We've done something similar with Feeder Dashboard. You can mix RSS feeds and Twitter feeds, with unread/read states you get a glance of what you've missed, making it hard to fall into infinite scrolling. An example is our public dashboard for the us elections: https://feeder.co/dashboard/us-elections

Interesting. It's a good, concise interface. Still not a fan of horizontal scrolling, though. I probably won't ever get used to that. :) But then I'm one of those weirdos who has a portrait-mode secondary monitor set up. (I'm the author of TwitGrid.)

TwitGrid now takes parameters passed in via the URL fragment (everything after the # in a URL).

For example, you can do this:

https://vilimpoc.org/research/TwitGrid/#columns=8;users=vili...

And it will parse that and layout the feeds in 8 columns per viewport. The viewport can then be paged up / down.

I run a few bots which notifies on Slack/Discord. I think a Tweetdeck like grid would make more sense for them. Since these notifications are sometimes a little too chatty, I think it sits in-between a Tweet and Notification.

There is now an explainer video -- https://youtu.be/S8JrmXg0BNo -- for why this exists in the first place.

"Loading... If you have an Ad Blocker or Strict Tracking Prevention enabled on Firefox or Edge, this may not work.

Chrome of course, by default, leaks like a sieve."

That made me laugh, also like the idea although I do not relly use Twitter.

I do it all for the laughs.

I wish this allowed you to follow # instead of @. To me twitter is a way to follow topics and get an idea of what's the buzz in my communities.

This is actually what TweetDeck is for. The columns run into infinite scrolling land pretty fast with # topic searches. Still, it's on my mind to parametrize what handles / topics to load via the URL hash, so stay tuned. Easier than downloading / editing the original file anyways.

This would also allow you to directly bookmark a collection of handles / copy / paste a URL to others.

It seems like Twitter handles hashtags as search results, but usernames as end points.

This is a cool project and something I'll definitely be using.

I was concerned that it was 7Mb of code but most of that is the animated png :)

Heck no, the single-page HTML is currently only 17KB with a ton of comments. I try to keep things light. :)

Maybe someday I'll use TypeScript, Babel, React, and Webpack to generate a 7MB version.

Can you expand on the issues you have with TweetDeck?

First line in the README specifies that this project is a response to Tweetdeck.

Yes, but why?

1. I hate infinite scrolling and algorithmic surfacing. It's maximizing engagement and not giving me content I care about. I don't care for the dopamine rush.

2. I want to leverage visual memory, you know what you've seen even over a few days, and it's easy to scan through a list of peoples' last posts to know you've already seen it. A quick scan with positional context that doesn't mutate every 30 seconds trying to keep you from bouncing.

Imagine reading a newspaper where the columns reloaded every 30 seconds.

This tries to be precisely the opposite of that, though individual user columns will update if they post, that's part of Twitter's iframe JS though, so out of my hands.