The Methods and Tools Special Interest Group will meet from 5 pm till 7 pm Central European Time.
Agenda: Issues addressed at CAPE-OPEN 2016 Annual Meeting where COBIA Phase I deliverables were demonstrated.
U.S. Environmental Protection Agency (represented by Bill BARRETT), AmsterCHEM (represented by Jasper van BATEN), Linde Engineering (represented by Hans-Günter SIMON), Bryan Research & Engineering (represented by Michael HLAVINKA), Michel PONS (contractor for CO-LaN as Chief Technology Officer)
Error Common interfaces: the need for such an Errata & Clarifications document was questioned at the CAPE-OPEN 2016 Annual Meeting. AmsterCHEM makes the point that no PME is known to really act according to the type of error raised by a PMC. The same takes place in reverse: there is nothing the PMC can do when the PME raises an error. The only use of the description supposedly attached to each CAPE-OPEN error raised, is to inform the end-user by displaying somehow the description. The context in which an error is raised is sufficient in most cases: a persistence error will be raised by a persistence interface when called. A calculation error will be raised by a Unit Operation within its calculation and the PME will already know which method has been called.Still enforcing the need for a clear description message together with any error raising is a valid request for an errata.
Another basic requirement in the communication over CAPE-OPEN between a PMC and a PME is that only CAPE-OPEN errors should be thrown by CAPE-OPEN methods. COM errors should not be thrown instead. So rather than an Errata & Clarifications, best practices could be formulated for COM/CORBA implementations: source should provide a description, receiver should check for an error condition.
AmsterCHEM advocates for nested errors to be taken on board in the COBIA version of error handling. It is a type of call stacks unwinding much like in .NET.
COBIA licensing: another issue raised at the CAPE-OPEN 2016 Annual Meeting. COBIA needs a license that does not impose any requirements like copyleft on users of the software. MIT and BSD licenses are likely choices. MIT license requires that people using the software provide copyright and permissions notice, but it can be used and included in other software. End users should be granted information about their rights with regard to the use of the COBIA code. Point raised: does CO-LaN need to release COBIA runtime with code signing?
COBIA Phase II: scope of COBIA Phase II needs established. For the interfaces to be included in COBIA, what is not supported by a SIG does not go into COBIA. Numerical interfaces are “on hold.”. AmsterCHEM points out that parameters and persistence need to be defined immediately in COBIA. Regarding Parameters we want to get rid of variant as part of COBIA.
Contact Bill BARRETT (representative of U.S. Environmental Protection Agency and leader of the Methods & Tools Special Interest Group) should you wish to attend (CO-LaN members only)