IRC channel logs
2025-02-18.log
back to list of logs
<jfred>dthompson: I was having trouble with some new propagators I was writing until I reread a paragraph near the end of your post and finally understood what the groups are for and why they're needed <jfred>I'm also kind of still convinced that propagators really want to be expressed in a visual programming language of some sort because of the way you wire up propagators in between cells <dthompson>visual programming would be nice for these, yeah <jfred>Maybe not a node-and-wire language, at least not always. Maybe something more like a schematic diagram, which sometimes omits connections further than one layer out from each component for clarity's sake <dthompson>a text representation is still crucial, though, imo. I like the idea of being able to switch between textual/visual forms of the same code. <jfred>Yeah, agreed, if only for smoother compatibility with things like version control <jfred>Allowing for edits on the visual side without clobbering the formatting on the textual side is probably tricky <jfred>especially with e.g. comments not being part of s-expressions <jfred>then again, the part of propagator programming that really lends itself well to visual layout isn't the propagator definitions themselves, but the links between them. the definitions themselves still want to be textual, I think <dthompson>I still think there is a much better DSL lurking here, though <Zarutian_iPad>look into how FlowBasedProgramming visual editors do these kind of things <jfred>the point about text being great for trees but not graphs is making me wonder if there's any old research to mine on textual graph programming <jfred>but yeah maybe there's a web-specific DSL lurking here that would be more straightforward <av0n>I remember wanting some sort of "S-expr, but for graphs" when I was researching propagators a while ago. I have no clue what that would look like though.