This is the Dutch contribution for BIP: 39, titled Mnemonic code for generating deterministic keys. See https://github.com/bitcoin/bips/blob/master/bip-0039.mediawiki and https://github.com/bitcoin/bips/blob/master/bip-0039/bip-0039-wordlists.md . The selection criteria used are discussed step by step below.
An exact number of 2048 words must be selected.
Words may consist only of lower case characters a
to z
.
According to https://github.com/bitcoin/bips/blob/master/bip-0039.mediawiki we
choose a minimum word length of four characters. According to the output of
./analyse.sh
in tools
, a maximum word length of eight is chosen.
At least the first four letters of each word must be unique. According to analysis, the least ocurring first letter should occur minimally four times.
Words will be preferred based on their word frequency from quality corpora.
Words with only the following POS-tags are considered: * * *
Words that have multiple POS-tags are ignored as to avoid unambiguous word meaning. TODO maybe not
Words which are very similar in spelling or sound should be avoided. TODO
By the type of characters found in the selected words, a normal sorting can be used. No special localized sorting is needed.