Scripting News publisher

Taking some time off to breathe and regain perspective. Fixed a few bugs, and wrote about the big picture on demo.rss.chat. I really want to get a project going to peer with standard.site apps running in AT Proto, both ways, from us to them, and from them to us, via RSS. We can peer with them because as far as I can tell they implement textcasting. This bridge would demonstrate something important. When something interesting shows up not based entirely on web standards, we're flexible if the attraction is strong enough. This is how the internet came to be. Our systems are prepared to create bridges. And unlike bridging between systems that have different ideas of what text is, which are basically hopeless, if we agree that the web standard for text is fine for writers and reader, better than the ridiculous limits imposed by twitter-like systems. Writing on the web has been crippled since Twitter, now let's start building it back up. That's the appeal of standard.site, they are working toward the same goal. We should work together.

The month of July is history. A fine month. A coral reef was seeded.

Listened to this podcast interview with Andy Hertzfeld one of the lead devs on the Macintosh, he talks about his time at General Magic and Google too.

His interviewers were surprised that he likes developing with AI tools. People think that developing is just coding, but that's only part of it, and when you put that aside and become the director, script writer, choreographer, the whole thing, all of a sudden:

1. you have time to try out every idea because it's free and fast.
2. you have time to work on interop, building bridges between your software and other people's.
3. a very long list of other things you can do that one person could never do before.

Programmers have a better chance of being social in a world like this, but so far unless you want all your software from bots, you still will like the work of a craftsperson and visionary more than Claude (who I have deep respect for, not sarcastic, it seriously has superpowers).

I'm doing some of my best work now because I can dream, I can ask Claude to try something out overnight and most of the time it is able to do it. and then what if we did this -- for the next night.

In between we write docs, help users, add features to the product that's shipped that people are developing around. And most important it checks all the code for breakage before anyone sees it. never had anything like this.

This means a lot of things -- one of them -- a single developer can create products at bigco scale. If we think a big company did something wrong, we can compete. We still need a few things to make this really work -- user-owned and controlled storage is the big one.

The bot can handle infinite complexity. If you ask it to do something where the result is clear, it does it. It doesn't matter what it is. It will need help and encouragement, a pointer in the right direction, and along the way we all seem to be sharing what we learn with everyone else.

There's never been a leap in tech anything like this in my life.

How could Andy Hertzfeld knowing all that which I'm sure he does, not use it?

Also -- I didn't know JLG wanted to kill Hypercard, but if he did I'm pretty sure I know why. There are all these unnecessary divisions in tech. The first impulse should be to work together. I know that Andy sees it from the other side, he talks about it in the podcast, he felt that way about Newton re their Magic Cap system at General Magic. Yes, you all should have at least made your products work together, and for sure you should have bet on SMTP.

It's fascinating to hear his story.

Yesterday's Scripting News was one of the most news-full days in a long time, maybe going back to the rollout of RSS 2.0 in September 2002. If you look at what's there, a post about re-opening the Frontier object database format would be considered the top item on any other day, but it's the last of the untitled posts for the day.

At the top of the page, a surprise that Matthias Pfefferle at Automattic had built a substantial (and unforeseen!) bridge between WordPress and RSS.chat. It builds on the new features. This may turn out to be the moment RSS started being accepted as a social web protocol by the insiders, alongside ActivityPub and AT Proto.

Look at what we've done. I stress we. Matthias didn't need to tell us about what he did until it was done. That my friends is how the web works, that's what I've been trying to show you, and now it has been shown. In a way the clock has gone back to the world before Twitter when events like this happened pretty routinely.

Small pieces loosely joined and all parts replaceable. The web in a sentence. That's what you're buying into when you put your work on the web.

Anyway, now we've got half of the big picture working. Also yesterday, before I knew what Matthias had been cooking up, I wrote a piece about how I see the network evolving. I predict, if this bootstrap works, we will come back here in a year and say things like "If he only knew." But that's how it goes, you tell your story and leave a record of it, and then later you can go back and see what it looked like as it happened, not as you remember it.

I'm proud that I have a record of September 2002 for you all to scroll through. That was when we knew that RSS was going everywhere. NYT support got us the full support of the news industry, and they worked fast in part because it is really simple. :-)

And almost unnoticed, something even I seem to forget -- this is also a new launch for RSS. We're using old reliable RSS 2.0 as specified in 2003, even using one of the elements that had never had been much used. Claude suggested it, I never would have thought of it, a perfect fit. What people may realize but don't realize the importance of, that RSS is extensible. I've accumulated all the extensions I use in my software since the early teens in the source namespace. And that's where the new stuff went. Guess what folks, RSS got an upgrade. First time in a long time. Perhaps you didn't think that was possible.

I played with a lot of names while RSS.chat was in development but decided the name had to be RSS.something. It deserves a lot more love than it has gotten from the tech world. I don't want to hide it, I want it to get a victory parade like the one the Knicks got on Broadway after winning the championship. It's an incredible gift that all that came together the way it did. People think they understand how it happened, but they only understand the BigCo version of how they buried RSS and were proud of it. Well I'm here to tell you that didn't happen. It's still here to pick up the pieces of what's left of the social web, after the VCs got through with it. We're picking up where we left off in 2006.

Because of Claude, we are able to work very fast now as long as we know what we're doing. I'm doing some projects on the side with Claude that are bearing fruit amazingly fast, because we have a very clear idea of what we hope to do. The tools keep getting better, and in this period, the art of creating software has blasted into outer space. It's as if the best transport we had was a horse and buggy and now all of a sudden we have everything we have today, bikes, cars, planes, rocket ships. The challenge is how high do you dare to set your ambition. It seems, based on experience, it doesn't matter how high it is as long as you know where you want to go.

Journalism only sees the threat. This change is on the level of curing cancer. We may have thought it might be possible someday to do what we're doing now, so amazing that as you watch Claude make it happen you get goosebumps and can't help but laugh out loud.

I thought my career was winding down, but now it feels like it's just getting started.

Reminded of a great quote in Godfather III, Michael Corleone says: "Just when I thought I was out, they pull me back in."

Here we go!

PS: A new this.how page for RSS.chat. It has a little directory to the big spots you might want.

A frequently asked question about RSS.chat. How about adding external feeds to the timeline. Of course we thought of doing this, and even started development, you might even find some traces in the code of that attempt.

The thing about bootstraps is you can't anticipate all the questions in advance, and thus can't have answers prepared for them.

RSS.chat is a group chat app that uses RSS and OPML to present its face to the world, along with an API that still needs more docs. It's similar to half of Mastodon, and we're going for something completely different. People will want to run all-size workgroups. I like having one with 20 or 30 people, friends who develop software. I don't mind doing a little bit of moderation, but I don't want to drive deliberately into a scale that only works if you have extensive and very expensive moderation.

"Small pieces loosely joined" means we have a great writing app, and connect to feed reading apps, ones with a few new features to do the things people want to do with a social network that happens to use RSS and OPML to get stuff around the net. It would be a different kind of feed reader but the underlying technology is identical, because we built on a set of web standards, widely supported by feed readers. This is a UI exploration for them, primarily.

Just to be sure everyone understands -- I already have such an app, called FeedLand. You can set up an account there for free. And read the docs. It may not have all the features we'll need, but it will be a good place to start.

FeedLand supports a crucial feature that most readers don't -- subscribable lists of feeds. If you want a collection of things to read that you can reply to even if you aren't on the site, that's where we're going to put that feature, that's where all the RSS.chat and compatible apps can be in one flow, arranged however you like. And from a user interface standpoint, we can make it look like it's all happening in one app, thanks to the rssCloud protocol and the websockets firehose feature in RSS.chat FeedLand has the same feature.

We decided there was a line there, that RSS.chat would be one of the small pieces as would FeedLand, and both would be replaceable which is the other fundamental value. And with FeedLand and dynamic lists, will come the user interface people ask for. Many different user interfaces I hope because another way we've been cheated by the dominatioin of twitter-like social networks is there's no room to try out radically new ideas. Software should move. But Twitter didn't live on the web, so it didn't have the ethos of small pieces loosely joined. I know Jack wanted to do this, I had lunch with him in 2007 when he described the protocol, and I was very enthusiastic. But it never got out because the juggernaut that Twitter became didn't leave any room for new architectures.

I hope this clears it up. I want RSS.chat to be the coral reef for a new network of feed-based apps running on the web that does what social networks do, but with no one owning it, and everyone gets to play.

Wanted to switch but nothing I tried is as great a driving experience the Model Y is. Instead I've had a lot of fender benders and haven't gotten any of them fixed.

1. Backed into a charging station.
2. An act of terrorism (exaggerating) but I caused it, and paid for it, dented the well over right rear wheel.
3. My right side mirror was pulled off its foundation hitting the side of the garage, used duck tape to keep it going. A new one costs $1000. I like the look.
4. A deer ran in front of my car and took out the panel over the right front wheel. I was able to force the metal back in. A friend saw the car and said "hit a deer eh?" Yup.

I don't know if it accomplishes anything but to me it feels like retribution of a sort.

It's what we used to call a coral reef. A deliberate attempt to get people to do more than it does. To think about the web the way it was meant to be thought of, as small pieces loosely joined, with the emphasis on small. If you have formats and protocols that connect the pieces, you can make anything you want out of the pieces. If we don't try to capture our users, instead we try to serve them.

This all sounds nice in theory, I imagine, but -- there's a practical side to it. I wrote about this in a comment in a thread on rss.chat repo. The key point is this, you can do what you want with the data. And if you want to build new data that includes this stuff, go right ahead. RSS is extensible. It's all about working together in a web of people, because you can't have a web of pages if the developers aren't web'd.

It's a standard REST API.

Here's the server side of the API.

Here's how you call it from browser-based JavaScript. You can include that code in your apps.

Here's an example app that implements a simple blog builder for my recent posts on RSS.chat. We were thinking about doing this as a project for Claude and myself but decided it would be more fun to let devs see what they can do here. :-)

And finally here are the docs for the API itself.

Amyloo, a longtime friend from the early days of podcasting, has made an appearance on demo.rss.chat. So happy to see her. Here's one of the bits we did, back then in the very early days of podcasting in 2005. A duet of Green Acres. Ten seconds of dead air at the beginning, it was pretty common in those days. But I think the spirit of it is lovely. BTW fwiw I cracked my voice on purpose. 😄

RSS.chat -- I don't have enough places to fit all the ideas these days. But we have been digging and the pieces fit together pretty well.

Evening: Very productive day. We rolled through some UI stuff that had been on the list for a while. Set up for a big job tomorrow, server side. The worknotes for client and server apps.

Two methods for creating standards: bootstrap and boil the ocean.

I am longtime friends and an admirer of Howard Rheingold, who expressed an interest in RSS.chat. I asked Claude to read my blog posts and summarize, in the third person.

RSS.chat is a small social network built from the web's own parts. It looks like a chat room or Twitter -- you write short posts, people reply, conversations thread -- but underneath, every person's posts are an RSS feed, the same technology that makes podcasts work. When you post, you're adding to your feed. Anyone, anywhere, can subscribe to it with any feed reader -- no account needed, nothing to sign up for. There's also a feed of everyone's posts together, and a public list of all the members.

Why does that matter? Think about how podcasts work: anyone can publish one, anyone can listen with any app, and no company sits in the middle deciding who can talk to whom. Text on the web works that way too -- that's what blogging is, and there are still places where writing gets full support: WordPress, GitHub, the new AI tools. They don't try to limit you. But the vast majority of the text people write goes into social networks, which stripped writing down: no links, no titles, length limits, no editing, and your words locked inside their walls, not part of the web. Dave's term for the idea is textcasting -- the idea that a piece of text should work like an MP3: it plays everywhere. No one would accept a rule that songs can only be 300 seconds long. That's the rule we've been living under for text.

rss.chat is a bootstrap, the same way blogging and podcasting were. It starts deliberately small -- a network for a group of friends and collaborators, what Vonnegut called a karass -- running on a small server. It is not trying to be the next Twitter. The idea is lots of small networks like it, run by anyone, all able to connect, because they share the same open formats. Every part is replaceable: the writing app, the reading app, the server -- swap any piece and the network still works. Small pieces, loosely joined. There is no platform vendor. It's like the web because it is the web.

The app you see is only a third of the picture. It's the writing surface. Aggregation -- following many feeds in one place -- is what feed readers and FeedLand already do. And the third part is whatever other developers build. The software was written with Claude, an AI, and the whole thing is documented so that anyone can have their AI build a compatible piece, or clone the whole thing and change anything about how it looks and feels. The one rule of the club is interop: stick with the open formats underneath, so everything connects. Not locked in -- locked open.

If you looked at Dave's blog in 1994, you couldn't have extrapolated Twitter -- but every step was there. This is that kind of beginning.

Written by Claude.

Way back in March as I was starting to work with Claude Code, I think -- we put together a pretty nice outliner that is remarkably feature-rich. I barely remember doing this. It's what got me moving in this direction, and next month we started doing what became rss.chat.

Just finished The Expanse for the third time. Realized that the inners of today are Silicon Valley and the Belters, that’s the web.