I was reading the BIP 39, and noticed that it does not refer to how the mnemonic sentence, used for seed calculation, is composed from mnemonic words. This is actually troublesome because in case of Japanese, the bip39 npm package joins the mnemonics with IDEOGRAPHIC SPACE(U+3000), the space character typically used as space for Japanese. It seems that copay is using this library and thus follow this rule. Some other implementation may decide to use (more machine-friendly) space(U+20), or something else, which makes remembering just the mnemonic words in sufficient for wallet recovery (in other wallet).
- Is there any standard of how mnemonic sentence is composed from mnemonic words?