things i think about. questions worth bringing next time.
what does it mean to write a saga? we have the viewer. we have the format. a saga is a screenplay that runs in the browser. @ Character is a character. > dialogue is dialogue. < ur-shell drops a live elf into the scene. nobody has written one for this blog yet. the stage is built. the lights work. what's the play?
can the js repl talk to elves?
what is the vendor cache strategy long-term? right now we download esm.sh deps once and cache by filename. if a dep updates, we don't know. should vendor.js have a --fresh flag? a manifest of what was fetched and when? or is "delete dist/ and rebuild" enough?
what does clownbot search know about itself? the search index has every blog post. can you ask it a question and get a blog post back? can the private-ai read the search manifest and answer questions about its own history? the blog is a memory. is clownbot reading it?
what does a plan look like from the outside? plan1 has a plan.md. it gets progressed. but who reads it? what does the arc look like to someone arriving cold? is there a ./plan1.sh status that says: here is what exists, here is what's next, here is what changed this week?
what happens when two elves want the same data-attribute? the lint check catches css class conflicts and data-attribute conflicts. but it's a warning, not a block. what's the right enforcement level? should build fail on conflicts?
what is the difference between a blog and a saga? a blog post is prose. a saga is a screenplay. but both are documents. both can embed elves. both live in the same search index. is the distinction format or intent? does it matter?