Content about daystream

Responsive image

I finally got around to connecting photo uploads to AWS Lambda. I've now offloaded version generation to Lambda, which is working nicely. I used this photo of Teddy for my testing.

One of the things I struggle with on Daystream is the difference between a post and an article. The main distinction is that an article has a title and a post doesn't. It's somewhat more extensive than that (different editors, articles have a special isolated view, etc.), but, conceptually, not much more.

Inclusion of both types of entries has been intentional from the beginning, but, to this day, I struggle with choosing the best bucket for a particular entry at the time of writing it. I make a choice in the moment and the content sits in that bucket forevermore. This in-the-moment decision is largely based on the expected length I have for a particular piece of content when I start writing it. Pretty short? Make it a post. Longer? Make it an article.

I suppose that framework works ok, but it certainly breaks down from time to time. For example, this morning I was looking back on [this *post* on the ways in which travel has changed in recent years]( It might be one of my favorite things I've written on Daystream. But, reading it today, I couldn't help but think that it should have been an article. It's somewhat long and certainly lends itself to a title.

At a minimum, I need to add functionality for switching a post to an article. I added that to the todo list right after re-reading that post this morning.

This, by the way, is a post and should remain one. I think.

I recently upgraded the Daystream web app to Ruby on Rails 7. Everything went rather smoothly, but I did discover a hidden issue a few days ago. Basically, images in articles weren't being rendered.

It took me a couple days, but I finally tracked the issue down to [a change in the default variant processor in Rails' Active Storage Module]( Turns out, Amazon's Linux doesn't seem to be configured with the new default processor (vips), so the servers couldn't render the images whenever a relevant article was requested.

Fixing the problem was straightforward - a simple settings change to use the default processor from Rails 6 (mini_magick) instead of the new default (I decided this was easier than installing vips on the AWS servers).

So, articles are now being rendered in their full glory again, letting me reflect on some of my favorites, like [*The Hustler* is not a Pool Movie](

responsive image It's even worse than it appears..

I'm mad at myself. I've neglected posting lately, mostly because I was doing some work on the back end of Daystream (moving it to the most recent AWS Ruby platform, upgrading Rails, etc.). That's probably the biggest downside of the solo setup—working on Daystream takes away from the time I have to use it, especially when it's not my day job, so to speak.

But, I wouldn't have it any other way.

The shame is that we've done a lot over the last couple weeks—moved Jonathan, Paige's wedding, a friend's wedding—and I haven't posted about any of it.

Maybe I'll make some posts to the past.

I completed a pretty major update to Daystream last night. I've been working on upgrading to Rails 7 over the last couple weeks and, while doing that, decided to switch from Bootstrap to Tailwind CSS. Last night I pushed everything to production.

Getting Rails 7 to work on Elastic Beanstalk is far simpler than the process for doing so for Rails 6. And deleting node and yarn from the process felt way better than I expected.

And I'm loving the flexibility Tailwind provides. There's still a lot to do on the design front, including a lot of cleanup, but the Tailwind's approach to CSS is more logical to me and, dare I say, has me excited about improving the design of the site.

So, the work continues.

Finally acting on a Daystream feature I've contemplated for some time. Working locally. Will push to the site over the next few weeks.

I did some maintenance on Daystream user RSS feeds tonight. Fixed a few minor issues that broke feed validation from time to time (depending on the items in the feed at any given time.)

Digging into RSS is always a source of frustration for me. It's beautiful and, yes, simple. It's amazing to me that more people don't use it to consume content from the web. I get why publishers moved away from it - eyeballs on sites are more valuable than quickly scannable content in a reader.

They won, we lost.

Just doing a little testing tonight.

It's even worse than it appears.

I've let Daystream go before, and I missed it. I don't want that to happen again.

Responsive image

This seems as good a photo as any to be the first posted through the Shortcuts app on iOS. I remember taking it at a gas station on a drive home from Florida. In southern Georgia, I think. The grammar police are everywhere...thankfully!

I struggled a bit with photo uploads through the iOS Shortcuts all. The interface for creating scripts is a little janky...still. After struggling with the logic of the photos script, I realized that it was a problem with the interface that was borking things for me.

Basically, the script calls a Get Contents of URL action to send a POST request to the API. The photo is included as a file in the form data. When you add a field to the form data in Shortcuts, it defaults to a text field, but allows you to change it to "file" when tapping on it. This doesn't actually change the field to a file, though. Rather, it allows you to pick text attributes of a file to send as text. I had changed it to File, and thought, for hours, that I was in fact sending a file.

I finally reconstructed the action from scratch, adding each field one by one. When initially adding fields, a separate text v. file option appears. Setting that to File actually sets the field as a file. In the list of fields, the file fields are right-aligned while the text fields are left aligned. I wish I would have known that a couple hours ago!

Anyways, the Shortcut is now working. First photo posted through the Drafts-Shortcuts-API conduit coming next!

I'm finally getting around to updating some of the scripts I have for Daystream in the iOS Shortcuts app. Shortcuts is still somewhat limited but has surprised me with what it can accomplish.

I've now got it working essentially as a conduit between Drafts and the Daystream API for posts. This allows me to publish to Daystream directly from my favorite iOS text editor, with tags and everything. My post earlier today about last night's football games was the first to use this setup. It should enable faster, and more frequent, posting.

I plan to create similar Shortcuts for links and photos, too. For those, I'll use drafts to write the main text elements for the content (the body and tags), and then use Shortcuts to get the other content (url for links and a photo for photos).

Then on to the watch.

Fixed an API issue this morning. Posting through iOS Shortcuts again!

updating to the latest Amazon linux 2 platform, including a ruby update. Blue/green, here we go!

Post for yourself. That's what this is about.

Ev Williams had [an interesting post on Medium last week]( that got me thinking about the difference between Daystream and social media networks–it's all about the decision to explore content. Basically, social media makes that decision for you, whereas Daystream empowers you to make that decision for yourself.

Quoting from Ev's post:

> "Trust is more important than ever and well-established editorial brands still have meaning. But today, credibility and affinity are primarily built by people — individual voices — rather than brands."

I happen to agree with this, but have to wonder how Medium, along with Twitter, Facebook, Instagram, and other algorithm-dependent systems, justify their approach to presenting content to people in this people-powered content revolution? They must either count themselves among the "well-established editorial brands" Ev mentions...or accept that their respective algorithm serves a primary purpose other than promoting the objectively best content.

By using an algorithm to create the feed of content a user sees when they log in, no matter the form it takes, these networks are making content decisions for their users.

Daystream is not algorithm-based. Log in, and you'll see who in your network is posting that day, that week, etc. You don't see their content right away, just *them*. From there, you, the user, decide which content you want to discover based on the identity of the creator.

I've learned that this is incredibly empowering as a reader because it allows me (forces me, really) to make the decision to jump into content based on the creator's identity alone. Internally, that forces me to make that decision to jump based on my view of that particular creator. This is more complex than it sounds–some decisions are based on my relationship to the creator (of course I want to see what my son posted today), while others are based solely on my view of the trustworthiness of the creator.

The best content algorithm is the one in your own head. We're building Daystream to empower you to use it.

responsive image The chat startup is leaning on its millions of passionate users to fuel growth after nearly tripling revenue last year.

We added [markdown]( support to posts tonight.

Articles and posts are different types of content on Daystream. Articles have titles; posts don't. Articles are composed with a rich text editor, while posts have always been plain text entries.

I've always liked this distinction, and felt it was useful because it seemed to help me write more (omitting the need for a title was surprisingly freeing). But, I've always felt like something was missing with posts.

*Emphasis* and [links]( are important aspects of web writing, whether the piece has a title or not. Markdown allows us to use these elements in posts without abandoning their plain text simplicity.

So take a review of the [markdown style guide]( and start marking down your posts!

I added audio capabilities to Daystream today. You can now upload mp3's to your journals. This is something I've wanted to do for awhile, and I have some bigger plans for the feature in the future. Should be fun.

So where's the first recording? I used our Post to the Past feature to add a recording of water draining into a sewer to my journal for last Saturday.

We asked two fundamental business questions for Daystream years ago:

Can we build a platform without running ads?

If we don't run ads, do we have to monetize the personal data of our users in some other way?

We answered "yes/no" back then, and today those answers guide our development of the platform.

"yes we can, no we won't" is behind every direction we take and every decision we make. This approach is what defines us and I couldn't be more proud of that as we watch other platforms answer both questions differently.

I've been experimenting with displaying images for links on Daystream. I though they would provide meaningful information alongside the title and any lede that the user adds. It was fairly easy to code, so I didn't think much of it.

After seeing them for a few days, though, I have to say I'm not a fan. I don't think they add much as far as context goes, especially when you consider we've got no control over the actual image associated with the link. They increase load time, too.

And, perhaps most importantly, the lessen the visual impact of images added by the user. Visually, these images should be the most prominent content on any given journal, not uncontrolled images associated with links shared by the user.

So, the experiment is over - images will no longer be displayed for links on Daystream.

I had some quick thoughts while clearing 11" of snow from the driveway that relate directly to Daystream.

In certain areas of technology, it seems ripe to ask the question:

Overall, is the field leveraging user/customer focused advances in technology to benefit users/customers or is it succumbing to a leveraging of *other* advances by a few, powerful corporations at the expense of users/customers?

I think those two points lie on a continuum and are in some level of tension as a field of technology advances. I also think fields get out of whack from time to time and need to correct, pulling back to the other side. Social media seems to be way out of whack right now, skewing heavily toward the "powerful corporations" side of the equation. I don't see those corporations doing anything to correct this. Indeed, they just seem to keep doubling down on the issue, extracting more personal information from their users and using it in new ways to maximize profit.

Daystream has always been about avoiding that. It is, in fact, a founding principle of our technology. I still remember the early conversations with Jonathan all those years ago. He was the first to point out to me that "the algorithm" is awful. Why should it decide what the users see? Shouldn't it be easy for a user to see the content they want to see and know that, in fact, that's what they are seeing? Boom - enter our elimination of the firehose and replacement with date-based accessors. What did Sally post on her last birthday? Well, let's go look. That has always formed the kernel of Daystream and we continue to extend that founding principle as we grow our features.

We'll always skew heavily to the "benefit people" side of the equation.

The trick is figuring out how to make a go at it while doing that. If we can't, we can't. We won't go to the other side just to keep it going, I know that much.

Daystream has been a pet project for more than ten years. It's time to take it beyond that. It's time to change social media by implementing all the things behind the Daystream concept. It's time.

Responsive image

I found an old iPod of Jonathan's as I was going through our electronics graveyard. It still had copies of the Chip Trading and Daystream apps he created! Non-functional, but boy did this bring back some wonderful memories. And look how small that iPod is!

Testing post entry from the web again.

Working on the API tonight...looking forward to that!