Marker interface seems to be a very elegant solution. I am just
wondering about a potential issue. Let us assume we defined two
interfaces, say
SymmetricMatrix, and BandedMatrix. User A writes a matrix class which
implements both interfaces. Meanwhile, user B implements an algorithm
which requires a symmetric, banded matrix. Presummably, user B will
define a new marker interface, which extends both SymmetricMatrix and
BandedMatrix. So we have on the one hand
class UserAMatrix implements SymmetricMatrix, BandedMatrix;
and
UserBAlgorithm.operate(BandedSymmetricMatrix).
One day, user A hears of the work of user B. The matrix he has
implemented has just the required features (symmetric *and* banded). But
it does not implement BandedSymmetricMatrix, so he cannot apply
UserBAlgorithm.operate to an instance of UserAMatrix. There is always
the possibility of creating a new class which extends UserAMatrix and
implements BandedSymmetricMatrix, but that would obfuscate the hierarchy
tree. Is this a problem? Is that a stupid remark?
Sebastien
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
