This morning I had some inspiration about my home in lambdaMOO. The last few weeks my goal statements have been something like: Well you know, there's common lisp, and MOO, and clim application frame commands, and emacs...! However I feel like I've put these into order a bit. Firstly, a MOO is not a MUD nor a MUSH. In a MOO, everything in the world is made in MOO. Maybe there are some primitive operations to start with that needed to be there before the first new objects were written in MOO. In contrast, a MUD is programmed in a host programming language. It is not self-hosted. Nor is it a MUSH: Where an in-game command language is implemented in the host language, then used for complex activities in-game. Changes of nature would be programmed in the host language. So I need to address the separation of MOO, common lisp and clim commands. To start with, to preserve the MOO as a MOO, the main functionality should just be accepting input and passing this through to the MOO server, and replying with the server's output to me, the player. In this case I am imagining a clim application-frame interactor with completions, but basically performing a pass-through to the moo server via rmoo-mode in emacs (my moo client). What is the relevance of common lisp to this MOOxperience? I've got a good one: As well as creating real history by passing entered text directly to the MOO server and getting the definitively real responses, which I identify as time contexts, add Sandewall's presentation of imaginary contexts as a generalisation of time contexts being conjectures about future times based on my domain model representations, but also conjectures about what could have been (imaginary past timelines). These are imaginary simulations of the actual MOO I am also literally sitting in, and the day dreaming advises what I will do in the real history of the MOO (when I finish daydreaming and do something). In particular, I expect my daydreaming to resemble my own second implementation of Sandewall's zookeeper lab from their 2010 cognitive agent intelligences lab, without the intent to recreate all of Sandewall's novice-5.zip (used in that lab). My engine will be the world-of-shouting. This simplification is acceptable because a highly local MOO history's performed verbs and property changes are more exactly like the imaginary zoo scenario's world. (The zookeeper wanders around the zoo feeding animals, cleaning, detecting and performing medical care for animals, assisting birthing animals, staying hydrated themselves, entertaining zoo visitors..). So the MOO remains a MOO. But in good lisp company, one of my activities while I am literally in the moo is writing a common lisp imaginary context simulator of highly local subsets of lambdamoo. I guess the code will sit in emacs, rather than being mailme'ed from moo notes.