Log In
Home
10/13/2021
Offline
Hi everyone. I have a couple of question for OPC UA experts:
1) Do you know or is there a document/resource which describes how specifications are named? E.g. OPC UA specs are named 10000-X with X<100 for certain specs and X>100 certain companion specs. What is the rationale there?
2) I understand OPC UA specification versioning for UA specs named 10000-X with X<100 evolves as follows: ..., 1.03, 1.04, 1.05.01, 1.05.02,.... Can an OPC UA stack of version Y work with the Base Information Models (i.e. Opc.Ua.NodeSet2.xml available from the OPC Foundation) of version Z? I suppose there shouldn't be problem when Y>=Z, but I am unclear on the other way around.
Hopefully you can help me with some pointers/help.
Thanks!
05/30/2017
Offline
1) The numbering for CS is based on classification. See the left panel here:
https://opcfoundation.org/deve.....documents/
2) Specs are generally forward and backward compatible. i.e. a 1.05 can work with all versions <1.05 because changes are additive and all applications are expected to handle extensions to an information model that they had not seen before.
10/13/2021
Offline
Thanks Randy for your prompt reply!
Regarding "all applications are expected to handle extensions to an information model that they had not seen before", can you explain or point me to specification text explaining this concept?
Regards,
Luca
05/30/2017
Offline
If a Server instantiates a ObjectType it can add additional nodes to the instance as it sees fit.
i.e. it is legal create an Object with B/C/Y components from ObjectType that only defines the B/C components.
Every client is expected to be able to deal with extra stuff it does not understand (i.e. ignore extra stuff without errors).
1 Guest(s)
Usage Policy