Category Archives: Computers/Internet

Dear Twitter: Please Ditch the Clutter

Have you ever been to a Las Vegas casino? The main floors tend toward sprawling layouts, with lots of shiny distractions to entice you to stay and spend more time and money on the slots instead of helping you get where you’re going. That’s what Twitter’s new layout feels like.

When Twitter started out, the home timeline would just show me posts from people I followed. Now it also shows me

  • Posts they liked, but didn’t like enough to retweet.
  • Posts from people they follow.
  • A “Who to follow” box that I can’t seem to get rid of, which is also on the sidebar.
  • Advertisements – I mean “Promoted” tweets.

I get that ads are the business model they’ve chosen, but what’s with the rest of it? It’s not like I’m going to get bored if I don’t have more suggestions shoveled in front of me.

And I am going to get frustrated if I can’t find the stuff I’m actually looking for. Let’s think back for a moment to the early 2000s, back when there were a lot of different competing search engines. Google won not just because it was fast and accurate, but because it had a simple, fast-loading, no-nonsense home page while everyone else was trying to cram everything imaginable onto a “web portal.”

With that in mind, let’s look at what happens when we look at a specific post. The logical thing to do would be to show you tweet itself and the context around it: If it’s part of a thread, show the rest of the thread. If it’s part of a discussion, show the discussion. And that’s how Twitter used to work. But now you have to click through another link to see that context, and instead it wants to show you “Tweets from people like so-and-so.” How is that a useful default?

It’s like going to a page in a book and finding not the previous and next pages, but ads for other books.

I actually do like the two-click retweet button functionality, where you click and get a menu asking if you want to retweet by itself or quote-tweet it. Normally a two-item pop-up menu is a terrible idea for usability, but this is a case where introducing some friction in the process might give people a chance to consider what they’re doing.

But the rest of it feels like they’re desperately throwing everything they can think of at me in hopes of broadening my engagement with the site. And that reaches a point of diminishing returns. When you can’t use the site for what you’re trying to do, it ends up making you much less interested in coming back.

I wish I could use TweetDeck on my phone.

It would be simpler.

My Rule of Thumb for Preventing Notification Overload

To keep myself from getting distracted by too many notifications on my phone, I ask myself the following questions whenever a new category pops up:

  • Will I need to act on it? (Likes/favorites are nice, but I don’t need to respond.)
  • How time-sensitive is it? (“Your ride is here” is more time sensitive than planning a get together for next weekend.)
  • How important? (“Server down” is more important than a project update. A conversation is more important than a newsletter.)
  • Is it actually for me, or is it an ad for the app service?

Then I turn off what I don’t need, turn off sound on the less urgent ones, and customize sounds for the most important ones.

So I hear when a text or instant message comes in, but not email or social media. When I pick up my phone I see emails, mentions & replies, but not favorites or boosts, etc.

It helps me a lot with alert overload. YMMV.

Random Thoughts on Self-Hosting

I’ve been thinking about what it means to self-host a service, and that there are degrees even within that.

I have a self-hosted WordPress blog in the sense that I manage an installation of WordPress, but I run it on a VPS at a web host. It’s not as self-hosted as someone running a server on a Lollipop Cloud or FreedomBox, but it’s more self-hosted than someone using WordPress.com. It’s also more self-hosted than someone using the managed WordPress hosting at the same web host.

The key advantages of self-hosting are privacy and control. Unless a service uses end-to-end encryption, the admins at each level can probably read your stuff – you have to trust that they won’t do it unless they have to.

And of course when you run your own service, you don’t have to fear losing control when Google Plus shuts down, or Flickr changes their pricing structure, or Tumblr changes their TOS, or MySpace botches a server migration.

The obvious disadvantage of self-hosting, of course, is that you’re on the hook for all the maintenance. Spam filtering, moderation, security updates, server migrations – those are all on you.

And unless you’re using your own software, even on your own box there’s still the risk that a project is going to shut down & leave you without security fixes, or pivot to a new direction that no longer fits what you want. (So glad WP’s block editor is still optional!)

I’ve settled on a balance where I manage the top-level web apps, but my webhost handles the hardware, the virtualization and the LAMP stack. (And email. Ugh, I’d forgotten how much of a pain a mailserver can be to handle until I tried to set one up on a Raspberry Pi a couple of months ago.)

I guess I’m kind of splitting the difference. 🤷

Verizon is Already Trying to Sell Tumblr

Wow, that shoe dropped sooner than I expected. Verizon is already shopping around to sell Tumblr. I figured it would be toward the end of the year, not the middle.

After Tumblr’s ham-handed ban on adult content last fall purged a bunch of accounts, sparked a lack of confidence, and triggered an wave of users leaving in digust, it became clear that Verizon had no idea what to do with Tumblr (not that Yahoo! had much more). If they hadn’t already started the death spiral, they’d at least knocked it out of orbit.

I’ve never been super-active on Tumblr, but I interact with a few people, and I used to occasionally post things there that weren’t reposts from my blog, or Flickr, or Instagram, or wherever. So, just in case, I backed up a full archive, imported some of the original posts, and pared down all the old duplicates and outdated signal boosts so that when Verizon inevitably gave up monetizing the site, it would be easier to find the pieces I wanted to keep.

Honestly, I’d rather Verizon sell it than shutter it and sell off the data (you think Verizon wouldn’t?). But it depends on who buys it. If anyone wants it.

Here’s hoping Tumblr finds a suitable buyer who understands what they’re getting and is willing to invest in the community, not someone who just wants to squeeze out the last few drops of cash before sending it to join GeoCities in the great Internet Archive in the Sky.

The Best Backups are Scheduled Backups

I’m thinking about social media backups again after Prismo’s database loss and one of my own test blogs crashing.

I can and do automate backups on the VPS where I host my main blogs.

I can manually backup my social media accounts, but IIRC none of them offer automatic scheduling. I have to remember to run a backup, log into the site, find the right control panel (which sometimes changes!) and request a backup.

I’d like to be able to schedule recurring backups on Mastodon, Twitter, etc. Send me an email each month with a link when it’s ready.

OK, you don’t want to keep generating backups for abandoned accounts. Here are some ideas:

  • Skip the process if I haven’t posted since the last archive.
  • Instead of scheduling a recurring job, schedule a new one 30 days out each time I download an archive.

Refine as needed.

Now, those of us with a little more tech savvy can automate some things with IFTTT. Not the native backup process, but we can set up rules to listen for new posts and automatically save the content somewhere else. But while I can reliably save the text of every post from Twitter, Mastodon, etc., saving media depends on what I’m saving it to. Often the best you can do with IFTTT is embed, not copy. (And that’s if the media is even available in the source feed. Pixelfed’s RSS doesn’t include image URLs, and Mastodon’s RSS/Atom includes them in a way IFTTT doesn’t recognize.)

Eh, maybe I should just read up on ActivityPub and see if I can make a subscribe-to-archive bot.