Greek and Arabic constructions of the regular heptagon (Q5895464)

[A critical comment can be obtained on request from the editorial office.] The project of research planned for this extensive paper might be well chosen some century ago, but appearing after the second half of the 20th century it could only lead to disappointments. With the discovery of the regular polygons, treated and computed in the 2nd millennium B.C., it is clear that these studies had to be based on the isosceles trapezium, the next step after ''Pythagoras'' to the relation of ''Ptolemy''. It replaces all which can be done by ''trigonometry'', governed by the rule that the square of the diagonal is equal to the sum of the square of the equal sides and the product of the parallel sides. The fact that al-Birûni asked to show that the side of the enneagon is a root of a cubic equation demonstrates that the equations were looked for a priori. In the babylonian ''geometry of the plummet'' [cf. the reviewer, \textit{Simon Stevin} 33, 38-60 (1950; Zbl 0089.003)] the problem of the heptagon was treated. The heptagon has three determining quantities, a,d,D viz. the length of the side and two diagonals. Specifying the isosceles trapezia by the sequence of numbered vertices one obtains four quadratic relations: \(d^ 2=a^ 2+aD, (1234)\); \(D^ 2=a^ 2+dD, (1245)\); \(D^ 2=d^ 2+ad, (1246)\) and from these by additions and subtractions a fourth relation, a ptolemaic relation: \(dD=ad+aD\quad (4).\) The first and the third relation are those mainly quoted in a ''lemma of Archimedes''. The present reviewer wrote since 1959 the quotation marks indicating his opinion that this attribution is not correct. A transversal in a square can be used to solve cubic equations. We see from completing rectangles - PESD, ERQT - that \(z^ 2=y(y+x),\) a relation of the form (1), and that the double of the area EAB is \(ay=S_ 1\). With \(BF=k\) we have \(ay=kz\) and two times the \(S_ 2\)-area CFG is \(x(a- k)=ax(z-y)/z.\) This leads for the condition \(pS_ 1=S_ 2\) to \(pyz=xz- zy,\) which is all we need to see that a second equation for the heptagon arises for \(p=1\) in (4), whence in x,y,z we have quantities D,a,d of the heptagon. In such homogeneous systems one can choose one ''unknown'' and is left with quadratic equations, which can be solved by intersection of conics, and two conics need a third in order to eliminate duely a ''parasitic'' solution. Through three points pass \(\infty^ 2\) conics and thus there are \(\infty^ 4\) possible choices for a fitting pair of conics. One can therefore only meet with ''particular cases'', and the more cases one has treated the ''more general'' the work is considered. We think that the blaming by the author of R. Rashed is unjust. The author's remark that the problem comes out to determine sin 180\(\circ /7\) is certainly futile. In fact till the time of Huygens and De Sluse one had no idea of pencils and nets of conics and curves, and one did not combine ''geometrical equations'' in reductions. It must be left to the interested reader to find - what the author neglected to indicate - how just specialising a chosen quantity the equations of the conics used by the Arabs are just a pair of the above given tetrad. This is unnecessarily complicated by the author's deliberately drawing mirror-images of some figures in the manuscript. This causes an, impossible for the Arabs, coordinate transformation \(x=- x'\), \(y=y'.\) As the author did not treat the possibilities in the several cases he mentions, we have to indicate here that al-Kuhi - just as Menaichmos - did not see the fact that the points of intersection of his first, and partial, hyperbola and parabola are concircular. By the work of the author we find put together lists of manuscripts and verbatim translations on solutions, quarrels between young geometers on errors, mistakes and priority, and have instead of the old about seven published solutions some 12 combinations of conics out of the \(\infty^ 4\) making the solution ''general'' and not only some particular case. This is hardly a contribution to the problem-history of mathematics itself. One might ask whether one should waste his time on such a project, just for completing some sections of not yet edited manuscripts, on an already well known situation and the level of knowledge of the period.