Page 102 - Geometric Modeling and Algebraic Geometry
P. 102

100    T.-H. Lˆ e and A. Galligo
                                                                ∗
                                                         ∗
                           This condition defines a scroll F(2, 2) in (P ) that we call the dual scroll of F(2, 2).
                                                              5
                                                                                  5 ∗
                           The parametric equations of the scroll F(2, 2) in the dual space (P ) were given
                                                                ∗
                           above.
                           6.3.4 Intersection of Π A and F(2, 2) ∗
                                                                 ∗
                           By replacing the parametric equations of F(2, 2) in the implicit equation of Π A ,we
                           see that Π A ∩ F(2, 2) is given by the intersection of two curves of bidegree (1,2) in
                                            ∗
                           the parameter space P × P :
                                             1
                                                 1

                                    ϕ 1 (t, u)= A 1 − 2B 1 u + C 1 u − D 1 t +2E 1 tu − F 1 u t =0
                                                             2
                                                                                 2
                                       ϕ 2 (t, u)=2B 2 u + C 2 u − D 2 t +2E 2 tu − F 2 u t =0.
                                                          2
                                                                              2
                           We have two cases: either ϕ 1 (t, u) ∩ ϕ 2 (t, u) is finite (4 points) or infinite. We first
                           consider the generic cases, i.e. the intersection contains 4 distinct points (t k ,u k ),
                           k =1,..., 4 and t k  = t j , u k  = u j if k  = j. This will give a classification of the
                           maps of bidegree (1,2) up to change of coordinates and a set of normal forms.
                           6.4 The generic case

                           We first recall the result for the generic complex case (see more details in the article
                           [13]).

                           6.4.1 The generic complex case

                           Generically, ϕ 1 (t, u) ∩ ϕ 2 (t, u) contains 4 distinct points; they correspond to 4 tan-
                           gent planes. They are tangent to all conics of S, along a special torsal line.

                           We can choose these 4 tangent planes in P to be the planes of coordinates (X =
                                                              3
                           0), (Y =0), (Z =0), (T =0).

                           We proved in [13] that, after a suitable change of coordinates and change of pa-
                           rameters, the surface S admits the parametric representation:
                                                 ⎧
                                                         X = tu 2
                                                 ⎪
                                                 ⎪
                                                    Y =(t − s)(u − v) 2
                                                 ⎨
                                                 ⎪ Z =(t − as)(u − bv) 2
                                                 ⎪
                                                         T = sv 2
                                                 ⎩
                           This normal form depends on two moduli a and b. Moreover, the singular locus of
                           the surface is a twisted cubic and has parametric equations (c.f [13]): (t : s)  −→
                           (X : Y : Z : T)
                            ⎧
                            ⎪ X = abt(bt − t − bs + as) 2
                            ⎪
                              Y =(a − 1)(as − bt)(bt − b t + b t + bats − bts − ats + as − bas )
                            ⎨
                                                                                  2
                                                                                        2
                                                      2 2
                                                  2
                                                            2
                            ⎪ Z = a(a − 1)b(bs − t)(bt − t + bts + ats − bats − b ts − bas + b as )
                                                                                       2
                                                                                   2
                                                       2
                                                                                          2
                                                   2
                                                                           2
                            ⎪
                              T =(at − bt − as + bas) s
                            ⎩
                                                   2
   97   98   99   100   101   102   103   104   105   106   107