ActivityPub would have been better if it used XML instead of JSON.
-
ActivityPub would have been better if it used XML instead of JSON.
/duck
-
ActivityPub would have been better if it used XML instead of JSON.
/duck
@tennoseremel activities need an extensible markup language for mentions, tags, quotes, media, and a lot more.
-
@tennoseremel activities need an extensible markup language for mentions, tags, quotes, media, and a lot more.
@tennoseremel we often end up with html-in-jsonld, but worry not: as json is inherently human-readable, and html (by the sole virtue of not being xml) is too, everything is great and the web developers feel all modern and trendy.
-
ActivityPub would have been better if it used XML instead of JSON.
/duck
@tennoseremel@gts.skobk.in fwiw if ActivityPub were XML instead of JSON I would not have integrated it.
-
ActivityPub would have been better if it used XML instead of JSON.
/duck
@tennoseremel And if it had used JSON instead of JSON-LD?
-
@tennoseremel And if it had used JSON instead of JSON-LD?
@fentiger@mastodon.social then there would be property conflicts!
But far fewer than expected and everyone would just adjust and defer to prior art... and most importantly nobody would use JSON-LD compliance as a cudgel.
Sounds nice actually.
-
@julian The ability to make up brand new, guaranteed-unique identifiers on the fly is probably the thing I like most about JSON-LD.
As you imply, though, it's not really a problem in practice, because devs are used to JSON, and know what happens when you add extensions to standards.
-
@julian The ability to make up brand new, guaranteed-unique identifiers on the fly is probably the thing I like most about JSON-LD.
As you imply, though, it's not really a problem in practice, because devs are used to JSON, and know what happens when you add extensions to standards.
@fentiger@mastodon.social the purportedly world-ending effects of property collisions are overblown, imho.
-
@julian The ability to make up brand new, guaranteed-unique identifiers on the fly is probably the thing I like most about JSON-LD.
As you imply, though, it's not really a problem in practice, because devs are used to JSON, and know what happens when you add extensions to standards.
>The ability to make up brand new, guaranteed-unique identifiers on the fly is probably the thing I like most about JSON-LD.
But it is not a JSON-LD thing, right?
A JSON document can have property names that are URIs, or reverse-DNS strings, or UUIDs, or longNamesThatIncludeSoftwareNameAndVersion. All of those can be globally unique.
-
Here's the first recorded collision in the history of Fediverse: https://github.com/bookwyrm-social/bookwyrm/issues/3731
-
>The ability to make up brand new, guaranteed-unique identifiers on the fly is probably the thing I like most about JSON-LD.
But it is not a JSON-LD thing, right?
A JSON document can have property names that are URIs, or reverse-DNS strings, or UUIDs, or longNamesThatIncludeSoftwareNameAndVersion. All of those can be globally unique.
@silverpill @julian Fair point.
I suppose I was thinking of the ability to assign it to a shortened name in the @'context - but this is probably relatively unimportant once it's been gzipped.
