Ted summarizes emails and calls on latest convergence attempts
PatrickL: are we in good shape for next week's meetings?
Ted: small f2f crowd but not critical mass for our main topic. ws & meetup going well
PatrickL: I think we should have sbranch further in the mail thread
https://lists.w3.org/Archives/Public/public-automotive/2019Aug/0029.html
Ted prompts Ulf to introduce rbranch
Ulf: I originally thought of this
for being able to bring ViWi data to VSS not the other way
around
... it is possible to help migrate data from a ViWi model such
as media to VSS
... without rbranch it is hard to identify what a child looks
like or if there is even a child
... a resource in ViWi doesn't need a child node but it is
possible and you can find out the metadata to see if there may
be children, rbranch introduces that
... in theory it should work and did some simple test
translations but haven't tried a full tree
Ted: anyone besides Laurent have thoughts on sbranch?
PatrickL: besides the email unsure about the details of the proposal
Ulf: I am not sure how sbranch would help with migration in the other direction
Ted: @@3level
PatrickL: every level within ViWi
expresses a certain type, service, resources and bundled data
set (nodes)
... they are not really comparible given the fixed meaning and
structure
Daniel: does every microservice have its own endpoint in discovery or everything from one connection?
<PatrickLue> /<service>/
<PatrickLue> /<service>/<ressource>/
PatrickL: the lowest level root would be service, next resource and third element[s]. it isn't about a tree really
<PatrickLue> /<s>/<r>/<element>/
PatrickL: there can be multiple services on the same level. elements have several properties
Daniel: for me, microservices are a thing of infrastructure. in VSS you could have subtrees served by different service implementations
PatrickL: there are multiple
things mixed in the idea. for microservice we expect well
defined apis and only allowing the service to interact through
it
... there could be multiple microservices, for instance say
anything below .../seat is handled by another service
instance
... the service resource element could be described further and
will do so by email
Gunnar: consider what the HTTP
request looks like, we should focus on that and figure out the
piece behind it
... a few goals: VSS tree has been expressed as a good
structure and there is a desire to have REST requests reflect
that structure
... the three depth limit could just be another way to serve
those resources
... we can leave aside eg uuid for now
[tangent on uuids and alternate uri paths]
Daniel: there were merits for
requesting unique ids for resources
... point of Gen2 is to be developer friendly
... main focus should be the path
Ted: @@alt_data_structures
PatrickL: so far people have
understood to make their service available there is a structure
they need to follow
... there have been times where they have had flatter views and
challenge to get to depth
Ulf: about ability to address unknown data models, the chance of being able to do that is greater with VSS model vs ViWi since former is more flexible
Daniel: you have graph interface there, how do you know if you are querying the third level?
PatrickL: for tracks you would request everything on the third level
Daniel: how do get down to track level in media example, that seems deeper?
PatrickL: yes that is the forth level I alluded to
Daniel: media player/media collection/album and then how deep can the property level go?
PatrickL: it is possible to have
tree below that but can only filter on properties
... URL always starts same
... if not given any filter arguments gives you full json
element
Daniel: service discovery for first url you don't have since it stays the same?
PatrickL: every service is
responsible for adding itself to registry, could be solved with
something like mdns
... everything is known and discoverable. all services bring
themselves to the service registry
Daniel: that could be different services from VSS branches registering themselves
PatrickL: correct
Daniel: you could just serve VSS there
Benjamin: why can't you have
resource and subresource?
... I don't understand this depth limitation as it doesn't
align with REST principles
Daniel: REST doesn't have such
limitations
... service can be defined by the host contacted and could use
different microservices transparently beneath it
... an option for Gen2 would be to simply remove the depth
limitation
PatrickL: we want to avoid adding
complexity
... easier for someone interested in a certain topic to filter
at a higher level than drilling deeper
<PatrickLue> Media Service Description https://www.w3.org/Submission/2016/SUBM-viwi-service-media-20161213/
Daniel: if a query is easier, we need to do a true comparison
Tuesday morning PST during f2f will continue this conversation
This is scribe.perl Revision: 1.154 of Date: 2018/09/25 16:35:56 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Present: John Ted Adnan Glenn Benjamin Ulf Joakim Gunnar Sebastian PatrickL Daniel No ScribeNick specified. Guessing ScribeNick: ted Inferring Scribes: ted WARNING: No "Topic:" lines found. WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Found Date: 03 Sep 2019 People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option. WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report WARNING: IRC log location not specified! (You can ignore this warning if you do not want the generated minutes to contain a link to the original IRC log.)[End of scribe.perl diagnostic output]