Hacker Read top | best | new | newcomments | leaders | about | bookmarklet login

Indeed, it has a completely different target. In fact, it was built as an experiment to see what could be built using reddit data without being Yet Another reddit Browser.


sort by: page size:

Speaking of, are there alternative Reddit backends? It would be interesting to be able to take one of these sites or Reddit apps and point it at an alternative backend that provides identical functionality with different data.

Reddit also has tons of unofficial clients that interact with it. It’s not quite the same since you still don’t own the data, bit it comes close.

Yeah it's custom. Perhaps the old version used to be based on the open-source bit of Reddit's codebase. But I think the new one is totally home-grown. I think its source code is available though.

Reddit's APIs are not user-based from what I can tell.

It's not really an exact Reddit-looking clone, but we built a tool called Newsy, which builds content aggregator like Reddit without code.

https://www.newsy.co

We focus on users with un-used domain names and try to spin up a news site like Reddit quickly with various features built in.


Why not just port to the Reddit codebase? The functionality seems similar.

Thanks for the info! I'm intrigued by the different interfaces for the same content. I'll definitely keep an eye on this when reddit dies in a couple weeks!

It's trivially easy to build a scraper for the new content too. And I think people really over estimate the relative amount of data reddit has compared to the literal rest of the internet. Like sure, it's nicely categorized and ranked (somewhat), but leaving reddit out of a corpus probably won't drastically effect your final model.

Yep, definitely. The API was only a small sample of the actual business logic inside the Reddit site as a whole.

This uses a 2017 fork of reddit’s server that reddit open sourced. The backend wouldn’t be reddit-like. It would actually be reddit, pre-new UI.

Posted recently: https://news.ycombinator.com/item?id=36203610

The p2p backend makes it fairly slow to load, but it's an interesting approach. They seem to have put in a lot of effort to make the UI an almost-clone. Which is sort of interesting since the reddit UI gets a lot of negative feedback. Probably lots anyone could learn from the source code though.


A new feature but it wasn’t built on a new codebase. Reddit is a monolith and a lot of things users think of as different “entities” live in the same set of tables.

Yeah, it's based on the old open source reddit.

I sort of explained this in another post but it does consists more features than Reddit (or similar). It consists a poll system, people can add more topic (if not existed), people can like topics (tags) and users. Users can view their liked users on their page view.

There are many more feature added into that I haven't mentioned. I also got more features waiting to be integrating in the nearby future.

Perhaps, I might focus more detailing information on it, so people can have a better idea of what's going on...

Thank you for your response.


I've put together a comparison of Reddit alternatives if anybody is interested:

https://gist.github.com/hanniabu/6f96c6e820d58d8736f3c15d4c0...

There's also some notes above the linked table


Yes, it is different. At the time he wrote that, there was literally one hosted FreeBSD box running all of reddit. It was called... 1.reddit.com. Eventually the database had to be moved to a separate machine called... 2.reddit.com.

reddit now runs on a few hundred Ubuntu EC2 instances.


This is such a cool idea! Does it scrape reddit, or use the API?

The authors of Raddit tried that, but found that actually running Reddit's software, especially on low-end hardware for a smaller community, was a pain in the arse and decided it'd be easier to build their own that fit their purposes.

This reminds me of how Reddit works internally. It’s not exactly the same, but I recall they had a “thing” table that aggregated every kind of object into one.

Pretty cool setup. Once again, not the same, but interesting

next

Legal | privacy