Minor updates to comments

README.md

@@ -61,7 +61,7 @@ Several scripts are distributed for the user to generate $C$ matrices for standa
 These scripts generate unweighted adjacency matrices, then the user should weight the edges.
 A sample generating a weighted, directed Watts-Strogatz ring network of $N$ nodes, with in-degree $d$,
 and rewiring probability $p$, without self-connections allowed from re-wiring, and ensuring 
-the network remains connected:
+the network remains at least weakly-connected:
 
 ```matlab
 % First generate the unweighted adjacency matrix without self-edges:
@@ -96,6 +96,8 @@ As such, there are two steps involved here:
 % Then compute sigma^2:
 exp_sigma_sqr = synchronizability(Omega_U);
 ```
+The call to `covarianceUGaussianNet` will throw an error if the matrix does not meet synchronisation conditions,
+or will not ensure guaranteed convergence of $\Omega_U$.
 
 ## 1.3 Batch experiments involving parameter sweeps and repeat runs
 

generateNewRandomRingMatrix.m

@@ -12,7 +12,7 @@
 %   For undirected graphs, we don't allow connection to an odd number of
 %   other nodes (too unconstrained to work out who to connect to)
 % - allowSelf - whether to allow rewired connected to be made to oneself.
-% - ensureConnected - only return a connected matrix.
+% - ensureConnected - only return an at least weakly-connected matrix.
 %     If this is set to true, then p must be >= 2/N(N-1) for undirected, 1/N(N-1) for directed, for the matrix to be connected.
 %     Note: if allowSelf==true, then (N-1) -> N here.
 %
@@ -133,7 +133,7 @@
     %printf('Rewired %d out of %d edges considered\n', rewirings, edgesConsidered);
 
     if (ensureConnected)
-        % Now check that the matrix is connected
+        % Now check that the matrix is at least weakly connected
         symmA = A | A';
         visited = zeros(1,N);
         visited = runDfs(symmA,N,1,visited);

parametersTemplate.m

@@ -32,7 +32,7 @@
 %   So (b - c) is the strength of coupling to previous state of the destination.
 %   For stability we require |b| < 1 when undirected, though b == 1 ensures
 %   zero-mode as an eigenvector (for our first sync paper we only use b =
-%   1).
+%   1). Using this then does not guarantee stability in this case.
 parameters.b = 1.0;
 % - c - as above; c is the total strength of non-self coupling of the sources to the destinations.
 %   For stability we require |b - 2c| < 1 when undirected (though I don't