Clause 6.4.3.3 describes the ability to obtain the potential for MULTIPLE modified values on a node for a specific timestamp (not a time-range, but a specific timestamp).

Clause 6.7.2.2 describes that inserting data is not possible when a value exists at the specific timestamp.

Clause 6.7.2.3 describes that replacing a value at a specific timestamp is allowed, but makes no mention of whether or not the server must/should archive the previous value.

Clause 6.7.2.4 describes that using update/replace is a combination of 6.7.2.2 and 6.7.2.3.

There seems to be a conflict here? On the one-hand we're saying that you may have multiple modified values at a specific timestamp, and on the other hand we're saying that we wont allow adding values when the historian contains values for that timestamp, or, you must replace the existing value (at a specific timestamp) with a new value. If the calls don't permit adding multiple values at a specific timestamp, and there's no requirement to archive prior values when being replaced, then how can there be a potential for having multiple modified values at a specific timestamp?

Additional Question: Does (or SHOULD) the spec mention the requirement (or not) for Servers to maintain a record of value that are replaced? (I didn't see it) If such a reference does exists then perfect, where is it and could we reference it in the clauses mentioned above?

We have a number of test-cases "on hold" right now awaiting clarification of the intended behavior with regards multiple modified values at a specific timestamp.

I hope this makes sense. Thanks.