make a tool to find out how many tweets someone has deleted?

No, my understanding is that Twitter has a "bucket" of the latest 3600 tweets you've posted / liked. This bucket is faster than their database to query, so once you empty the bucket, the API returns no result, and the bucket is never "refilled" from the database.

Those are not technical terms, sorry, but this is how I currently understand it so it might help you too?

Re: A new and improved Twitter API

#103

Post by munificent »

An interesting side effect of making your product ad-driven is what it does to the incentive structure of treating your product as a platform.

Twitter is just one example of many of companies who have created an API, asked developers to build on it, and then later burned the whole thing down.

Fundamentally, if you let developers build products that access your data in ways that users want, the very first they're going to do is remove the fucking ads. And, obviously, if that's how your company makes money, you can't let that happen. Given the choice between killing your ads or killing your API, you're going to kill the API every time.

Imagine running a casino and offering a "slot machine API". If you let the developers discard all of the losing spins, you're gonna have a bad time. This is essentially what an ad-driven company does when they have an API that lets developers separate the data users want (actual data) from the data they don't (ads).

Re: A new and improved Twitter API

#106

Post by draw_down »

toyg wrote: post #41
Nav_Panel wrote: post #40

Earlier quoted context omitted.

> conversation threading Actually, this was a major point of API failure when I last tried to use it. Did you know that, with the current API, there's no way to retrieve the replies to a tweet ? This makes it impossible to build tools that, say, allow you to visualize a tweet thread, because you simply can't fetch the thread. So I'll say that, as a casual, social, and informational user of Twitter (i.e. not business,…

> there's no way to retrieve the replies to a tweet My crusty old version of Tweetbot has been doing that for years...? Maybe not perfectly for tweets with hundreds of replies, but well enough.

Tweetbot actually doesn't do this very well. Look at a tweet that's a week old or so, you won't see replies. Open it in Safari and the replies appear. (The limit may be a little longer, perhaps a month, but either way you won't see replies past a certain freshness.)

Re: A new and improved Twitter API

#107

Post by ehsankia »

benlumen wrote: post #33

I'd love access to the Twitter API for research purposes for my writing, but I'm priced out for that purpose. I can't prove it, but it feels like they only want big businesses researching their markets, and not people like me asking questions about Twitter itself - even though letting people do that would result in negligible additional traffic. This might be controversial and I know people will say that they're a pr…

This is just my own vague experience, but I don't think this specific to Twitter, nor does it really have to do with pricing people out.

I think that public APIs in general are a thing of the past, mostly because of bad actors, security problems and scale. Look at things like Clearview AI, just slurping up every photo they can get their hands on. All it takes is one or two abusers and then everyone has to be limited.

The simple instagram API that I used to just get my own most recent post 3-4 times per day, now requires me to submit an application, with multiple size app icons, justification for my use, privacy policy, ToS, etc. All to I can fetch my most recent post from my own profile.

It's just ridiculous how every large website is clamping down on their APIs, but with abuse left and right, they don't really have many options.

Re: A new and improved Twitter API

#108

Post by jMyles »

Lots of (seemingly justified) complaints here. Is anybody on Mastadon? I have joined, but I have no followers and don't know who to follow. I'd love some connections over there.

I'm @PintSizePorcupine@mstdn.social

Re: A new and improved Twitter API

#109

Post by lacker »

esnard wrote: post #102
londons_explore wrote: post #96

Earlier quoted context omitted.

Meaning the API is leaking the number of deleted tweets? Could you make a tool to find out how many tweets someone has deleted?

No, my understanding is that Twitter has a "bucket" of the latest 3600 tweets you've posted / liked. This bucket is faster than their database to query, so once you empty the bucket, the API returns no result, and the bucket is never "refilled" from the database. Those are not technical terms, sorry, but this is how I currently understand it so it might help you too?

Rather than “buckets”, I bet they just have a flag on each tweet in the database for whether it’s deleted. So if you delete the most recent 3600, next time they do a query to see recent tweets, it retrieves 3600, but they’re all deleted, so it filters them all out. Buckets are a little less likely implementation wise because you’d either sometimes see that the most recent bucket had very few tweets in it, or they’d have to be constantly shuffling tweets between buckets.

Re: A new and improved Twitter API

#110

Post by alicemaz »

picodguyo wrote: post #21

This is what they cite as the key benefits: "A cleaner API that's easier to use, with new developer features like the ability to specify which fields get returned, or retrieve more Tweets from a conversation within the same response Some of the most requested features that were missing from the API, including conversation threading, poll results in Tweets, pinned Tweets on profiles, spam filtering, and a more powerfu…

if they reintroduce user streams I'll be pleased, if they ease off the per-app user limit I'll be elated. it doesn't really mean anything if they promise such and such policies or initiatives otherwise we'd all be holding our breath for bluesky. what's important is what they actually end up doing, which I guess we'll find out next week
Post reply on HN