I can see that. There was (and is) a plethora of object systems, which could be seen as a double (triple?) edged sword. It provided choice, with many alternate features and methods (no pun intended), but perhaps at the expense of developer clarity - which one best serves my needs? Will I end up in a dead end? What are most others using? What serious problems might I be adopting? The "third edge" though, is a positive, though took time to arrive: the TclOO that was finally developed was informed by the technical and social impacts of years of experience with various OO systems, and indeed emerged mature and well-formed because of that. In fact, it was initially developed to only be an underpinning for other flavours of OO sugar to be bolted on top, in support of existing systems like [incr tcl] and Snit. It ended up being received as system good enough to use directly, in it's own right.
I've never have the chance to use the new core TclOO, but I really think they've made mistake at that time by implementing another new object system instead of promoting an existing one like [incr tcl] (even if not perfect/incomplete).
