[en] Models of composite weak bosons, the top-condensate model of electroweak interaction and related models we surveyed. Composite weak bosons must be tightly bound with a high compositeness scale in order to generate approximate puge symmetry dynamically. However, naturalness argument suggests that the compositeness scale is low at least in toy models. In the top-condensate model, where a composite Higgs doublet is formed with a very high scale, the prediction of the model is insensitive to details of the model and almost model-independent Actually, the numerical prediction of the t-quark and Higgs boson masses does not test compositeness of the Higgs boson nor condensation of the t-quark field. To illustrate the point, a composite t_R-quark model is discussed which leads to the same numerical prediction as the top-condensate model. However, different constraints an imposed on the structure of the Higgs sector, depending on which particles are composite. The attempt to account the large t-b mass splitting by the high compositeness scale of the top-condensate model is reinterpreted in terms of fine tuning of more than one vacuum expectation value. It is difficult to lower, without a fourth generation, the t-quark mass in the composite particle models in general because the Yukawa coupling of the i-quark to the Higgs boson, _t2/4π = 0.1 for m_t = 200 GeV, is too small for a coupling of a composite particle