cindi wrote: > From the original case for the fault manager, the command line options > for fmdadm(1M) were classified as evolving and the output, unstable. > It seems straightforward to map evolving to committed for the command > line options. What is not so clear is the commitment level of the > output. The new interface taxonomy suggests that such output is not an > interface at all and should be classified as 'Not-an-interface': > > "In the course of reviewing or documenting interfaces, the situation > often occurs that an attribute will be present which may be inferred > to be an interface, but actually is not. A couple of common examples > of this are output from CLIs intended only for human consumption and > the exact layout of a GUI." > > The project team is happy to adjust the classification level to what > is appropriate and customary. Please advise. > > Cindi Not-an-interface is a bit strange. Its design point was to make clear things which people might mistake as being interfaces. Perhaps this taxonomy was a mistake, but it really doesn't matter.
Make the output either Not-an-interface or Volatile. Then again, if the original case declared the output as Unstable, the new name would be Uncommitted (People felt that Unstable was pejorative.) That has the only allows change in a Minor release, like S11, as if that's going to ever happen. :-) Maybe there is a little clean-up with the old classifications. Probably not a big deal. - jek3
