Buzzword bonanza!
Announcements
Community updates and announcements.
Admins will post any updates here so be sure to follow!
Important updates will be pinned to the local feed.
I know, I tried to make it sound friendly and not anti-Lemmy.
I feel so old. I read that entire announcement three times, and still have no idea what Sublinks is.
Basically, it's a replacement for Lemmy. Ground-up rewrite of the source using a language with a much larger community.
I thought I was pretty familiar with the fediverse (joined mastodon in 2018) but I don’t understand what some of that means. What is a Link Aggregation Social Network and why is it capitalised?
Lemmy, Reddit, Sublinks, Kbin are all Link Aggregation social networks. They mostly share links to articles and the like. It's just the category they're in.
Oh, God. This must be the moment when I realize I'm over the hill for real. You're clearly assuming I know what Lemmy is, which implies that most people in this setting would, in fact, know what it is.
But I don't.
Jesus, I'm gonna need some Chivas after this.
it's funny, because I'm reading your comment from lemmy! to me, you're already "on Lemmy" so good job.
In old man terms: it's just a bunch of websites that talk to each other. we share links and memes, using some sites with a reddit aesthetic (lemmy), some with a twitter aesthetic (mastodon), but all (most of) the content gets posted to each other. me, I found my way here from the announcement being linked on Lemmy.
and you're not missing much from the announcement. they didn't really say anything substantial. it's all just corporate speak - bedazzled promises yet to be delivered. we'll see what they launch when they launch it, but i stopped caring by the time i read "innovate" and "revolution". at the end of the day it's gonna be pizzas and cats.
Lemmy is the software that runs discuss.online, lemmy.world, etc.
Brother, you're makin' it worse...
Lemmy is to Reddit as Mastodon is to Twitter.
There are so many buzzwords in that announcement it makes my head hurt.
In fact, went to sublinks.org and the about section is also full of buzzwords. It’s not clear at all.
Well, I didn't mean to be particularly critical. I just didn't know what it was announcing. Now that you mention it, though, the wording is a bit corpo for the Fediverse.
Agreed, not critical just puzzled.
You’re going to revolutionize how we share by being API-compatible with Lemmy?
We want to capture existing websites that run Lemmy. We'll have a migration tool to convert from Lemmy to Sublinks. Users will still be able to user their favorite Lemmy phone apps, etc.
So this service is coming to sum instead of divide?
It's just forking Lemmy, but it will be fully compatible with it for federation, etc. It's not meant to create a ruckus. I simply wanted to move faster with some features and I cannot do that with Rust.
Great!
Hang on, you’ll switch discuss.online to this sublinks.org ? What if I don’t want to?
The change won't be noticeable until we start adding new features. The main reason to create Sublinks is to move quicker with features & functionality that the current Lemmy team cannot maintain for various reasons.
It’s just that I’ve had to create new accounts before (because of incompatibility) and recreate subscriptions, loose post history etc. Also because of instances not being maintained.
I just thought discuss.online was different and a more stable place to be. If you do migrate discuss.online will we still be able access and contribute to our subscribed Lemmy communities?
Yes, it’s a drop-in replacement for Lemmy. The only thing you may notice is having to reset your password because the password hashing is currently different on Sublinks. Everything else will be the same or better.
It’s a migration to Sublinks not a switch. That means all data will be transferred over.
Let me just check I’ve got this right.
Sub links is an enhanced version of Lemmy with some extra features. It works with normal Lemmy Clients. We’ll still be able to access our existing Lemmy communities but if our account is on a sublink instance then we’ll be able to take advantage of the enhancements in sublink communities.
We won’t need to migrate anything across manually, just log out and log in.
Yes, that sums it up pretty well.
Thank you
It's neat how your breathless description makes it sound like you've discovered fire but then it reads like a "devs not implementing our pet features" fork.
You'll be - of course - committing changes back to a feature branch to enrich the project better than Kay Sievers did, right? This isn't some petulant land-grab like Bender going off to make his own casino?
It’s not a code fork it’s a completely new codebase in a different language.
It’s not just about implementing “pet features”. I’ve worked closely with admins of all major Lemmy instances to build the feature set for this and the roadmap plan.
Will it compete with https://join.piefed.social/ ? Or does it have some stand out features ?
You can see the full feature roadmap here: https://github.com/orgs/sublinks/projects/1/views/6
Sounds very cool. Hope all the best for you/SL!
So general question ... why not contribute back to or softly-fork Lemmy?
While I'm sure you've got a lot to offer here and that SL may very well come to be awesome (especially, IMO, with the attractiveness of the tech stack to would-be contributors), I can't help but wonder if it'd be better in this moment for the fediverse to focus more on building on what's got momentum rather than splitting efforts. There are, of course, many counters to that argument ... so I'm wondering what your thoughts are in general and behind this project?
Main reason (or at least one of them) is the technology stack, choosing Java instead of Rust, to move fast with development, and (hopefully) to be more accessible for others to contribute.
All good reasons! Thanks!
It would be interesting to see benchmarks for different mock scenarios (regular user interaction, federation etc...). My understanding is that Lemmy has had a poor database design and bad SQL queries for a very long time, not sure if that improved (but since response times are definitely low with recent versions I guess yes), but it would be really cool if the database could be designed for performance from the bottom up instead of having it as an afterthought which led to the huge downtimes that we experienced last summer when servers with AMD EPYC CPUs and 100s of GBs of RAM couldn't handle a few ten thousand users.
Our goal is to fix the database. I've reengineered it. This has caused us to need a migration path rather than drop-in replacement. I didn't want to inherit their schema.
That's pretty cool, best luck with the project!
Thanks!
Good point