Proposal for JSON Feed version 1.1

[manton]

There was a lot of great feedback on JSON Feed. It's time to update the spec to incorporate a few suggestions. We're thinking this will be a minor update, version 1.1.

• New type application/feed+json (Suggest alternate media type. #32). When discovering the feed, apps should still fall back to looking for application/json if nothing else more specific is present.
• Multiple authors (Behavior for multiple authors #6). Allow an array of authors instead of just one wherever we had a single author before.
• Add language (Allow markup and/or specifying the language in all human readable strings #62). This can be a top-level default or specified on each item.
• Update link to more recent version of JSON spec (Reference to obsolete version of JSON spec #53).
• Clarify that id should really be a string (Why allow non-string values for a feed's id? #60). Move the tip about gracefully handling a number to the "suggestions for feed readers" section and not in the main part of the spec to avoid confusion.

Anything else that should be included? Thanks!

[jbrayton]

I would really like to see Issue #89 included. It is a proposal to help RSS readers determine that two feeds (a JSON feed and a feed in a different format) contain identical content.

When discovering feeds, this would help readers avoid presenting identical feeds as if the user really needed to choose between subscribing to a website‘s RSS feed, its JSON feed, or both. It would also help aggregators (such as Feedbin and Feedly) avoid polling and storing articles for identical feeds.

I am happy to create a pull request if that would be helpful.

[letrastudio]

I just posted #127, suggesting a via_url property for feed items.

If url links to where you’re talking about a thing, and external_url links to the thing you’re talking about, then via_url links to where you heard about it.

Via links are a well-established pattern used in many blogs (I use them myself). They're great for discovering new blogs or people to follow. Making them part of the spec might help make the idea even more popular, and could enable special handling of those links in feed readers.

[jameysharp]

I'd love to see #129 (which I just opened) addressed in a new version of this spec.

After filing that I found #100 which brought this up a couple years ago and was closed recently. The existing next_url attribute is the bare minimum so I'd like to see some more consideration given to this.

[manton]

I've posted a draft of version 1.1 here: https://github.com/manton/JSONFeed/blob/master/pages/version/1.1.markdown

[fboes]

I really like the new 1.1 draft, as it does not break backwards compability but for deprecating author, and adds language which I dearly missed.

Does the version property need to change for 1.1? Currently it still says https://jsonfeed.org/version/1 in the 1.1 spec.

As for more feature requests, I propose to wait for later versions to incorporate these, so the changes in 1.1 do not get blocked waiting for more features.

[dshanske]

Can we see about pushing this for the mime type alone?

[manton]

Thanks y'all. Yes, it's time. I'm pushing the last tweaks.

[fboes]

Couldn't wait any longer… https://journal.3960.org/feed.json 😁

[tedder]

So are you explicitly forking this, or is there hope @brentsimmons will update and/or add maintainers and keep this as the canonical repo?

[brentsimmons]

We will be updating the spec. We’re just slightly overwhelmed by other things at the moment, and we appreciate your patience.

[manton]

The web site has now been updated for version 1.1. Thanks everyone!

[brycewray]

The validator is a wonderful tool, but it considers a reference to Version 1.1 as an error:

Warning: The "version" field should have the value: https://jsonfeed.org/version/1

Any chance that'll be revisited?

[manton]

@brycewray Yes, I've been meaning to fix this. Thanks for the reminder!