OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

coel message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: [OASIS Issue Tracker] (COEL-149) Section 8.2.3 'Exceptions' in BAP needs clarification

    [ https://issues.oasis-open.org/browse/COEL-149?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=67182#comment-67182 ] 

Paul Bruton commented on COEL-149:

Dave added this text "The Data Engine MUST distinguish non-identical Atoms. The Data Engine SHOULD NOT store duplicates of Atoms."

I'd question the meaning of 'distinguish'. Perhaps we need to specify what we mean from the point of view of the interface:

DE must accept duplicate atoms without error (allows a device to transmit more than once)
DE shuold not return duplicate atoms in a query (caller can always process filter them out)
DE must not count duplicate atoms in an aggregate query (caller would get wrong answer to an aggregate if duplicates were counted).

> Section 8.2.3 'Exceptions' in BAP needs clarification
> -----------------------------------------------------
>                 Key: COEL-149
>                 URL: https://issues.oasis-open.org/browse/COEL-149
>             Project: OASIS Classification of Everyday Living (COEL) TC
>          Issue Type: Task
>            Reporter: Paul Bruton
>            Assignee: David Snelling
> See comments in revision 41.
> Let's be sure we make it clear what we mean here: In some of these cases the error response is specified above. The question is whether the DE provides any additional services like quarantining the atoms or post-processing them later to see if the consumer/device ID becomes valid a bit later….
> I don't see how a DE could state that it would return an error in this case: Checking each incoming atom against it's store for duplicates would be quite a burden and what would be the value? I think it is either duplicate or overwrite
> If we mean 'syntactically invalid' – i/e/ not a pseudonymous key, then the DE has to return a 400 in both of these cases. 

This message was sent by Atlassian JIRA

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]