Woof

There are hunters with dogs nearby, Claudio's really agitated.

Before getting out of bed, in my waking-up research session I got annoyed (again) at my notetaking setup. On the phone I currently use Markor markdown editor, which is autosync'ed with my desktop using SyncThing. I have dev in progress on #:squirt for improving this, but the frustration did give me an idea: in-browser editor, two-pane, using #:semtags as hyperlinks that when double-clicked in one pane, they open the corresponding URL in the other pane. Worth a #:spike! So I did a little dev with Claude (while still in bed). Got some starter code, and with my initial prompt and Claude's handover doc there should be enough to take it forward as and when. Even if the idea is a good one, the implementation won't be properly useful until I have a bit more infrastructure - #:tbox, specifically #:semem - in place. So I've dumped all the in the #:squirt part of the #:farelo repo for now. Now at desk, onto that infrastructure.

One thing that led me to the above is that I always struggle to remember where I'm up to with things. Being able to nav to (sub)project status/next steps quickly is must-have. My org is lousy right now, finding things hard. The idea of having One Big Top Level (yay! got a name for my reggae soundsystem) doc crossed my mind. Maybe... But hypertext, for heavens sake...patience Danny. I need #:semem to do search asap.

  1. make a #:transmission to walk fs, post files off to SPARQL server (in #:tbox, but not tightly coupled)
  2. bit of UI - HTML form with Claude's help I guess quickest
  3. get #:semem to build index
  4. search form

I ended up spending the morning tidying up #:semem, getting tests in sync with code. Ran out of Claude tokens. Dogwalk.

semem CLI

semem create store hyperdata

creates a SPARQL dataset at default URI (Fuseki API how, GSP?)

semem index [URI]

I want articles stored as named graphs. How best to do containment?

Hmm. Graphs that only contain indexes? Stick indexes in default graph?

Woof