RSS > ActivityPub

RSS is better than ActivityPub1.

Photograph of a boxing match, but with the heads of the competitors replaced with the ActivityPub and RSS logos (and "AP" or "RSS" written on their clothes, respectively). RSS is delivering a powerful uppercut to ActivityPub.
A devastating blow by RSS against a competitor 19 years his junior! For updates on this bout as it develops, don’t forget to subscribe… using either protocol.

When I subscribe to content, I want:

  • Resilient failsafes. ActivityPub has many points-of-failure. A notification might fail to complete transmission as a result of downtime, faults, or network conditions, and the receiving server might never know. A feed reader, conversely, can tell you that an address 404’d or the server was down.
  • Retroactive access. Once you fix the problem above… you still don’t get the message you missed: it’s probably gone forever – there’s no retroactive access. The same is true when your ActivityPub server connects with a peer for the first time: you only ever get new content after that point. RSS, on the other hand, provides some number of “recent” items the moment you first subscribe.
  • Simple subscriptions. RSS can be served from a statically-hosted single file, which makes it suitable to deploy anywhere as well as consume using anything. It can be read, after a fashion, in anything from Lynx upwards.

RSS ticks all these boxes. If I can choose between RSS and ActivityPub to subscribe to your content, and I don’t need a real-time update, I’m probably going to choose RSS.

Footnotes

1 I feel like this statement needs a few clarifications and caveats, but my hot take looks spicier if I bury them in a footnote!

  • By RSS, I mean whichever pull-based basic HTTP you like, be that Atom, JSON Feed, h-entry, or even just properly-marked-up HTML5: did you know that the <article> element is intended to be suitable for syndication use?
  • Obviously I appreciate that RSS and ActivityPub are different tools for different jobs, and there are doubtless use-cases for which ActivityPub is clearly the superior solution.
  • I certainly don’t object to services providing both RSS and ActivityPub as syndication options, like Mastodon does, where both might be good choices.
Photograph of a boxing match, but with the heads of the competitors replaced with the ActivityPub and RSS logos (and "AP" or "RSS" written on their clothes, respectively). RSS is delivering a powerful uppercut to ActivityPub.×

Automattic Acquires ActivityPub Plugin for WordPress

This article is a repost promoting content originally published elsewhere. See more things Dan's reposted.

Automattic has acquired the ActivityPub plugin for WordPress from German developer Matthias Pfefferle, who will be joining the company to continue improving support for federated platforms. Pfefferle, who is also the author of the Webmention plugin, said his new role is to see how Automattic’s products can benefit from open protocols like ActivityPub.

This is so exciting I might burst. Want to know why?

  1. Matt Mullenweg‘s commitment to ActivityPub makes me happy. WordPress made Pingback and Trackback take off, back in the day, and I believe that – in the same way – Automattic can help make ActivityPub more accessible and mainstream too.
  2. Matthias Pfefferle is both an IndieWeb and an ActivityPub star; I use (and I’ve extented upon) a lot of code he’s written every day and I sponsor him on Github! The chance that we get to work directly together is pretty slim, but it’s a chance right?

Susan A. Kitchens expressed concern that this could increase the level of ActivityPub spam out there (which right now is very low). I worry about that too. But I’m still optimistic that we can make something awesome off the back of this acquisition and keep the interpersonal Web federated, the way it ought to be.

When Twitter Loses, WordPress Wins

This article is a repost promoting content originally published elsewhere. See more things Dan's reposted.

You know who’s having a killer month? Automattic. Everyone who’s leaving Twitter seem to fall in at least one of these three camps:

  1. They have gone back to the blogosphere. (using WordPress, or WordPress.com)
  2. They have gone to Tumblr
  3. They have gone to the fediverse (of which a fairly large percentage are WordPress installs)

In all of these cases, Automattic wins.

Some smart observations here by Alex. A fourth point worth noting is that Matt has openly suggested that former Twitter engineers might like to come join us in Automattic and help make the web a better place. We’ve changed our careers pages a little lately but we’re still the same awesome company!

Alex went on to say:

I’ll be downright shocked if Matt isn’t working very hard to get Tumblr on the fediverse ASAP. He has so much to gain in supporting this movement, and very little to lose.

That’s definitely on his mind too, which I can safely say without leaking anything because he’s hinted at it himself. Exciting times.

Reply to Decentralization and verification

In “Decentralization and verification”, Derek Kedziora said:

…Mastodon by its very nature as a decentralized service can’t verify accounts.

We’d still need some trusted third party to do offline verifications and host them in a centralized repository.

Let’s not sell Mastodon short here. The service you compare it to – Twitter – solves this problem… but only if you trust Twitter as an authority on the identity of people. Mastodon also solves the problem, but it puts the trust in a different place: domain names and account pages.

If you want to “verify” yourself on Mastodon, you can use a rel=”me” link from a page or domain you control. It looks like this:

Screenshot showing @dan@danq.me's Mastodon account as the verified owner of website DanQ.me.
The tick is green, not blue, but I can’t imagine anybody complains.

A great thing about this form of verification is you don’t have to trust my server (and you probably shouldn’t): you can check it for yourself to ensure that the listed website really does state that this is the official Mastodon account of “me”.

You can argue this just moves the problem further down the road – instead of trusting a corporation that have shown that they’re not above selling the rights to your identity you have to trust that a website is legitimate – and you’d be right. But in my case for example you can use years of history, archive.org, cross-links etc. to verify that the domain is “me”, and from that you can confirm the legitimacy of my Mastodon account. Anybody who can spoof multiple decades of my history and maintain that lie for a decade of indepdendent web archiving probably deserves to be able to pretend to be me!

There are lots of other distributed methods too: web-of-trust systems, signed keys, even SSL certificates would be a potential solution. Looking again at my profile, you’ll see that I list the fingerprint of my GPG key, which you can compare to ones in public directories (which are co-signed by other people). This way you’d know that if you sent an encrypted DM to my Mastodon inbox it could only be decrypted if I were legitimately me. Or I could post a message signed with that key to prove my identity, insofar as my web-of-trust meets your satisfaction.

If gov.uk’s page about 10 Downing Street had profile pages for cabinet members with rel=”me” links to their social profiles I’d be more-likely to trust the legitimacy of those social profiles than I would if they had a centralised verification such as a Twitter “blue tick”.

Fediverse identify verification isn’t as hard a problem to solve as Derek implies, and indeed it’s already partially-solved. Not having a single point of authority is less convenient, sure, but it also protects you from some of the more-insidious identity problems that systems like Twitter’s have.

Screenshot showing @dan@danq.me's Mastodon account as the verified owner of website DanQ.me.×

Wake Me Up When September Ends

Mastodon’s Eternal September begins

In the light of the so-called “Twitter migration”, I’ve spent a lot of the last week helping people new to Mastodon/the Fediverse in general to understand it. Or at least, to understand how it’s different from Twitter.1

If you’re among those jumping ship, by the way, can I recommend that you do two things:

  1. Don’t stop after reading an article about what Mastodon is and how it works (start here!); please also read about the established etiquette, and
  2. Don’t come in with the expectation that it’s “like Twitter but…”, because the ways it’s not like Twitter are more-important (and nobody wants it to be more like Twitter).

The experience has filled me with feelings, for which I really appreciated that Hugh Rundle found such great words to share, comparing the surge of new users to September(ish) 1993:

The tools, protocols and culture of the fediverse were built by trans and queer feminists. Those people had already started to feel sidelined from their own project when people like me started turning up a few year ago. This isn’t the first time fediverse users have had to deal with a significant state change and feeling of loss. Nevertheless, the basic principles have mostly held up to now: the culture and technical systems were deliberately designed on principles of consent, agency, and community safety.

If the people who built the fediverse generally sought to protect users, corporate platforms like Twitter seek to control their users… [Academics and advertisers] can claim that legally Twitter has the right to do whatever it wants with this data, and ethically users gave permission for this data to be used in any way when they ticked “I agree” to the Terms of Service.

This attitude has moved with the new influx. Loudly proclaiming that content warnings are censorship, that functionality that has been deliberately unimplemented due to community safety concerns are “missing” or “broken”, and that volunteer-run servers maintaining control over who they allow and under what conditions are “exclusionary”. No consideration is given to why the norms and affordances of Mastodon and the broader fediverse exist, and whether the actor they are designed to protect against might be you.

I’d highly recommend you read the whole thing because it’s excellent.

Photo showing fragile dewey spiderwebs on blades of grass.

Part of the solution

I genuinely believe that the fediverse is among our best bets for making a break from the silos of the corporate Web, and to do that it has to scale – it’s only the speed at which it’s being asked to do so that’s problematic.

Aside from what I’m already doing – trying to tutor (tootor?) new fediversians about how to integrate in an appropriate and respectful manner and doing a little to supporting the expansion of the software that makes it tick… I wonder what more I could/should be doing.

Would my effort be best-spent be running a server (one not-just-for-me, I mean: abnib.social, anyone?), or should I use that time and money to support existing instances directly? Should I brush up on my ActivityPub spec so I can be a more-useful developer, or am I better-placed to focus on tending my own digital garden first? Or maybe I’m looking at it all wrong and I should be trying to dissuade people from piling-on to a system that might well not be right for them (nor they for it!)?

I don’t know the answers to these questions, but I’m hoping to work them out soon.

Addendum

It only occurred to me after the fact that I should mention that you can find me at @dan@danq.me.

Footnotes

1 Important: I’m no expert. I’ve been doing fediverse things for about 3 years but I’m relatively quiet on Mastodon. Also, I’ve never really understood or gotten along with Twitter, so I’m even less an expert on that. Don’t assume that I’m an authority on anything at all, and especially not social media.

Photo showing fragile dewey spiderwebs on blades of grass.×