Update payloads for relationships

Clarified the rules for the different relationship types: single child, list of children, reference, association
Allowed two modes for lists: full list and delta, with sdata:deleteMissing to distinguish between the two (full mode was only a compat mode in 1.0, it is now back in the standard)
Reserved sdata:index for future use in the XSD (no mention in the wiki)

Clarification

Schema tagging

Introduced sme:tags attribute to allow contracts to add tags to schema elements

New Feature

Schema version

Added a standard format for schema versions: major.minor.revision, with associated compatibility rules.

New Feature

Sync tick management

Clarified sync tick management section

Clarification

2.2 Tick Management

Sync requires complete payloads

Clarified the fact that sync feeds must transport complete payloads

Clarification

4.4 Synchronization Feed

Standard conflict priority values for sync

Added a table with standard conflict priority values that endpoints should use.

New Convention

4.2 Synchronization Digest

sme:syncOrder attribute

New schema attribute to indicate order in which resource kinds should be synched

New Feature

4.5 Synchronization Metadata

Polymorphic relations

Added schema rules for polymorphic relations

New Feature

Protocol filter variables

Added protocol filter variables like $uuid and $updated to query language and sme:protocolFilters attribute

New Feature

Query parameters for named queries

Named queries can take query paramters (where, orderBy, include, select)

New Feature

Case sensitivity rules for URLs

Clarified the rules for case handling in URLs

Clarification

SData Registry

The registry contract is used to register and discover application endpoints

New Feature

select query parameter

New parameter to control the properties returned in SData payloads.

New Feature

returnDelta query parameter

New parameter to optimize POST and PUT operations by asking the service to return the delta instead of the full payload.

New Feature

URL extensibility syntax

Convention that applications should use to extend the SData URL syntax

New Feature

Template for service operations

$template can also be used to retrieve default values for input parameters of service operations.

New Feature

Named Queries

Variant of service operations, for queries that take complex parameters and return a feed.

New Feature

sme:mandatory renamed

sme:mandatory has been renamed sme:isMandatory for consistency.

<sdata:schema> element for inlined schemas

Inlined schemas have been moved to an <sdata:schema> element as the previous solution did not comply with the Atom rules (<link> elements cannot contain elements)

Schema import and include

Schemas may import or include other schemas.

New Feature

Reliable posting supported on regular POST operations

Previous version did not allow the trackingID parameter to be used directly on POST operations that create resources. It imposed an additional service operation.
This limitation has been lifted

Restriction lifted.

Internationalized URLs

Precise rules have been given for non-ASCII characters in URLs

Clarification

New mechanism to express relationships between resources

3 new attributes have been introduced to encode links between resources directly in the payload: sdata:url, sdata:key and sdata:uuid.
This mechanism supersedes the <link> mechanism from the previous version.
sdata:lookup and sdata:descriptor have also been added to support generic interactive clients.
sme:hasUuid has been introduce to advertise if resource payloads carry an sdata:uuid attribute or not.

New protocol to update child resources

The protcol to update child resources has been modified. The consumer should not send the complete list of child resources any more. Instead, it should flag deleted resources with the sdata:isDeleted flag.
A flag (sdata:deleteMissing) has been introduced to allow providers to support the old protocol.

sme:maxLength
sme:averageLength
sme:totalDigits
sme:fractionDigits

New attributes to advertise type constraints in the schema.

New Feature

Rules for parent/child relationships

More precise rules have been given to differentiate parent/child relationships from reference and association relationships

Clarification

Link & Sync

Complete protocol to link and sync resources

New Feature

Link & Sync

Atom conformance

More precise rules to make SData feeds fully conformant with Atom rules

Clarification

Schema conventions

Added the --enum suffix convention for enum types

New Feature

MAY/SHOULD/MUST language
Compliance paragraphs

Used MAY, SHOULD and MUST language, as specified in RFC 2119 to make the specification more precise.
Added compliance paragraphs to clarify what is mandatory and what is optional for SData compliance.

Clarification

All wiki pages

Delete concurrency handling

Relaxed the concurrency handling rules for delete operations.
Provider should not return an error if the ETag value is absent.

Relaxed existing rule

SSO

Added Single Sign On section stating that SData will support the Sage Single Sign-On protocol. Technical details will be provided later.

Placeholder for upcoming feature