As a way of avoiding using script shifts (selma'o BY1), and of having to memorise Old Church Slavonic names for Cyrillic, and heaven knows what we'd do with Arabic "teh",

I (nitcion) propose the following sleazy lerfu cmene scheme:

• Cy/V name of lerfu, per normal Lojban
• + CC or CVC for script provenance. With the following list pulled out of thin air:

• Roman: l (from lo'a)
• Greek: g (from ge'o)
• Cyrillic: r (from ru'o)
• Arabic: x (from xrabo)
• Hebrew: j (from je'o)
• Armenian: dj (from xadjestan?)
• Georgian: k
• Devanagari: d
• Katakana: p (from ponjo)
• Hiragana: pix (ponjo xiraganan... Oh, whatever)

Which would mean

• Roman b: by or byl.bu
• Greek beta: ge'o by or betas.bu or byg.bu
• Cyrillic be: ru'o by or bukys.bu or byr.bu
• Arabic beh: jo'o by or byx.bu
• Hebrew bet: je'o by or bet.bu or byj.bu
• Devanagari ba: zai.devanagar.bu by or byd.bu

I mean, it can't be any more broken than the current system...

----

What about just using Unicode character codes? me'o se'erenodaugai would be the euro sign. Arabic beh would be me'o se'enoxarebi. And so on. mi'e filip

we can already do so; but numbers have no mnemonic value. The Unicode mailing list sees a fair bit of codes, but I think the majority of usage is still names (and the standard provides both.) — n.