It doesn’t matter if RSS is ‘dying’

A protocol is the gift that keeps on giving. Even if it's 'dying'. So people should keep building on RSS.

Infected by 2010s SaaS thinking, the first version of my curation app (Banquet) used custom data types. It did read RSS feeds. But the internals and output were custom. So when a user wanted a weekly email digest, I had to prioritise building that myself (I didn't).

I rewrote Banquet as https://zacusca.net for a few reasons. But one was that it should be RSS in, RSS out.

When people asked me why I was rewriting to make better use of a dead protocol, I did feel a bit silly.

But then when demoing Zacusca last week, I was asked about email digests. So I went to https://feedmail.org/ (found via this sub), pasted in a Zacusca URL, and bang we were getting email digests.

This isn't just about RSS. It's the underappreciated power of protocols:

• You get import/export features for free. (By contrast I still haven't finished this for Banquet v1.)
• Users get mobile apps and alternative clients for free. (I'm reading my Zacusca through Inoreader on Android.)
• Other product features -- like the email digests but maybe also search and format conversion -- can be plugged in via a software package or service that already exists.
• LLMs understand old protocols like RSS. So if you use LLMs to build new features they will be far more useful than with your custom data type.

Other examples of flawed but well-understood protocols:

• Education: .apkg
• Files: WebDAV
• Documents: .docx
• Finance: Plain text accounting

Now, I get that these can be irritating at first -- .apkg has some weird conventions.

But I increasingly believe the irritation is worth it for the gifts you receive later.

RSS is dead. Long live RSS!