No amount of log entries can ever prove things are working; a single log entry can prove things are broken.
Haha so I'm watching Sidekiq logs and every time I toot a *lot* of jobs get kicked off 😅
@ashfurrow in my opinion it succeed because of what Mastodon is and how it works not so much because of how the data is store and query. If we leave everything about Mastodon the same but swap postgrees with a graph data Mastodon would still be same. The invers is not true.
@ashfurrow agree. Sometimes I think it's because the top management is incapable of taking a hard look at what they are building and accept that they are gone of the path and work to fix that. Instead they look at what tech they can trow there money at that will automagiclly fix there problem
@ashfurrow when I look at mastodon as a blackbox, I see that what works is the data model. there are users and who they follow, it must mainly be a 2-level subscription mechanism. So it does not need the advantages a graph DB offers with deep relations, because the limit is preset at 2 levels. If it were to offer you liked this guy maybe you will also like that gal kind of suggestions, doing that and similar in postgress without a preset limit would be more taxing.
@Otuk sure, so let’s examine that a bit. My toot was that Mastodon was “successful” using only Postgres – so the user suggestion feature you mention sort of begs the question of “does a social network need that to be successful”, and the answer is apparently not.
It’s a super-interesting idea. I think that the use of traditional database/structures kind of reveals that traditional social networks (twitter circa 2010) are sufficient.
(I just woke up, hope that all makes sense!)
@kiview right – all that stuff, and more. I think the fediverse’s biggest strengths (distribution of responsibility, purpose-built instances, many people working together) will also raise new challenges that existing social networks like Twitter never had to deal with. That’s gonna be a problem because, while Mastodon can rely on OSS technology built by those companies, it can’t rely on their solutions to these social problems.
@ashfurrow Regarding Mastodon I'm always reminded about diaspora*, with a very similar approach, but never getting enough traction to be relevant (at least from my short encounter with it). I think the Internet (and society) needs free and open alternatives for social networks, if these will be Mastodon or diaspora*, I'm not sure yet... A walled garden is too comfortable for the normal user I fear.
@kiview yeah, we will find out! The benefit of using open protocols is that we can build momentum. When whatever replaces Mastodon get built, it won’t start from scratch; it’ll start from where Mastodon left off.
Case in point: d* is a part of the fediverse, this instance federated with some d* users. Mastodon didn’t start from scratch, it picked up where d* left off ☺️
@ashfurrow true and still this is a super big challenge for most OSS projects as well (regarding interactions on GitHub). I see a general lack of empathy of users towards the software authors, which at one points also poisons the relationships of the authors towards the users and the community.