Imagine having an XML spec with countless enums having minOccurs=0. with a corresponding enumSpecified boolean flag. A class is generated from the spec, which can then be easily persisted with the framework... Very handy indeed.
However - the framework refused to generate a database schema using CodeFirst - the scaffold migration generator threw error 2019 for each enumSpecified flag.
To work around this issue, I had to spend several mind-numbing hours marking the flags [NotMapped], and then writing boilerplate code by hand to add over 50 custom properties that are essentially a useless facade, in order to coerce into the database all of the enumSpecified flags through an int. Not fun, and certainly could have made better use of my time.
I understand the reason for this limitation, and have no defense for IBM in not adding a boolean data type, but at the same time, this limitation is not practical nor reasonable to emphatically enforce with errors. It would be fine to default to "enforcement" mode, but please, give us developers a break and let us turn this checking off...
Hopefully there can be a change made. It is my opinion that the framework needs to support reading a boolean value from a database integer type... Hibernate has it, and Entity Framework needs it also.