Invalid HTML on "schema.org" schema pages.

[John-Nagle]

"http://schema.org/Thing" and its descendants are no longer valid HTML. This seems to have been broken in Q1-Q2 2014, based on archived versions in the Internet Archive. In particular, the list of "More Specific Types" uses "li" items which are no longer within a "ul". (There's also an extra "/table" and a bogus "/span" tag in there.) This makes it harder to machine-process that data.

I assume those pages are automatically generated from "https://github.com/schemaorg/schemaorg/blob/master/data/schema.rdfa", which was new in 2014, so this is probably just a bad template.

W3C validation report: http://validator.w3.org/check?uri=http%3A%2F%2Fschema.org%2FThing&charset=%28detect+automatically%29&doctype=Inline&group=0

[danbri]

Thanks, yes we moved to a new codebase early last year. We ought to fix this, indeed...

[John-Nagle]

Thanks. The RDFa version of the same data, "http://schema.org/docs/schema_org_rdfa.html" (which comes from "https://github.com/schemaorg/schemaorg/blob/master/data/schema.rdfa") could also use some work. It's OK up until line 4536; then it gets weird. There are "href" attributes on "span" tags, starting at "action The movement the muscle generates." It looks like some schema of medical information in a similar, but not quite compatible, format was pasted in there. There's similar cut and paste trouble near "series" and "wikidoc". Some of this won't even parse properly as HTML5 in a browser.

Validator:
http://validator.w3.org/nu/?doc=http%3A%2F%2Fschema.org%2Fdocs%2Fschema_org_rdfa.html

[danbri]

Thanks. Don't spend too much time on the schema.rdfa file's compatibility - in its current form it is an implementation detail. It turned out not to be ideal to use RDFa for this, and we are looking into migration to JSON-LD anyway. But I need to integrate and test a decent parser for that first...

[timbl]

Could you folks please make Turtle an option.

• It is simpler than rdf/a or json/ld
• It is a native graph language not a tree language like json or xml
• It is the one required common language in the linked data platform.
• It can be read by things old libraries
  Tim

[chaals]

• reply@- notifications@  13.03.2015, 22:53, "Tim Berners-Lee" notifications@github.com:Could you folks please make Turtle an option.You mean getting a turtle version of information we have, or reading Turtle from pages (how)? cheers It is simpler than rdf/a or json/ldIt is a native graph language not a tree language like json or xmlIt is the one required common language in the linked data platform.It can be read by things old libraries Tim—Reply to this email directly or view it on GitHub.   --Charles McCathie Nevile - web standards - CTO Office, Yandexchaals@yandex-team.ru - - - Find more at http://yandex.com 

[John-Nagle]

I have no strong preference on format. Please pick something parseable and implement it. Thank you.