30 April 2014
John Clyne, Alan, Scott
@mention a person to add them as an attendee and they will be notified.
We reviewed the Error Handling documented and agreed to pursue option (3):
Avoid use of MyBase's global error code for detecting errors. This would imply guaranteeing that constructors never fail (easily done by adding a class initialization method when needed, e.g. open() or initialize() ). Hence, method/function return values, not MyBase::GetErrCode(), would be the preferred mechanism for detecting errors.
We also reviewed the WASP API. Alan raised questions about how the API might scale with larger and larger data sets, and how whether it would support unstructured data like that produced by MPAS. John responded that there are a couple of extensions that could be made to improve scalability (e.g. more flexible control of grid resolution). The WASP extensions would not, however, readily support unstructured data.
We also discussed the action item list. Alan is completing item (3), before moving on to 4 and 5. After that time Alan feels we should review the prototype from end-to-end.
Item | Who |
---|---|
Review WASP API proposed by John | Alan, Scott |