I have deeply mixed feelings about #ActivityPub's adoption of JSON-LD, as someone who's spent way too long dealing with it while building #Fedify.
-
and then
1) machine gets thread (cause is JSON-LD by known/allowed `generator`)
2) machine fetches or gets cached wikidata entry about Berlin and displays the card (kind of "infobox" then).
.. from the named "SpecialEntitiyData" of wikidata which is JSON-LD as well.
3) machine is happyapart from our tools, I need to credit Max Lath who is doing inventaire, the federated book library and did a lot of previous work for wiki JSON-LD like the wonderful https://github.com/maxlath/wikibase-sdk
PS - just btw;
about inventaire I am sharing currently photobooks for free rent in Dortmund, Germany
https://inventaire.io/users/sl007
they have also nice use for JSON-LD re. books /authors etc. https://data.inventaire.io/ like so many software in fedi.
If you ask the redaktor Service Actor for Place (`Question`) to find you a waffle restaurant in Amsterdam serving blue syrup near a train station then we do also use SPARQL like them - without the AI bullshit - just cause millions of friendly humans contributing to wd and OSM … -
I have deeply mixed feelings about #ActivityPub's adoption of JSON-LD, as someone who's spent way too long dealing with it while building #Fedify.
Part of me wishes it had never happened. A lot of developers jump into ActivityPub development without really understanding JSON-LD, and honestly, can you blame them? The result is a growing number of implementations producing technically invalid JSON-LD. It works, sort of, because everyone's just pattern-matching against what Mastodon does, but it's not correct. And even developers who do take the time to understand JSON-LD often end up hardcoding their documents anyway, because proper JSON-LD processor libraries simply don't exist for many languages. No safety net, no validation, just vibes and hoping you got the
@contextright. Naturally, mistakes creep in.But then the other part of me thinks: well, we're stuck with JSON-LD now. There's no going back. So wouldn't it be nice if people actually used it properly? Process the documents, normalize them, do the compaction and expansion dance the way the spec intended. That's what Fedify does.
Here's the part that really gets to me, though. Because Fedify actually processes JSON-LD correctly, it's more likely to break when talking to implementations that produce malformed documents. From the end user's perspective, Fedify looks like the fragile one. “Why can't I follow this person?” Well, because their server is emitting garbage JSON-LD that happens to work with implementations that just treat it as a regular JSON blob. Every time I get one of these bug reports, I feel a certain injustice. Like being the only person in the group project who actually read the assignment.
To be fair, there are real practical reasons why most people don't bother with proper JSON-LD processing. Implementing a full processor is genuinely a lot of work. It leans on the entire Linked Data stack, which is bigger than most people expect going in. And the performance cost isn't trivial either. Fedify uses some tricks to keep things fast, and I'll be honest, that code isn't my proudest work.
Anyway, none of this is going anywhere. Just me grumbling into the void. If you're building an ActivityPub implementation, maybe consider using a JSON-LD processor if one's available for your language. And if you're not going to, at least test your output against implementations that do.
#JSONLD #fedidev
@hongminhee How hard would it be for a future version of ActivityPub to simply back out JSON-LD support? Would there be a downside to this?
-
@hongminhee How hard would it be for a future version of ActivityPub to simply back out JSON-LD support? Would there be a downside to this?
@mcc@mastodon.social asking the important questions
-
@silverpill @hongminhee @mariusor
I don't remember the details, but
Mastodon and Takahe were definitely giving responses in different shapes for a question object. -
@julian Yes, POST to personal inbox sends a 404, POST to group inbox sends a 202 (I guess group inbox is how we communicate now).
@silverpill@mitra.social okay! That's a little concerning because all of those routes just end up going to the same controller... but this gives me enough to debug, so thank you!
-
@silverpill @hongminhee @mariusor
I don't remember the details, but
Mastodon and Takahe were definitely giving responses in different shapes for a question object.@silverpill @hongminhee @mariusor
> There might be small variations in how a protocol feature (such as polls) is implemented, but they are not difficult to deal with.
This becomes a matter of opinion. You say "not hard to deal with" and "dealing with JSON-LD is a burden".
I think that being able to throw (valid) json-ld in any shape and letting the processor deal with it and have only one set of class to represent the types is well off the trade off of having a few extra bytes.
-
@silverpill @hongminhee @mariusor
> There might be small variations in how a protocol feature (such as polls) is implemented, but they are not difficult to deal with.
This becomes a matter of opinion. You say "not hard to deal with" and "dealing with JSON-LD is a burden".
I think that being able to throw (valid) json-ld in any shape and letting the processor deal with it and have only one set of class to represent the types is well off the trade off of having a few extra bytes.
@silverpill @hongminhee @mariusor
> To be honest, I can't even imagine how JSON-LD could help with that instead of making it harder.
-
@hongminhee How hard would it be for a future version of ActivityPub to simply back out JSON-LD support? Would there be a downside to this?
@mcc@mastodon.social I'm not sure, but that would break some ActivityPub implementations relying on JSON-LD processors.
-
@hongminhee How hard would it be for a future version of ActivityPub to simply back out JSON-LD support? Would there be a downside to this?
@mcc @hongminhee Mastodon, Fedify and other implementations that treat LD as mandatory (
MUST) even if it's optional (SHOULD) will be non conformant. As Mastodon is the biggest implementation by far margin, deprecating it is no small feat. -
@mcc @hongminhee Mastodon, Fedify and other implementations that treat LD as mandatory (
MUST) even if it's optional (SHOULD) will be non conformant. As Mastodon is the biggest implementation by far margin, deprecating it is no small feat.@cochise @hongminhee But Mastodon famously doesn't actually *support* LD right? That's the point of the thread? So wouldn't they be the easiest to convince to stop supporting the thing they never supported?
-
@cochise @hongminhee But Mastodon famously doesn't actually *support* LD right? That's the point of the thread? So wouldn't they be the easiest to convince to stop supporting the thing they never supported?
@mcc @hongminhee Don't really support, but discards activities without
@contextanyway.I suspect JSON-LD was a way to have extensibility and escape XMPP's XEP hell with servers and clients not supporting or disabling features in an infinite matrix.
But seems community favors FEPs describing JSON schemas and hardcoding it over getting them from a server and mapping the object at runtime. -
@hongminhee I had a similar realization early on when implementing Pinka. I almost went full JSON-LD but found that to properly expand the document I might need to make network calls. I stopped worrying about unknown terms and just hard coded a list of well-known AS and APub terms for interoperability.
-
I incentive moved this topic from Uncategorized
-
@hongminhee what i have found necessary (sadly) is to sometimes ignore what @\context a software produces and simply inject a corrected @\context describing what they *actually* meant instead of what they said they meant. x_x
the "incorrect" mastodon context in use right now (or equivalent), which can be swapped out for the "correct" mastodon context to be more compatible with generic json-ld (and more semantically correct)
the "incorrect" mastodon context in use right now (or equivalent), which can be swapped out for the "correct" mastodon context to be more compatible with generic json-ld (and more semantically correct) - mastodon-context-correct.jsonld
Gist (gist.github.com)
it would be an Exercise to sit down and map out the actual contexts of softwares like mastodon 4.5, mastodon 4.4, misskey 2025.12, akkoma 3.10.2, and so on...
for all else, there's shacl i guess, if you want to beat things into the correct shapes.
-
@hongminhee I had a similar realization early on when implementing Pinka. I almost went full JSON-LD but found that to properly expand the document I might need to make network calls. I stopped worrying about unknown terms and just hard coded a list of well-known AS and APub terms for interoperability.
@kanru @hongminhee ironically this is what you're supposed to do! preload the terms you understand into local contexts. newer jsonld-adjacent specs (vc, cid, and so on) tell you that you MUST NOT fetch the contexts over the network at runtime, and instead MUST treat them as already fetched with a given sha256sum. https://www.w3.org/TR/cid-1.0/#json-ld-context
-
@julian @mcc @hongminhee the downside is that you now need a central registry of allowed terms and what they mean.
the way to avoid that is to always use "expanded" form, i.e. use full IRIs as property keys (and types) and {"id": "foo"} over "foo". in effect, you treat the http(s) authority as the social entity defining the term.
-
@mcc @hongminhee Don't really support, but discards activities without
@contextanyway.I suspect JSON-LD was a way to have extensibility and escape XMPP's XEP hell with servers and clients not supporting or disabling features in an infinite matrix.
But seems community favors FEPs describing JSON schemas and hardcoding it over getting them from a server and mapping the object at runtime.@cochise @mcc @hongminhee mastodon is one of the "better" ones in that regard, but famously requires you to have the same context as it (instead of expanding shorthand terms to the full IRIs and comparing those...)
-
@hongminhee if i can give one piece of advice to devs who want to process JSON-LD: dont bother compacting. you already know the schema you output (or you're just passing through what the user gives and it doesn't matter to you), serialize directly to the compacted representation, and only run expansion on incoming data
expansion is the cheapest JSON-LD operation (since all other operations depend on it and run it internally anyhow), and this will get you all the compatibility benefits of JSON-LD with little downsides (beyond more annoying deserialization code, as you have to map the expanded representation to your internal structure which will likely be modeled after the compacted one)generally agreed except
> you have to map the expanded representation to your internal structure which will likely be modeled after the compacted one
this is compaction but manual instead of using a jsonld processor to do it. maybe the more precise argument is "don't bother with auto/native compaction"?
with that said: you also lose out on flattening and framing, which are pretty cool features for transforming the serialization. if you don't care about those, ok fine
-
@julian @mat atproto lets you section things off by "app" roughly, which is something that could be done with "plain old http" using content-types and well-known uris.
json-ld makes it so that you don't have to use those -- the uris can be anything you'd like, including more natural names.
the problem is that people can and will disagree. "talk it out" is not a complete solution. the "talk it out" solution is things like central registries managed by the IANA which most treat as consensus.
-
The problem is rarely in parsing as2 context, but dealing with how different implementations decide to create projections from the data.
Take a simple poll. The 3 diffferent servers I saw were generating the text, the choices, and the replies collection in completely different ways. Without JSON-LD, each separate system would be fighting to figure out how to present the data.
@raphael @silverpill @hongminhee @mariusor most often the trouble i see is with ignoring the fact that everyone is using the same terms with different meanings, and pretending that we all agree when we actually do not.
the second most common issue i see is with the complete lack of any guarantees beyond "this thing is probably an activity" (which even that small bit is often discarded!)
json-ld is so far down the list of pain points, and the pain comes from ignoring it or misusing it.
-
@raphael @silverpill @hongminhee @mariusor most often the trouble i see is with ignoring the fact that everyone is using the same terms with different meanings, and pretending that we all agree when we actually do not.
the second most common issue i see is with the complete lack of any guarantees beyond "this thing is probably an activity" (which even that small bit is often discarded!)
json-ld is so far down the list of pain points, and the pain comes from ignoring it or misusing it.
@raphael @silverpill @hongminhee @mariusor example: if you took lemmy's use of as:Group you might assume that every as:Group is a lemmy-style "community" and that it always produces 1b12-style Announce activities, and that "Announce" means how they use it and not its actual definition.
now if lemmy had used their own vocabulary, it might be easier to understand that "this is a lemmy-style community".
the activity processing model shouldn't care what lemmy properties are used.
