Update text on vocabulary mapping and @vocab to allow relative IRIs

[gkellogg]

which are resolved relative to the base IRI. This subsumes the previous usage of allowing just the empty string.

This is for #72.

---

Preview | Diff

[BigBlueHat]

Some weird whitespace happening in example 25 for the Turtle:

<http://example.org/places#BrewEats> a <http://example/document#Restaurant>;
<http://example/document#name>
"Brew
Eats"
.

Don't think that's intentional though. Thoughts?

[gkellogg]

@BigBlueHat It seems to be an artifact of the diff view; it's okay in the source and in the normal preview.

[iherman]

This issue was discussed in a meeting.

• RESOLVED: address security concerns related to relative URIs for @vocab in current PRs before closing #72

View the transcript

Allow relative IRIs for @vocab
Benjamin Young: #72
syntax - #114
api - w3c/json-ld-api#58
Gregg Kellogg: when we decided to satisfy the need for using the base as the @vocab,
… I came up with the idea of using an empty string as @vocab.
… When I implemented it, what I ended up doing was resolve @vocab as a relative URI against the base.
… So why not allow that in the spec…
Ivan Herman: ship it! :-)
Gregg Kellogg: It is uncontroversial, as it does not change past behaviour.
… Now you can simply use @vocab: “#”, to use relative URIs everywhere.
Rob Sanderson: –> #72 (comment)
Rob Sanderson: is this the issue we discussed this during TPAC?
… if you set vocab to ../# and you had example.org/ns then you get example.org/ns../#
Gregg Kellogg: I have to improve the text to indicate when standard URI resolution is used, and when string concatenation is used
Proposed resolution: address security concerns related to relative URIs for @vocab in current PRs before closing #72 (Benjamin Young)
Ivan Herman: +1
Benjamin Young: +1
Ivan Herman: please close issue 72 when you clarify that and merge it.
Gregg Kellogg: +1
Rob Sanderson: +1
Ivan Herman: Let’s reduce the number of open issue
Pierre-Antoine Champin: +1
David I. Lehn: +1
Resolution #4: address security concerns related to relative URIs for @vocab in current PRs before closing #72
Simon Steyskal: +1

[gkellogg]

Latest changes fix #72.