On Sun, 29 Nov 2015 22:12:27 +0000, Isolde Helmut <[log in to unmask]> wrote:

>>  For a (2-person) dialogue, there would have to be 2 superenclosures,
>>each with its "tell" instance. The 2 speeches wouldn't be synchronized, so
>>there couldn't be any words for "yes" or "no"; answers would be full
>>sentences. In fact, a speech wouldn't have to have an internal
>>chronological order.
>You could use more than two "tell" instances with a time ordering, if you
>wanted a true linear conversation.

That's true; there need only be 1 instance for each name phrase, avoiding the need for a nickname strategy (nicknames are an unimplemented idea I had where an unenclosed glyph is identified with a name phrase). There would be 3 lines linking the "tell" instances, which would probably be down the middle, with quoted sentences on the sides.

>> Even with these strategies, I will have to figure out specifics. Any
>I suppose you might be treating it like a spoken language, where you'd have
>to come up with some semantic roots at some point. Perhaps a 'root' would
>be two of your three components. Arbitrarily chosen, I suppose, for SW0,
>unless you have aesthetic or ergonomic (is there a more suitable word for
>'relating to ease-of-writing'?) considerations.

2 components doesn't give me quite enough roots, so I think root + extension will be limited to certain word sets. I don't have any ergonomic considerations, unless that includes the order the components are written, assuming they're drawn, not just displayed as a whole.