Re: [ubl-dev] OrderChange.SequenceNumberID

[David Goodenough <david.goodenough@broadwellmanor.co.uk>]

A rule it is then.  Just wanted to be sure that I was not missing something.

 

David

 

On Monday, 1 June 2020 14:54:59 BST you wrote:

 

A PEPPOL kind of profile would be able to restrict an ID to an integer to make it work within a user community (always the intention of UBL) 

 

On Mon, 1 Jun 2020 at 14:53, Stephen D Green <stephengreenubl@gmail.com> wrote:

 

Mmm. No idea how that sequence number ended up as an ID. Design by committee had quite a few pain points. As much political as technical 

 

On Mon, 1 Jun 2020 at 14:22, David Goodenough <david.goodenough@broadwellmanor.co.uk> wrote:

 

Yes you are right that Date only covers one change per date, but there are no ordering semantics for an ID.  Never did understand why Date and Time were used separately rather than DateTime - but that is another question.

 

Were it a numeric field that would be fine, but this is text.  So if for instance a buyer were to use "first", "second", "third" (fine up to here) and "fourth" as the IDs - it is after all supposed to be buyer assigned with no semantic constraint - then all attempts to find the most recent by any kind of algorithm is going to be dead in the water. 

 

So the process has to become manual, which rather defeats the point of UBL.

 

David

 

On Monday, 1 June 2020 13:35:54 BST you wrote:

 

Correction: invitation to treat plus offer (not offer to treat)

 

On Mon, 1 Jun 2020 at 13:35, Stephen D Green <stephengreenubl@gmail.com> wrote:

 

You have to remember that these are legal documents too. So issue date semantics are mainly for legal aspect of order-related documents since an order has implications for contract and âoffer to treatâ. Sequence semantics are more about process aspects of change documents. 

 

On Mon, 1 Jun 2020 at 12:36, David Goodenough <david.goodenough@broadwellmanor.co.uk> wrote:

 

When would it be possible/sensible for the ordering of OrderChange documents by IssueDate and SequenceNumberID to be different? 

 

Both are mandatory fields, and both seem to convey the same semantics.  Surely one of them (the SequenceNumberID seems the obvious candidate as IssueDate is mandatory just about everywhere and is deterministically sortable provided it is always stored in UTC whereas an ID is not) is redundant and worse possibly contradictory?

 

David

--

 

----

Stephen D Green

--

 

----

Stephen D Green

 

--

 

----

Stephen D Green

--

 

----

Stephen D Green