Conway groupoids and completely transitive codes (Q722342): Difference between revisions

Let \({\mathcal D}=(\Omega,{\mathcal B})\) be a supersimple \(2\)-\((n,4,\lambda)\) design, so any two blocks in \({\mathcal B}\) intersect in at most two points. For any two points \(a,b\in\Omega\), a permutation \([a,b]\in\text{Sym}(\Omega)\) is defined: \[ [a,b]=(a,b)\prod_{i=1}^{\lambda}(a_{i},b_{i}) \] where for \(i=1,\dots,\lambda\), \(\{a,b,a_{i},b_{i}\}\in{\mathcal B}\). Then for \(a_{0},a_{1},\dots,a_{k}\in\Omega\), \(k\geq1\), define \[ [a_{0},a_{1},\dots,a_{k}]:=[a_{0},a_{1}][a_{1},a_{2}]\dots[a_{k-1},a_{k}]. \] Further, fix a point \(\infty\in\Omega\), and set \[ {\mathcal L}({\mathcal D}) =\{[a_{0},a_{1},\dots,a_{k}]\mid k\geq 1,a_{i}\in\Omega\text{ for }0\leq i\leq k\}, \] \[ {\mathcal L_{\infty}}({\mathcal D}) =\{[\infty,a_{1},\dots,a_{k}]\mid k\geq 1,a_{i}\in\Omega\text{ for }1\leq i\leq k\}, \] \[ \pi_{\infty}({\mathcal D}) =\{[\infty,a_{1},\dots,a_{k-1},\infty]\mid k\geq1,a_{i}\in\Omega\text{ for }1\leq i\leq k-1\}. \] The set \(\pi_{\infty}({\mathcal D})\) is called the hole stabilizer. A Conway groupoid \({\mathcal C}({\mathcal D})\) is the small category whose object set is \(\Omega\) and for \(a,b\in\Omega\), \[ \text{Mor}(a,b)=\{[a,a_{1},\dots a_{k-1},b]\mid k\geq1,a_{i}\in\Omega\text{ for }1\leq i\leq k-1\}. \] So the set of all morphisms in the category \({\mathcal C}({\mathcal D})\) is equal to the set \({\mathcal L}({\mathcal D})\). In this article, two infinite families of \(2\)-\((n,4,\lambda)\) designs \({\mathcal D}\) such that some elements of \(\text{Alt}(n)\) do not occur as morphisms of \({\mathcal C}({\mathcal D})\). Such Conway groupoids are referred to as ``not full''. The first one arises from the \(2\)-transitive actions of \(\text{Sp}_{2m}(2)\) on sets of quadratic forms over a \(2m\)-dimensional \(\mathbb{F}_{2}\)-vector space for \(m\geq3.\) The second one arises from the natural action of the affine group \(2^{2m}\cdot\text{Sp}_{2m}(2)\) on \((\mathbb{F}_{2})^{2m}.\) Then two main theorems were proved which give some implications: 1. They give classifications of Conway groupoids subject to certain extra suppositions; 2. They give new characterization of the Conway groupoid determined by \(M_{13}\); 3. They show that for each \(\lambda>0\), there are only finitely many supersimple \(2\)-\((n,4,\lambda)\) designs with the associated Conway groupoids not full. We make a remark here. A \(2\)-design is supersimple if and only if that three distinct points of the design are contained in at most one block. Such a design is said to be circular by \textit{J. R. Clay} [Ann. Discrete Math. 37, 95--105 (1988; Zbl 0684.05008).