Fantastic. I don't have a twitter account and it's really annoying to open a link and find out twitter thinks it's adult content (often not really) and get a login page. Or just scroll too far into the replies get hit with a login.
feel like it's half not their fault (scrolling is hard to customize in the DOM), half mostly of their fault (you can use columns to make infinite-scroll interfaces less neurologically grating), and half totally their fault (giant dom hierarchy that ruins my ram, loading recirc before content, zero confidence when ctrl-clicking something whether it will actually open in a new tab)
Nitter provides superior UX even though it takes 10 times longer to load. Think about that, I’m waiting 15 seconds because that is more enjoyable than Twitter’s native interface. No nags, no autoplay videos, no repeated clicking to see a single thread, reasonable amounts of whitespace.
Nitter is an excellent alternative. The only thing stopping me from 100% using it, as I have for Reddit->Libreddit and Youtube->Invidious is the lack of some sort of timeline. I tried setting up an RSS feed for it, but I mostly use Twitter to follow artists I like, and it wasn't easy to browse. Fortunately, there is a pull request for this feature (https://github.com/zedeus/nitter/pull/363) but it's been sitting for a while.
One minor annoyance I have with one of the "open" Nitter hosting sites I use to follow Twitter accounts is that Nitter seems (at least on iOS) to save bookmarks with the original twitter URL instead of the Nitter one I'm currently viewing, and I then have to manually edit the bookmark URL to correct it...
Does anyone know if it's possible to control this (with an option?) if I were to host Nitter myself (open source I guess, but still)?
This does negate some of the privacy from Twitter advantages, especially if hosted in your home network. Of course nitter also has the privacy from instance operator question, and your two options are basically to use a small list of instances you trust or such a large pool of instances that the instance operator has very little to link together
The instructions provided in the README (about manually bookmarking a bookmarklet link) reflects the careful process that has to be followed to use bookmarklets on mobile browsers. The instructions to drag and drop this bookmarklet (or even long press and copy the link destination) will fail for anyone who's not using a traditional desktop/laptop form factor computer system.
Bookmarklets work on all devices. Deskop, iOS and Android.
I can easily read and understand the code.
They only run when I invoke them. While extensions have access to everything you do an any page, bookmarklets only have access to the page where you invoke them.
They are easy to change. For example in this one, I will probably update the array of Nitter instances regularly.
I can organize them just like my other bookmarks.
I can put the bookmark in just a line of text in a text file with an explanation. As I did here. No need to dabble with extension repositories of Firefox and Chrome, as you would have to with extensions. I don't even know how it would work for other browsers.
Regulation could give them a reason to, but there's a conflict of interest when the controlling interests in government would prefer to ask the FBI to suppress inconvenient speech on the dominant platforms.