Research for cooling and trapping of atoms in the mid 1980s was mainly motivated to enable precise spectroscopic measurement. However, perturbation by the trapping force introduces sig-niﬁcant inhomogeneous broadening, and the long interrogation time did not directly translate into narrow lines. We have devised methods to eliminate the broadening in both optical and ground hyperﬁne transitions of alkali metal atoms by using magic wavelength and magic polarization, re-spectively. I will discuss our experiments on cesium atoms using magic wavelength, and on lithium atoms using magic polarization. In addition, the narrow hyperﬁne transition of lithium atoms, which are trapped in a 1D optical lattice with magic polarization, allows site-speciﬁc addressing when a magnetic ﬁeld gradient is applied. I will talk about our experiment to manipulate two single atoms - or qubits - in adjacent sites of the lattice in a coherent and independent way.