- From: Markus Lanthaler <markus.lanthaler@gmx.net>
- Date: Mon, 2 Apr 2018 21:12:45 +0200
- To: "'Hydra'" <public-hydra@w3.org>
Thanks Pavlik for scribing. The minutes from this week's telecon are
now available at
http://www.hydra-cg.com/minutes/2018-04-02/
The full text of the discussion is below, including a link to the audio
transcript.
-------------------------------------------------------------------
Hydra W3C Community Group Telecon Minutes for 2018-04-02
Agenda:
https://www.w3.org/community/hydra/wiki/Conference_Calls#Agenda
Topics:
1. Actions with explicit target (#154)
2. Usage of Hydra vocabulary in Heracles.ts
3. Usage of actions in ApiDocumentation
4. Subject URL in POSTs
Action Items:
1. Markus to write down a proposal on how a expects "request
shape" could look like
2. Karol, to extend use cases to include ApiDocumentation
examples
3. Pavlik, to create issue regarding subject in POSTs etc.
Chair:
Markus Lanthaler
Scribe:
elf Pavlik
Present:
Markus Lanthaler, elf Pavlik, Karol Szczepański
Audio:
http://www.hydra-cg.org/minutes/2018-04-02/audio.mp3
elf Pavlik is scribing.
Markus Lanthaler: we don't have an agenda, what would you like
to discuss?
Karol Szczepański: I don't know the state of the 154 (actions)
PR
Topic: Actions with explicit target (#154)
Markus Lanthaler:
https://github.com/HydraCG/Specifications/pull/154
elf Pavlik: I made a very small commit based on what we
discussed previously [scribe assist by Markus Lanthaler]
... as I pointed out before, use cases 5 and 5.1 has more
differences than necessarz
... this PR makes both use a UUID
... once the server creates the UUID, once the client
... both of them use the same pseudo-code now
... (like the one Markus posted on GitHub)
Karol Szczepański: I implemented that feature for Herakles.ts,
it has component for filing up iri template
Karol Szczepański: I see issue with specification not explaining
how to get values for variables in a template based on properties
in the mapping
my approach takes 2 resources, first goes into payload and second
gets used for values in IRI template variables
Markus Lanthaler: at some point we need to define how IRI
templates relate to request
... have you thought more about that give your implementation
xp in heracls?
Karol Szczepański: in case where you have iri template and the
payload, the template can have variables which values will not
come from the payload
elf Pavlik: I think we miss some use cases [scribe assist by
Markus Lanthaler]
Markus Lanthaler: ... so far we have only two: one using UUID,
one using slugs
... maybe we should create a few more use cases
... the UUID, e.g., appears only in the URL but not in the
payload
... Linked Data Fragments use subject, predicate and object in
their URL templates
... it uses them as kind of reified statements.. not properties
of the resources
Markus Lanthaler: yeah, that's a good point [scribe assist by
Markus Lanthaler]
Karol Szczepański: we may miss something in the spec explaining
how to 'craft' the whole request
... we have details on how to 'craft' body
... but not enough clarity on IRI templates, or request headers
... HTTP request can have headers but spec doesn't account for
that
Markus Lanthaler: i thought that we should make 'expect' some
data shape which would instruct what goes in the payload etc.
Karol Szczepański: I recall Tomasz suggesting something in this
direction
Markus Lanthaler: we could for example use SHACL
ACTION: Markus to write down a proposal on how a expects "request shape"
could look like
Sorry, but no Tracker is associated with this channel.
elf Pavlik: maybe we could build around UC5 and UC5.1 in two
versions 1 - could use UUID like in my last commit to PR 154 and
the other one could use *slug* for both 5 & 5.1 as attempted in
previous version of UC5.1
elf Pavlik: I would like to see more use cases that require
actions with an explicit target [scribe assist by Markus
Lanthaler]
... currently it seems most of them could be handled with
memberTemplate
Markus Lanthaler: something related to IRI templates in issue
about 'filtering enumerations'
https://github.com/HydraCG/Specifications/issues/161
Karol Szczepański: I have question related to PR on Herakles.ts
Topic: Usage of Hydra vocabulary in Heracles.ts
... you suggested something about using official Hydra JSON-LD
context
Markus Lanthaler: namespace IRI responds with JSON-LD context
Markus Lanthaler: http://www.w3.org/ns/hydra/context.jsonld
Karol Szczepański: Markus, we can't hear you
Karol Szczepański: I will embed this context statically for now
and make issue to load it dynamically
Markus Lanthaler: In general it takes a while go get through
those PRs, do you have suggestions on how to improve it
Karol Szczepański: I'll try to continue with having multiple PRs
open at the same time and let them land as it progresses
https://github.com/HydraCG/Specifications/issues/160
Topic: Usage of actions in ApiDocumentation
Markus Lanthaler: for ApiDocumentation one would need to create
dedicated property for that
Karol Szczepański: we could try to make ApiDocumentation for
existing use cases
elf Pavlik: we should also decide whether ApiDocumentation is
supposed to support everything that can be done inline [scribe
assist by Markus Lanthaler]
Markus Lanthaler: I would find it nice if they both have the
same scope but we may not find it possible
Karol Szczepański: let's try with our existing use cases first
and see how it looks
ACTION: Karol, to extend use cases to include ApiDocumentation examples
Sorry, but no Tracker is associated with this channel.
Topic: Subject URL in POSTs
elf Pavlik: when POSTing something, the client doesn't provide a
URL (@id in JSON-LD) [scribe assist by Markus Lanthaler]
... is the server supposed to return the subject URL in the
response
Karol Szczepański: POST should provide IRI in Location header
... quote from RFC
Karol Szczepański: we discussed possibility of including `@id`
or not based if you create or update
Markus Lanthaler: if you send POST with multiple addresses
server could split it into multiple resources
ACTION: Pavlik, to create issue regarding subject in POSTs etc.
Received on Monday, 2 April 2018 19:13:17 UTC