@julian

That should be easy enough to get working as-is to begin with, and we can then progressively enhance with schemas?

@mayel @jonny @smallcircles

@smallcircles @julian @bonfire @mayel @jonny

for illustrative examples, see https://linkedresearch.org/calls#contributions-and-review and https://csarven.ca/linked-data-notifications

go ahead, inspect the html, try to fetch it as json-ld, the works

@trwnh
@smallcircles @julian @bonfire @mayel
Ya ya. Adding LD to documents is easy. Making use of it at scale is hard.

@jonny @smallcircles @julian @bonfire @mayel i am finding that the challenge isn't so much "making use of it as scale" as it is specifically "being understood at scale". it's a problem that we have even with natural language. you have the capability to make use of information that you understand, but it's the understanding that's the hard part. and also discovering that information in the first place, and being able to find it when you need it. compared to that, use cases are straightforward.

@trwnh
@smallcircles @julian @bonfire @mayel
I think basically we are saying the same thing because understanding is a prerequisite to use for me so yes, agreed.

@trwnh
As a side note, I think RDFa is a great idea, I tried to use it in a document and make a markdown preprocessor to be able to use it with semantic mediawiki-style wikilinks [[predicate::object|display text]] where the document is the implicit subject, or [[subject::predicate::object]] and have those render out as a elements with RDFa. But then I learned that rdflib had dropped support and there weren't any actively developed parsers, so I gave up because I wasn't gonna write one and what's the point in writing something that can't be parsed easily.

@jonny depending on how things go, i might end up writing one for myself if i decide i want to use html as a canonical content format which gets converted into other serializations. i'm not particularly enamored with markdown as a canonical content format (although i recognize why many people like it) so i don't know how much energy to devote to markdown extensions if i won't end up using them

but i still have research to do re: content modeling and management...

@jonny one thing that seems potentially promising with markdown is to try to put the data in frontmatter instead of content? then when you parse frontmatter as toml/yaml/json you can basically take the json and make it json-ld by adding a context depending on the "type" of content it represents. imagine something like hugo and mapping the default frontmatter attributes to some vocabulary... then establishing equivalences as needed

@trwnh
Hey if you can write HTML, then dont bother with markup formats. I write code for scientists and if I were to ask a scientist to write their document in HTML they would just be like "no" and "never" lol

@trwnh
There is a Jekyll plugin that does this, but one of the things I wanted with RDFa is to be able to annotate the text within the document as well as the whole document. Its a shame it seems to be quietly being deprecated

@jonny @trwnh If you’re interested in a critical history of RDFa: https://aeshin.org/static/pubs/Shaw%20-%202022%20-%20A%20Dispute%20over%20Embedding%20Linked%20Data%20in%20Web%20Pages.c90822e73bae.pdf

@jonny yeah rdfa is mostly stuck in 2013, it never really got fully ironed out because most people already preferred turtle or the new hotness of jsonld. it's a shame because it does have some advantages that the other formats lack (namely, you can markup text *and* make some of that markup extra-semantic, write once instead of having to maintain two separate things and risk them getting out of sync

@rieyin
@trwnh
I certainly am, and have done my own archaeology here, so very interested in finding an actual history

@jonny @rieyin currently reading -- thanks for the link!

@jonny @rieyin ok whew that was a meaty read... i have a lot of notes i took that aren't suited for the mastodon format but i can say that while this is a very good paper and it helped put into words a lot of things that have been brewing in my mind for a long time, it doesn't address what happened *after* the rdfa debate at whatwg and the introduction of microdata, schema.org, open graph protocol. there's probably more at play if you consider this later time period (2010-2013??)

@jonny @rieyin this later time period seems to support my earlier assumption that one of the bigger reasons rdfa stagnated a bit is that it was kind of usurped by jsonld (started development in 2010) and turtle (2011 working draft)

@trwnh
@rieyin
I haven't had time to read yet, but you gotta cover the semtech 2011 meeting where microdata and RDF maintainers were like "we might disagree about formats, but fuck letting google decide"