If it's not federated, why Lemmy?
this post was submitted on 24 Jun 2023
8 points (100.0% liked)
Self Hosted - Self-hosting your services.
11419 readers
1 users here now
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.
Rules
- No harassment
- crossposts from c/Open Source & c/docker & related may be allowed, depending on context
- Video Promoting is allowed if is within the topic.
- No spamming.
- Stay friendly.
- Follow the lemmy.ml instance rules.
- Tag your post. (Read under)
Important
Beginning of January 1st 2024 this rule WILL be enforced. Posts that are not tagged will be warned and if not fixed within 24h then removed!
- Lemmy doesn't have tags yet, so mark it with [Question], [Help], [Project], [Other], [Promoting] or other you may think is appropriate.
Cross-posting
- !everything_git@lemmy.ml is allowed!
- !docker@lemmy.ml is allowed!
- !portainer@lemmy.ml is allowed!
- !fediverse@lemmy.ml is allowed if topic has to do with selfhosting.
- !selfhosted@lemmy.ml is allowed!
If you see a rule-breaker please DM the mods!
founded 3 years ago
MODERATORS
We're looking for a Community Management platform that has the features and usability of Reddit. Unfortunately we are not experts in Community Management tools, so are doing analysis a bit blindly.
We initially liked Discourse, as it is a pretty common open-source platform for online forums. However after standing up a proof of concept we are now highly skeptical that it scales horizontally enough for our needs...the concept of "infinite user-created sub-reddits" are an absolute must-have, and has been the major sticking point on the tools we have reviewed and piloted so far (including a number of proprietary non-open source options). Discourse themselves told us to limit the number of Categories that should be used, so my sense is that this is an inherent architectural limitation that can't be overcome without significant investment and branching of the code base. Comment voting is also a must-have, as well as a few other things. We are finding that these seem to be strengths of Lemmy and other Reddit-style platforms. And Lemmy seems to be the most popular...with the most community support. However if there is another platform that might better option, I'm open to it.
Meaning no offense to Lemmy devs, Lemmy still has some things to iron out until they can safely say "we're at version 1.0" - it's in beta, being battle tested as we speak, and it's not afraid to push feature breaking changes. For example, the latest version, 0.18, fixed a very annoying bug that pushed new posts in real-time (websockets) when using the web interface, but disabled captcha. Jerboa, the main Android client, pushed an update that removed compatibility with Lemmy below 0.18. lemmy.world, who competes with lemmy.ml for the first two instances by number of users, refused to upgrade to 0.18 because it needs captcha to ward off bots - now you can't use Jerboa with lemmy.world, and the web interface has the bug with new posts. So, all this rant should give you the idea that Lemmy is far from production ready, and it's not a desirable choice, yet.
I've been the engineering manager for a Twitch clone that got sold for the tech part a couple of years ago - check your DM, sent a personal link, you can use the email there if you want me to help.
Aren't those specs that would be better served by designing your own tool than introducing a complexe codebase that solves way more than the feature set you're after and whose knowledgeable people will be difficult to find and hire?
I would say what you really need is to take the time of specifying what are the very concrete features you put in the idea of a community management platform, then see if there is not a simple path to implement them incrementally (protip : remove the "platform" in the label, it immediately looks less scary and it serves no purpose anyway, that's a word you use to impress investors).
Of course, I say that from an external and naive viewpoint after reading your short description, sorry if I'm far off.
We have a detailed feature roadmap/prioritization laid out. I only mentioned the two most-critical items that I am seeing as real differentiators between the tools we have reviewed. And we have evaluated quite a few in addition to Lemmy: Discourse was (is?) the front-runner, but also proprietary, paid, non open-source ones like Khoros, Verint, and numerous others.
The challenge is that:
a) My business partners are keen on the Reddit style interface, but it must be a standalone instance and white-labeled.
b) Our business case requires near-infinite sub-reddits, which most of these tools can't provide
c) Our unique user base and business model is the special sauce of our investor pitch, not the tech. In a pinch, any of these tools can work, but we need something that scales in the right way. Replatforming down the road is expensive and impactful to users, whereas spending a bit of time up front to do tool eval is much cheaper. You can't "fail fast" when it comes to significant strategic decisions like this.
d) We don't yet have the funding for a full-spectrum, full-time dev team. We can afford one or two tech people part-time right now, with the assumption that standing up a pilot can be done part-time if the person has done it before. Once we are funded though, we can share fixes/features that we build back into the Lemmy community. That level of control is why I like open source tools over proprietary (where you don't have the ability to modify code or define the roadmap priorities).
Hence we don't want to build our own tool from scratch if Lemmy can check enough of the feature boxes. But I want to pressure test that, as I am concerned greatly about its overall lack of maturity as a platform (as @FleaCatcher also mentions below).
I don’t think you’re looking for architects or developers if you just want to run an instance. Those would be for developing custom features for Lemmy. I believe you are looking for a Cloud Engineer.
You're right...however in a perfect world I'm looking for more of a high-end generalist that can do it all, as I definitely expect that development will be necessary. If there aren't any proverbial "purple squirrels" out there that can (or want) to do both though, then we'll need to separate the roles and hire multiple.
Actually he needs a developer to disable federation. It's on be l by default and to my knowledge there is no way to disable it.
There are two config options that control this:
- private_instance
- federation_enabled
Federation is disabled by default, and needs to be enabled either through the online admin panel or directly through the config.json file.
view more: next ›