valueflows

valueflows docs
git clone https://s.sonu.ch/~srfsh/valueflows.git
Log | Files | Refs | README

commit 81f4c4e4f6edf148de54f5073dd0e19af46a22b1
parent ddbbf38e4a4491732dbb13e0d2a82d55495f729e
Author: Bob Haugen <bob.haugen@gmail.com>
Date:   Tue, 20 Oct 2015 16:26:50 -0500

Create 2015-10-20-minutes.txt
Diffstat:
Aminutes/2015-10-20-minutes.txt | 262+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 262 insertions(+), 0 deletions(-)

diff --git a/minutes/2015-10-20-minutes.txt b/minutes/2015-10-20-minutes.txt @@ -0,0 +1,262 @@ +at +Lynn Foster joined group chat. +To list all available commands enter "/?". + +Michael Williams 3:00 PM +https://github.com/valueflows/valueflows/wiki/Meetings + +me 3:00 PM +round trip + +What do we need to do to complete a round trip example that works? Who should do what? + +@id vs. url + +https://github.com/valueflows/valueflows/issues/45 + +owl:inverseOf + +https://github.com/valueflows/valueflows/issues/52 + +types of relationships: existing vocab vs user defined types of relationships + +https://github.com/valueflows/agent/issues/38 + +how specific is our relationship? (rdf:Property, vf:relationship, vf:agentRelationship) + +https://github.com/valueflows/valueflows/issues/63 + +subclasses of Agent, which to support and which to use existing vocab for + +https://github.com/valueflows/agent/issues/51 + +formatting relationship jsonld for discovery + +https://github.com/valueflows/valueflows/issues/59 + +separation of concerns + +documenting current structure of repositories and which one stays responsible for which concerns +Connor Turland joined group chat. + +me 3:04 PM +round trip + +What do we need to do to complete a round trip example that works? Who should do what? + +@id vs. url + +https://github.com/valueflows/valueflows/issues/45 + +owl:inverseOf + +https://github.com/valueflows/valueflows/issues/52 + +types of relationships: existing vocab vs user defined types of relationships + +https://github.com/valueflows/agent/issues/38 + +how specific is our relationship? (rdf:Property, vf:relationship, vf:agentRelationship) + +https://github.com/valueflows/valueflows/issues/63 + +subclasses of Agent, which to support and which to use existing vocab for + +https://github.com/valueflows/agent/issues/51 + +formatting relationship jsonld for discovery + +https://github.com/valueflows/valueflows/issues/59 + +separation of concerns + +documenting current structure of repositories and which one stays responsible for which concerns +https://github.com/valueflows/valueflows/wiki/Meetings +Who's here: Mikey, Connor, Pavlik, Lynn, Bob +Agenda: hard ones first: +Round trip +Types of relationships +how specific is our relationship +subclasses of Agent +.... +Did we close off ID vs URL? +Lynn thinks yes +... +What do we need to do to complete a round trip example that works? Who should do what? +See the whole technology loop work. +Is holodex far off? +Mikey says, a lot of work to do it the way we want. +Holodex is the longer range one, let's start with something simpler +Put out from one app, read in to another. +Then follow our nose. +Get request. +Later content negotiation. +Simple get first. +Do some simple HTML, +simple lists, +avatars. +CORS headers +3-4-5 servers on different subdomains. +integrate data from multiple sources into a simple interface. +Then add more complexity. +Elf will play with Polymer, make something basic, simple graph. +Lynn will get hers working, too. +Holodex will come in later. +Connor made simple profile form. +Then display it. +Small app to make relationships. +Valuable, not a lot of work. +Publish code, show how to do it. +Data, schema, data, apps, to communicate to different audiences. +Connor and Ishan discussing links between platforms, Connor working on the links, want graph platform to work with this kind of data. +Browse links thru URL to URL. +Regardless of how you keep the data, the navigator knows how to navigate and present it. +Starting point, then follow your nose. +http://en.lodlive.it/ +follows the linls +links +Connor: something like metamaps + +elf Pavlik 3:27 PM +ACTION: elf to write minimal requirements for browser + +me 3:27 PM +Mikey requests more visions, what is the end state we are shooting for +.... +Mikey agrees ID vs URL is figured out. +Data for browser to work with. +json-ld for a particular resource is the missing part + +Lynn Foster 3:29 PM +@id vs url - need the @id to link to a URI that returns LOD (jsonld) + +me 3:30 PM +... +types of relationships: existing vocab vs user defined types of relationships + +https://github.com/valueflows/agent/issues/38 +Places where we have vocab in development but not finished. +In our app, all relationships are user defined, but have a behavior field that is usually a common relationship. +Pavlik says, let's publish the data and get more experience and then revisit. +When we want to integrate and query across multiple sources, we will confront the issues that need to be solved. +Lynn mentions issue of filtering. +Asks do we want to agree on a set of types? +Pavlik: work thru examples first. +We have different perspectives and backgrounds, make examples, discuss, propose alternatives. +Lynn proposes small set to discuss. +Sensorica Affiliate same as Member. +DHen has role-based relationships: Harvesting site, Harvester, Drying site +Harvesting site sounds like a resource. +Lynn says it's modeled as a farm, which is an agent (group agent). +Pavlik urges formalizing examples. +Lynn says if we start getting into resources, we won't get agents done. +So let's work with things we can agree are agents. +http://dhen.webfactional.com/accounting/agent-jsonld/ + +Michael Williams 3:41 PM +https://github.com/valueflows/valueflows/tree/master/use-cases + +me 3:43 PM +Agents should have agency, are they all really agents +Lynn suggests discuss Affiliate +is it a valid relationship? +They might change it again. +In our software, we know it is like a member. +Is there a label, preferred label? +Pavlik returns to the idea of waiting until we need to integrate data. +Lynn asks does it make sense to make define a type of relationship called Member +or a small set of relationships that have known behjavior and sub-property off those? +Or deviate from those if necessary? +Mikey agrees with putting everything out for now and then look for the overlaps later. +We are agreed, integrate and refine later. + +Lynn Foster 3:49 PM +relationship is instance of rdf:Property + +me 3:49 PM +premature to amke superclasses + +Lynn Foster 3:49 PM +super-properties + +me 3:50 PM +we all agree that: "relationship is instance of rdf:Property" +How specific are relationships? +I.e. are Agent<->Agent relationships necessarily different from Agent<->Resource relationships? +Pavlik suggests focusing on Agent<->Agent +Once we move on the Agent<->Resource we can look at those and know more precisely what relationships we want. +Lynn is concerned about backtracking after holodex, tibor, etc etc are already using it +Pavlik says we can make it clear that they are experimentatl +don't promise stability +wait for experience with integrating data across sources +....next topic.... +subclasses of Agent + +Lynn Foster 3:59 PM +we agreed on keeping general for now, get more experience with how the LOD technology works, then be more specific as we need to +so rdf:Property for now + +me 3:59 PM +which ones do we want to support +Mikey's opinion, don't worry about it yet. +Let people define from their own data and see what the commonality is. +We agree on Agent as the top level. +vf:Agent for awhile instead of FOAF:Agent? +yes, convert later +Can we agree on Person and Group as subclasses? +Pavlik wants to stick with those two for now. +Lynn says we in NRP land will need more subclasses +Pavlik says subclass in data as required +says we can use inference to distinguish e.g. Network +We all agree on user-defined types +Pavlik reiterates, publish formalized examples, integrate, then discuss how to create the common vocab. +As long as we don't modify the common context, we have complete freedom to publish examples. +We all agree "As long as we don't modify the common context, we have complete freedom to publish examples." +... +New topic: separation of concerns + +documenting current structure of repositories and which one stays responsible for which concerns +Connor explains why he is here: +listening to different conversations in different groups +helping to create bridges +elf's statement will help to create links +feeling from different platforms +wanting to create mappings between groups +nodes in graphs +taxonomies +events, intentions will also show up in hylo +implementing them in different platforms +Connor will be traveling around trying to make bridges between platforms +Mikey likes it +Lynn thinks its fabulous +Pavlik says thanks, circles are closing +Connor Turland left group chat. + +me 4:19 PM +Wrap up time +Mikey says, keep the rhythm going +Checkouts: +Mikey is appreciating the energy +some gravity puling everyone in +moving to a tighter orbit +Lynn totally agrees about the energy and the diverse opinions, +glad we got some agreements. +feel like I can go forward +Pavlik also glad we met again +likes puce bubbles +likes agreement on workflow +very optimistic +excited to work on the json-ld of this meeting +An adventure +Lynn appreciates elf keeping us on track +Mkey things there is no way we could go wrong + +elf Pavlik 4:23 PM +Bob - agrees with everybody + +Lynn Foster 4:23 PM +bob: agrees with everybody +happy about what connor is doing + +me 4:24 PM +teeshirts!