[7:54] Ram (Microsoft): Self-registration link: https://www.oasis-open.org/apps/org/workgroup/odata/event.php?event_id=41444
[7:54] Ram (Microsoft): Conference call details: https://www.oasis-open.org/apps/org/workgroup/odata/download.php/56760/TC%20meeting%20dial-in%20details.htm
[7:58] anonymous morphed into Mark Biamonte (Progress)
[8:00] anonymous morphed into Matt Borges (SAP)
[8:00] Ram (Microsoft): Good morning
[8:00] Ram (Microsoft): Waiting for Ralf to start the conference call
[8:03] Room information was updated by: Stefan
Please register yourselves at https://www.oasis-open.org/apps/org/workgroup/odata/event.php?event_id=41444
[8:03] Stefan: Info: We have Quorum: Voting Members: 8 of 13 (61%) (used for quorum calculation)
[8:04] Hubert Heijkers: Good Morning/Afternoon everybody, Happy New Year!
[8:04] Mike Pizzo: Happy New Year Hubert!
[8:04] Ram (Microsoft): I assume all of you dial-in are parked waiting for the host to initiate the call.
[8:05] Ram (Microsoft): Matt or Gerald: Do you see Ralf online?
[8:05] Matt Borges (SAP): As far as I know, Ralf is on vacation.
[8:06] Mike Pizzo: Can Matt or Gerald admit folks to the meeting?
[8:06] Hubert Heijkers: Ram, a bunch of us are in already, Matt, Gerald and myself are chatting away :D
[8:06] Ram (Microsoft): Can one of you admit the others on the caller queue?
[8:06] Matt Borges (SAP): Try using this link: https://sap.emea.pgiconnect.com/odata-tc?launcher=false
[8:07] Ram (Microsoft): I still see this: "This is a private meeting. Your request to enter has been sent to the host. Please wait for a response."
[8:08] Mike Pizzo: me too.
[8:08] anonymous morphed into Jeremy Crawford
[8:08] Hubert Heijkers: I simply dialed in!
[8:08] Hubert Heijkers: https://www.oasis-open.org/apps/org/workgroup/odata/download.php/56760/TC%20meeting%20dial-in%20details.htm
[8:08] Mike Pizzo: So no screen sharing? I think that's fine for today...
[8:09] Hubert Heijkers: Hold on
[8:09] Hubert Heijkers: Gerald just got in and is accepting everybody!
[8:11] Ram (Microsoft): Seems like everybody is on the call now. Thanks to Gerald.
[8:15] Ramesh Reddy(Redhat): can you share the address
[8:20] Ram (Microsoft): Potential agenda items for F2F: 1) JSON Schema 2) Errata 3 3) OData 4.1 4) OData Extensions for Data Aggregation 5) Extensions for Temporal Data
[8:27] Ram (Microsoft): Section 6. v4.0 Errata 3 issues
[8:28] Ram (Microsoft): ODATA-882: Mismatch between 11.2.6 Requesting Related Entities and 11.2.7 Requesting Entity References [Open]
[8:54] Ram (Microsoft): In errata 3 - be lenient and allow for both to accommodate current implementations. This issue should be fixed in v4.1 (potential breaking change) - we need to determine the exact semantics.
[9:00] Mike Pizzo: Unclear whether behavior should be 204 or 404. We return 204 for Employees('TheBigBoss')/Manager because the question is valid and the answer is that there is no related resource (all agree).
Argument for 404: When asking for Employees('TheBigBoss')/Manager/$ref, you are asking for the (absolute) reference to something that doesn't exist. This is similar to asking for the address of an object (i.e., the address of an object-valued property of an instance) that doesn't exist and should be an error (404).
Argument for 204: It's the same question, asked in a different way. In both cases you're asking for the resource. In one case you're asking for the resource, in another case you're asking for an alternate representation (the $ref) of the resource.
[9:08] Ram (Microsoft): We need to specify a one or the other behavior, so as to set the expectations for the client.
[9:10] Ram (Microsoft): ODATA-784: Need to specify the behaviour of arithmetic operators on Decimal type [Open] [Proposed]
[9:18] Mike Pizzo: So all of the implementations follow Evan's proposed rules for add/sub/mult. None of the implementations appear to follow Evan's rule for div.
[9:18] Susan Malaika (IBM): IEEE 754-2008 and ISO/IEC/IEEE 60559:2011
[9:22] Ram (Microsoft): For v4.1, should we require scale up for div?
[9:27] Mike Pizzo: Most of the Div examples seems to return the value in the max scale of the underlying type.
[9:28] Mike Pizzo: It appears that Evan's proposal for div (of Scale(A)) does not agree with his later comments (Scale(A)+Scale(B)).
[9:51] Mike Pizzo: Added to Issue:
[9:51] Mike Pizzo: All of the implementations follow Evan's proposed rules for add/sub/mult. None of the implementations appear to follow Evan's rule for div.
Most of the Div examples seems to return the value in the max scale of the underlying type.
=>How would a client know this value?
=>If an implementation is aggregating data from multiple underlying sources, how could it return (or even correct for) a single value for max scale? Would it be forced to convert everything to the min max scale?
There seem to be at least 5 proposals for div:
Scale(A)
Scale(A)+Scale(B)
Scale(A)-Scale(B)
>= Scale(A) (doesn't seem to address Evan's second case (A div 3) gt 0.3)
max precision of decimal
Susan will look to see what IEEE 754-2008 (ISO/IEC/IEEE 60559:2011) specifies.
[9:56] Ram (Microsoft): Next meeting: Jan 14
[9:57] Ram (Microsoft): Next week: Focus on issues on Enums, simplified syntax, payload processing, and other issues that may have proposals.
[9:58] Ram (Microsoft): Next week: We will discuss status of the various docs and timeline, and what we plan to accomplish by June 2016