Connected Text may not be dead yet!

Posted by SmallDog
Nov 29, 2018 at 06:03 PM

The one thing I most want from CT is better, fuzzier autocomplete.

For one thing, it should default to substring search. As it currently stands, unless you remember the exact words a topic’s name begins with, you simply can’t get to it in the autocomplete dropdown. You have to use the topic search.

Most of the time when I insert a link it’s purely to tie bits of my text together semantically. For example, I might write:

“This reminds me of that experience I recently had in which [blah-blah-blah]”

Now let’s say at some future point I recall that experience, and think: “hey! I remember several times in the past month I pointed out how what I was then experiencing reminds me of that particular experience. It’d be nice to find those references.”

But how? Mostly likely you don’t remember the exact words you used in those references. Worse, you most likely didn’t use the exact same words each time you refer to it. So it looks like you’ll have to grep a word or two and look through the results. A very unpleasant prospect! This where CT’s autocomplete come in handy. It helps you standardize on a single phrase to use when referring to a given idea. Same meaning, same language. Finding all references to some idea is just a matter of finding all backlinks.

For anyone who like to put links to this kind use, linking is probably something you do very often. I can easily have 8-10 links in a single paragraph. But this workflow isn’t supported very well by the current behavior of CT’s autocompletion, simply because you can’t search *inside* a topic name. Sure, you can do a topic search, but it’d much nicer to be able to minimize the context-switch, not having to leave the editor at all.

I keep some of my notes in sublime_zk, which has just this ideal (from my pov) autocompletion behavior (inherited from sublime text itself). But I miss so many of CT’s rich text and scripting capabitlies.