Re: [ISSUE-103] Hard vs silent fail on missing delete triples

Sent from my portable device. 

> On 11 Dec 2014, at 23:24, Alexandre Bertails <alexandre@bertails.org> wrote:
> 
> Hi Tim,
> 
> The group opened ISSUE-103 [1] after the feedback you provided. We
> discussed it during the past two meetings [2].
> 
> Here are the solutions we considered:
> 
> a) We add a new Assert operation to assert that a triple exists, as
> defined by Pierre-Antoine in [3]
> 
> b) We add a new mode strict/lax operations (syntax to be defined) to
> change the semantics of Add and Delete on the fly. (question: what
> would be the default?)
> 
> c) We define 4 operations (I made up the new names, they are not set in stone):
>  * Add to add a triple, doesn't fail if the triple was already there
>  * Delete to delete a triple, which fails if the triple didn't exist
>  * AddNew to add a triple, succeeds only if there was not such triple
>  * DeleteAny to delete a triple, never fails
> 
> d) We recommend the use of etags, and Add/Delete never fail.
> 
> There was no strong consensus (maybe a _little_ preference for b or
> c?), so the group decided to let the editors to come back with a
> proposal, and to check with you if you had a strong preference.
> 
> Do you have any preference?
> 


Yes c by a long way 
Modes bad. Make it difficult merge documents in a language . Difficult to find errors 

Different commands with clear individual semantics good.  

Scuse brevity
Timbl

> Alexandre
> 
> [1] http://www.w3.org/2012/ldp/track/issues/103
> [2] http://www.w3.org/2013/meeting/ldp/2014-12-08#line0136
> [3] http://lists.w3.org/Archives/Public/public-ldp-wg/2014Nov/0096.html
> 

Received on Friday, 12 December 2014 10:59:11 UTC