David CoxA proposal for how we could make a Twitter to Mastodon migration actually work (let me know what you think, this is something we could actually do this if people were willing to invest in such a thing):
1/n
First: there needs to one (1) place where people land, that can take as many people as want to sign up. No leading w/ explanations of servers, that comes later and/or on an "advanced sign up" tab or something. The pitch is basically Twitter, but not ruled by corporate interests.
To the extent that decentralization/federation is part of the pitch, the message is that that's because it keeps you from being locked in. The instance (which may or may not carry the name "mastodon" at all) is run by a nonprofit with a proper nonprofit board and governance.
To the extent that a default mastodon server can't handle huge numbers of users, probably fine to just randomly assign people to "subinstances" under the master umbrella. This is just the cohort you entered with, no particular significance to it, and ideally no one even notices.
They can change to a different "home community" later or to a different Mastodon server later as they get their footing. All that stuff is there if/when people start actually caring about it.
A few enhancements to core Mastodon are needed, either as contributions, or as a fork if needed. Proper threading, quote tweets (yes, quote tweets), basic text search, and algorithmic ways of finding other users. I know these are controversial.
It's not algorithms per se that make Twitter bad, it's non-transparent algorithms that exist to serve a business model. We can build algorithms that are open, and user-tunable, with sensible defaults that encourage better behavior.
But without recommendation algorithms or even search, discoverability is just dreadful. We don't need to be Luddites to encourage a good experience, we just need to be transparent and working in the interests of the public good.
Moderation needs to be a funded, professional, first class function. That's why we need a nonprofit that actually raises real money to coordinate everything. There need to be full-time employees and an actual governance structure.
Abdicating these functions to random private citizens on a volunteer basis just doesn't work.
Anything built for this new Mastodon-compatible site should be made available to the rest of the mastodon world, and they are free to pick and choose whatever parts they do or don't want. This isn't embrace-extend-kill, it's embrace-extend-share. That's possible for a nonprofit.
The landing/sign-on experience should also acknowledge that people are landing from Twitter and make it easier to migrate. There are nice services for cross-posting and finding friends, but these need to be integrated directly into the sign up process.
We also need to fix the problem of being able to migrate to a new instance, *including past posts.* Maybe you should even be able to import a Twitter archive?
These features needn't all be available at launch. It's sufficient to just signal they are coming. It's hard to know what's coming on most Mastodon instances, and I feel like the whole experience was built by people who never really liked or "got" why Twitter is so important.
And make no mistake, Twitter is really important, whether you think it is a "hellsite" or not. Extremely valuable and important information exchange is happening there, and we have a chance to decouple that from corporate overlords.
We could make this happen in real life if we had some money and ideally some big fish to help lend credibility (and probably dollars 😅) to make it happen. I even suspect that some of the most important changes (e.g. to the onboarding/landing experience) could be done swiftly
Is this something that is of interest to people? Is someone out there already moving in this direction?
Raito Bezarius@neurobongo I already worked on https://socialhub.activitypub.rocks/t/guidelines-to-implement-federated-recommender-systems/2184 but I did not get any traction at that moment.
Mangaki.fr has some algorithms which could be applied in a federated setup, but there would be a need for some efforts to integrate into standards + server implementations.
Anyway, the technology is here though and even some funding should be available.
Guidelines to implement federated recommender systems?
Hi there, I am a core dev of GitHub - mangaki/mangaki: Site de recommandation de mangas et d'anime which is anime and manga recommendation website based on our own library of recommendation algorithms: GitHub - mangaki/zero: Mangaki's recommendation algorithms Our team is currently interested in how Mangaki and our library could become more federated / privacy-friendly, in particular, we see two avenues which are complementary IMHO: recommender system based on homomorphic encryption (using El...
@raito yeah, I suspect this is very much a thing that the community can do, and in many cases, already has. I think the trick is finding out if people actually would want to rally around such a thing (or if there already is a thing that we can rally around)
Kristin Branson@neurobongo I think it's going to be essential. While I personally love figuring out how new sites work & it has been lots of fun to try to figure out e.g. why searching for some handles was failing, but I'm pretty sure I'm in the minority there :).
I've spent a lot of my career trying to make usable UIs, and I think there will be many people who will struggle with the onboarding. The two options I see are:
1) A smoothed-out interface for non-experts built on mastodon's code base and design choices, analogous to how Ubuntu has a UI, and you can choose to use the terminal or you can choose to click on things. I think you've pointed out some of the major issues for new users -- choosing a server, finding posts.
2) A separate system, but one that federates with Mastodon servers. Some servers could choose to block these more user-friendly servers if they want to.
I personally like communicating with people who may not be as good at figuring out the intricacies of websites, but are brilliant in other ways, so I'm really hoping this happens. I also hope it is not based on blockchain?? because, why?
My main worry right now is that the current moderation model may not be able to handle the large influx of users. But, Wikipedia seems to do a good job with this, despite being all volunteer-based.
I also want to say how much I appreciate the patience that the long-time users have had with this invasion. It says a lot that so many of us new users are welcomed.
PS some servers allow 65K character posts, eliminating the need for threads. What do you think about that option?
Patrick Vanhoucke@neurobongo If we had some money? Would you set up a crowdfunding to buy Twitter? If the ‘main bird’ would want to sell it at all? Currently, Twitter is in a monopoly situation where the owner is also the CEO and the sole director. He can do anything that is not prohibited by law. No one will stop him. And he will rather push Twitter into destruction than compromise on his narcissistic ideas. Twitter represents ‘only’ 20% of his wealth.