@jens

> The "baz" namespace determines what "quux" means, and the definition of "quux" determines the semantics of finding a "foo:bar" underneath it

ok, but how do you deal with a naive processor that doesn't know what any of this means?

i am describing the case where you see this:

<some activity> <has an actor> <me>.
<some activity> <is a> <Like>.
<some activity> <has an object> <ghostbusters>.

then you dereference <ghostbusters> and get:

<ghostbusters> <has an actor> <bill murray>.

@jens an aware agent can merge these two graphs while preserving context. <has an actor> in the first graph gets qualified with a prefix or namespace.

an unaware agent doesn't know what they're doing. they just blindly copy <has an actor> without qualifying it at all. this is what introduces the issue.

@trwnh As I said: no. JSON-LD is not plain JSON, period. The terms used in one JSON-LD document are only ever valid in this one document. The document is the boundary, you cannot make any assumptions from one document for another one. It's true for all RDF sources, and even laid out explicitly in RDF Syntax and Abstract Concepts. That goes even further: Semantics of one document may change every time you read it (not applicable here, but please just never ignore that JSON-LD is RDF).

@nik JSON-LD is not plain JSON, but it can be parsed as plain JSON. this is a core part of ActivityStreams 2.0 -- and most of fedi takes this approach. fedi largely doesn't care about RDF at all, and will happily ignore that...

@trwnh I *know*. I know that most fedi devs don't understand the Social Web, and I know ActivityStreams explicitly promotes flawed implementations. And I know that everyone and their dog follow Eugen's example of making ActivityPub a zoo of proprietary JSON.

@nik right, i also dislike the fragility of fedi implementations. i'm trying to bridge the gap for the LD haters to at least produce valid documents.

@trwnh What issue?

The issue only exist if you insist on interpreting the merged document (! not knowledge graph!) as linked data.

"has an object" isn't something JSON knows. JSON knows keys that point at values. Your value may be a reference, if you go that way, or it may be any other JSON.

Perhaps the key is "hasObject". Perhaps it's "attachment". Either way, the path ".outer-key.actor" is different from ".actor", so whatever sits there as values doesn't have the same semantics.

@trwnh Thanks for making band-aids.

But maybe get @Gargron to publish his proprietary extensions as RDF-parsable context first instead of letting everyone outside Mastodon guess what his JSON-LD is supposed to mean.

We're making band-aid for things not some small individual Fedi hacker did, but to support corporate decisions of Fedi leaders that get in the way of individual Fedi hackers.

@nik i'm not sure we can expect people who don't understand RDF at all (and indeed don't even *want* to understand it) to suddenly become RDF-friendly good citizens.

@trwnh

This very thread proves that ignoring RDF makes handling JSON-LD so incredibly difficult that it cannot be expected from fedi hackers either.

If you are competent enough to handle that complexity yourself, go and learn RDF instead, it's simpler.

If you aren't, use a library, in which case it should handle RDF *and* help you ignore the complexity.

@nik we can say this as many times and as loudly as we want, but it won't change anyone's minds nor their software.

rather than declaring it all to be junk, it can be preprocessed (or fixed) into something properly semantic. that's because meaning is descriptive, not prescriptive. any arbitrary JSON payload has implicit semantics that can be made explicit with enough description -- that's what the JSON-LD context is.

yes, we can also write more/better libraries at the same time. let's do both.

@trwnh

If we want to allow people to ignore RDF, then at least do it correctly (resolve all terms into full IRIs). The plain JSON form of a JSON-LD document is not the original JSON-LD with the context key stripped away, but with all keys containing full IRIs.

If we do that, I am perfectly fine with people just using it as plain JSON (with really odd keys).

@jens it is possible to merge in a way that results in a valid graph with no semantic confusion. the goal of the thread was to describe such an approach.

> the path ".outer-key.actor" is different from ".actor", so whatever sits there as values doesn't have the same semantics

the ".outer-key" here has no inherent semantics for the value; it only has semantics for the subject. for example, the `object` of an Activity can be anything. you can't assume anything about `object.actor` vs `actor`.

@trwnh That depends on what the definition of "outer-key" in the context of the root object means. If the root object defines the key to point to any arbitrary value, that's an incomplete definition.

@jens well, that’s activity streams for ya

an actor can be anything, an object can be anything. the most minimal way to serialize an activity is to just use references, but you often want more information about the actor and/or object, so you dereference those. but a lot of naive json producers want to be “helpful” and save you the extra http requests, so they embed the dereferenced json documents… naively…

@trwnh But if it's an incomplete definition, then it's hardly the naive JSON processor's fault

@jens well, syntactically it’s either a JSON string (meant to be parsed as an id reference) or a JSON hash… or a JSON array containing either of these. (except the naive processor is so naive they probably don’t understand arrays because they expect a single value in all cases.)

semantically, it’s saying “jens created something”, where the object of the sentence “something” can be whatever. you need to dereference “something” to get… 100% arbitrary information about it

@trwnh Just to distinguish this *a little*, the syntax is JSON, and JSON processors will likely only process the document if it's valid JSON.

Whether the value is an object, array, string or other JSON type is technically still part of the key's semantics, which should be defined.

FWIW, I'm not hung up on whether you count that as part of syntax or semantics. It's basically a layering issue. If you look top down from meaning to bit stream, it seems more like syntax. But if you look bottom...

@trwnh ... up, once you've turned a bit stream into structured, typed data, it's semantics.

The key thing is how much meaning you assign to the notion of type here. You essentially treat type as "the full, abstract meaning is encapsulated" whereas I treat it as "I can distinguish it in code well enough to keep it apart from other types".

The point being, it should be part of the specifications of the key whether a string, array or object is *valid* here.

It's totally fine to e.g. say it has..

@trwnh ... to be an object with a "type" field that specifies other valid fields, or an array of such objects, or a JSON reference pointing at such an object or array.

And then I can still write a naive-ish processor that can handle this, or a very naive processor that doesn't care at all about the value.

But I do not have to know all possible type values or what they imply for the value.

This is where you can go and say "well, but if I use LD for those objects, I can reason about stuff...