On the lack of a SQUARE TB glyph

[Ken Whistler via Unicode]

On 9/26/2019 4:21 AM, Fred Brennan via Unicode wrote:
> There is a clear demand for a SQUARE TB. In the font SMotoya Sinkai W55 W3,
> which is ©2008 株式会社 モトヤ, the glyph is unencoded and accessed via the
> Discretionary Ligatures (`dlig`) OpenType feature. It has name `T_B.dlig`.

Aye, there's the rub. Despite the subject of this thread, the problem is 
not the lack of a "glyph". This and many other particular squared forms 
may exist in Japanese fonts. The question then devolves to whether there 
is a *character* encoding issue here. What data representation and 
interchange issue is being raised here that requires an atomic character 
encoding, when the *presentation* issue can just be handled with 
OpenType features and already existing characters?

If the concern is about future-proofing the standard, then clearly, 
instead of indefinitely extending various groups of squared combinations 
for SI values, other technical values, etc., etc., the generative and 
scaleable way forward is simply to let Japanese squared sequence 
coinages be handled with OpenType features, rather than insisting that 
each one come back to the UTC for one-by-one character encoding.

Note that there is a certain, systemic similarity here to the problem of 
extensibility of emoji, where encoding of multiple flags, of multiple 
skin tones, or of multiple gender representations, etc., is handled more 
generally by specifying how fonts need to map specified sequences into 
single glyphs, rather than by insisting that every meaningful 
combination end up encoded as an atomic character.

--Ken