[OASIS Issue Tracker] Commented: (ODATA-288) Change default decimal Scale to variable

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

    [ http://tools.oasis-open.org/issues/browse/ODATA-288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=32723#action_32723 ] 

Evan Ireland commented on ODATA-288:
------------------------------------

Some additional considerations...

V3 assumes Scale=0 for unqualified Decimal. Backwards compatibility would seem to be an issue. However is it "unexpected" to be forced to use a facet to allow a fraction.

If Decimal with Scale=0 is considered such an important use case, then we might consider:

(1) Introducing an Integer type (like XML schema "integer"), with optional Precision facet but no Scale facet, or

(2) Allow Precision to be optionally specified for SByte, Int16, Int32, and Int64.

I think the two most important considerations are:

(1) Decimal with variable Precision and Scale shouldn't require the use of a facet.

(2) Integer with variable precision shouldn't require the use of a facet, but we might decide that Int64 has sufficient range for the use cases that an Integer type would be intended for, so in that case we might just recommend using Int64.

My personal preferences would be:

(a) Decimal defaults to variable Precision and Scale.

(b) Add an Integer type which defaults to variable Precision.



> Change default decimal Scale to variable
> ----------------------------------------
>
>                 Key: ODATA-288
>                 URL: http://tools.oasis-open.org/issues/browse/ODATA-288
>             Project: OASIS Open Data Protocol (OData) TC
>          Issue Type: Improvement
>          Components: OData CSDL
>    Affects Versions: V4.0_WD01
>            Reporter: Evan Ireland
>
> In order to avoid data modelers having to make excessive use of facets, the Decimal primitive type should allow variable Scale by default, rather than defaulting Scale to zero, which effectively makes unqualified edm.Decimal equivalent to xsd:integer.
> We could thereby also eliminate the VariableScale facet.
> (A parallel consideration might be that if Scale=0 is common, then we should add an edm.Integer type).

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://tools.oasis-open.org/issues/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira