Note: the following article is entirely fictional. Its use is solely as a sketch of the sort of thing I'd write if/when I eventually succeed in my work. How to create your own introspective sandbox from scratch in 1 hour Today, we will create a web page that: Is made of 2D SVG shapes Which … Continue reading What am I aiming for again?
A while ago, I abandoned my graph-drawing track in BootstrapLab and started over in a new direction. This was productive, but now that I'm coming to a stage of increased friction with the sorts of things I'm trying to add, I feel this is a good time to pause and discuss what I've been doing. … Continue reading Reflections on the evolution of a toy system
Shortly after starting my SVG work, I realised I just had to read Ivan Sutherland's 1963 thesis. (I really ought to have already done so, and this was the final kick.) Sadly, it is somewhat hard to read owing to the fact that many fundamental programming concepts had not even been invented back then. And @Pavel, … Continue reading Reading the Sketchpad Thesis
Periodically I must snap myself back into recognising my real aim: to experience the traditional process of bootstrapping, but by doing as much as possible in a "live" system that I can see, feel and persist without touching a single line of "code". After re-orienting myself towards this, I am carefully designing a way to … Continue reading Back to bootstrapping
What distinguishes the computer medium from all the others? What constitutes a "simulator"? What's the minimal "machinery" needed? The Church-Turing thesis says: it's the Turing machine. This does depend on what you mean by "minimal" or "simplest". I actually think that the random-access Register Machine is the best place to start, and I view the … Continue reading Ideas for a “human-friendly” “dumb” substrate
OROM is a mixed bag. On the one hand, it "brings forth" many fruitful ideas I wouldn't have had otherwise. On the other, it is hampered and held back by its rather low-level implementation. There may have been an excuse for this when it was published, but I do not have one today. Consider "the … Continue reading The untapped power of universal simulation
(EDIT: unfortunate title. It represents more of a desire than something that I actually accomplished - lol) I think that version 0 of my self-improving system is finished. It is a hacked-together yet working version of OROM in the browser. Now I would like to reflect upon and analyse what I've done, because I now have … Continue reading 4. Towards a general theory of bootstrapping
Considerations, and rough devlog, for implementing the Id object system described in Open, Reusable Object Models (and related material) in an already-existing visual interactive platform: Objects and vtables An object is a bunch of mutable state (1). In every object A, there is a special part of this state that points to another object (2). This is called the vtable of … Continue reading 3. Hacking together OROM/(DOM+Ctrl+Shift+J)
Up till recently I'd had roughly the following plan: Use JS, the console and Canvas to build a "mark 0" retained-mode graphics system Use the graphics system to build stuff, e.g. the Id object model. But now I realise that I was mistaken on two counts. First, given what I'm aiming for in point #2, … Continue reading [Draft] Make use of what already exists!
Now that I have a way to persist my JS console development, it's time to get to basic graphics as soon as possible. Something I must keep in mind here: I should be able to get away with horrible, hack-ish solutions in the short-term, because the whole point is that I should be able to … Continue reading Canvas: the assembly language of Web graphics