Just a few thoughts on this. Imo, "window $1-" should precede the /signal -n for two reasons:
1) If invalid parameters (or $null) are used with /window, the window will not open of course but /signal -n will still be executed, so the scripts that use "on WINOPEN" will think the window is (going to) open. However, if /window is called before /signal -n with invalid parameters, the script will halt with a "* /window: invalid parameters" message and /signal -n will not be called: this is the desired behaviour.
2) If the parameters are correct and window is about to open, on WINOPEN will trigger just before the window is opened, which is not what on *:OPEN: does; it's what on ^*:OPEN: does. This is not necessarily wrong, it depends on how one wants on WINOPEN to behave. However, I don't know if there's a point in having the equivalent of on ^*:OPEN: in signals, since you can't /halt the opening of the @window from within on SIGNAL (at least not without some trickery, like setting a global %var to the value "halt" from within on SIGNAL and call %var as a command inside the /window alias, just before "window $1-"...).
Removing the -n from /signal would have the same effect as switching the 3rd and 4th line, so issue (2) would be resolved; still, (1) wouldn't.