Add bip39 Indonesian wordlist #621

pull perlancar wants to merge 2 commits into bitcoin:master from perlancar:master changing 3 files +2059 −0
  1. perlancar commented at 3:26 am on January 1, 2018: none

    How the wordlist is produced:

    1. Download and uncompress https://dumps.wikimedia.org/idwiki/latest/idwiki-latest-pages-articles.xml.bz2 (the version used when producing the wordlist is 2017-121-30).
    2. Count the words in all the articles inside articles.xml using this script https://github.com/perlancar/perl-WordLists-ID-Common/blob/master/devscripts/count-words-in-mediawiki-articles . The result is https://raw.githubusercontent.com/perlancar/perl-WordLists-ID-Common/master/devdata/words.txt .
    3. Curate the words manually (mostly removing non-Indonesian words). The result is https://raw.githubusercontent.com/perlancar/perl-WordLists-ID-Common/master/devdata/words-curated.txt . You can diff this two wordlist text file to see the difference.
    4. Generate the BIP39 Indonesian wordlist using this script https://github.com/perlancar/perl-WordList-ID-BIP39/blob/master/devscripts/gen-wordlist . This script basically picks the most frequent words in words-curated.txt that are not already in the English, Spanish, French, and Italian BIP39 wordlist.
  2. Add bip39 Indonesian wordlist 2b35f485a4
  3. dabura667 commented at 6:38 am on January 1, 2018: none

    Technical Checklist ACK 2b35f485a449619060af724ea88a6ef0e1f44e43

    Checked:

    1. Is NFKD normalized list
    2. No BOM
    3. Has single trailing LF line break and is separated by LF line breaks
  4. perlancar commented at 7:13 am on January 1, 2018: none
    @dabura667 Like English, Indonesian only uses 26 Latin letters and encodable in ASCII so no need to put in BOM and asciibetical sorting suffices. Please advise about trailing LF and LF line breaks, because that’s how most of the other wordlist files use too.
  5. dabura667 commented at 8:07 am on January 1, 2018: none
    ACK means “ok to merge” and “Technical Checklist” means I checked for formatting errors and everything is OK.
  6. perlancar commented at 8:13 am on January 1, 2018: none
    Ah ok, thanks.
  7. Add Perl bip39 implementation: Bitcoin::BIP39 4fcfed70d2
  8. nym-zone referenced this in commit b2f66ba1a4 on Jan 5, 2018
  9. nym-zone commented at 3:18 pm on January 5, 2018: contributor

    Has this any independent review from other native speakers or experts in the language? The earlier BIP wordlist addition pull requests witnessed much lively discussion and examination. One was superseded by a new proposal after significant problems were found.

    Since this is said to be ASCII, it is easy to check some basic characteristics in addition to those checked by @dabura667:

    0$ grep '^[^a-z]' indonesian.txt
1$ grep -Eo '^[a-z]{0,3}$' indonesian.txt
2$ grep -Eo '^[a-z]{4}' indonesian.txt | sort -s | uniq | wc -l
32048

    Sample 24-word mnemonic (outside of code-tagging to permit line breaks):

    $ easyseed -b 256 -l id keuskupan utuh kegunaan serta pesisir mungkin reguler cermin langsung enam parkir lari gaib bensin babak dinilai meluncurkan mandiri bijaksana keamanan domestik bercerita prefektur legislatif @perlancar, a question from an implementer: Are these strings for identifying this wordlist sensible and appropriate for Indonesian-speaking users?

    0+	LANG(indonesian,	u8"Bahasa Indonesia",	"id",	ascii_space ),

    I have proposed on bitcoin-dev that native language strings and short ASCII codes should be standardized: https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2018-January/015498.html

  10. nym-zone commented at 3:37 pm on January 5, 2018: contributor

    @perlancar, I recommend that you split 4fcfed70d2761c10c2cd228db8c9310af15f64e6, “Add Perl bip39 implementation: Bitcoin::BIP39”, into a separate pull request. That should be reviewed separately from the wordlist; and it’s not even mentioned in the title or description of this pull request #621. People who glance through the list of pull requests will not even realize that this exists.

    It is good that you did atomic commits, so that 2b35f485a449619060af724ea88a6ef0e1f44e43 (the proposed Indonesian wordlist) and 4fcfed70d2761c10c2cd228db8c9310af15f64e6 (listing of the Perl module in bip-0039.mediawiki itself) can be handled separately.

    I am not a maintainer here; therefore, I can only make a “recommendation”.

  11. perlancar commented at 11:15 pm on January 5, 2018: none
    @nym-zone I did separate into two pull requests, but admittedly I didn’t create a branch for the first PR. And then when I created the second PR, Github merged the two.
  12. perlancar commented at 11:23 pm on January 5, 2018: none
    @nym-zone Thanks for the input. I did publish the proposed Indonesian wordlist as a Perl module (mentioned above) in hope of gathering some input, but of course the number of Indonesian Perl users is negligible. I will try to solicit input from more communities. For your information, I am a native Indonesian speaker. The wordlist were selected from most common Indonesian words which I have curated manually so I would say they are sensible. But that is my opinion.
  13. nym-zone referenced this in commit d03ddae008 on Jan 6, 2018
  14. nym-zone referenced this in commit 8aaa6f37e8 on Jan 7, 2018
  15. nym-zone referenced this in commit c7d698a35f on Jan 11, 2018
  16. ubunteroz commented at 3:59 pm on May 27, 2018: none

    Hello, native Indonesian speaker here. Thanks @perlancar for your effort! Quick review from your wordlist:

    1. It’s better to remove conjuctions (such as atau, tetapi, yaitu, and yakni) to avoid confusion.
    2. Also, non-root words can cause another confusion (peperangan -> perang, pedesaan -> desa, menyebabkan -> sebab).
    3. Limiting wordlist to max. 8/9-letter words can help us to remember easily.
  17. perlancar commented at 9:42 pm on June 2, 2018: none
    Hi Surya (@ubunteroz), thanks for the comments. All are good. I will manually construct an edited new wordlist when I have some free time, but PR’s or patches are welcome :)
  18. luke-jr commented at 5:19 am on July 5, 2018: member

    @dabura667 Only authors are supposed to ACK changes to BIPs.

    In this case, these people: @slush0 @prusnak @voisine @ebfull

  19. luke-jr added the label Proposed BIP modification on Jul 5, 2018
  20. dlwhitehurst cross-referenced this on Jul 11, 2018 from issue Add Indonesian Word List once Accepted (bip39) by dlwhitehurst
  21. DonaldTsang cross-referenced this on Dec 24, 2018 from issue Binary Lists by DonaldTsang
  22. in bip-0039/indonesian.txt:1989 in 4fcfed70d2 
    1984+unsur
1985+untuk
1986+upacara
1987+upah
1988+upaya
1989+upeti
    heri16 commented at 3:55 am on July 3, 2019:
    This word is unfamiliar even to native speakers.
    perlancar commented at 8:27 am on July 3, 2019:
    Sorry, by “this word” which one are you referring to? Because GitHub Web UI is showing 4 words: upacara, upah, upaya, upeti. I would argue that they are familiar to native speakers. What is your basis to say that they are not? Let’s take “upeti” for example, this word is found 792 times in the Wikipedia article (see PR description which I have just updated to describe the process of producing the wordlist).
    perlancar commented at 8:32 am on July 3, 2019:

    Native indonesian speaker here. Instead of the Wikipedia Corpus (which contain too many technical terminology), have we considered the “Sari Kata Bahasa Indonesia” book?

    This book is used nationwide in junior/primary schools. It would ensure anyone with a basic education would be able to understand our wordlist.

    This same book is also used as a reference to test if a foreigner can be permitted to work in Indonesian companies.

    This is a good idea, but what about license? And is there an online version for it?

  23. in bip-0039/indonesian.txt:1954 in 4fcfed70d2 
    1949+tumbuh
1950+tunai
1951+tunduk
1952+tunggal
1953+tuntutan
1954+turbin
    heri16 commented at 3:56 am on July 3, 2019:
    This word is unfamiliar even to native speakers.
    perlancar commented at 8:28 am on July 3, 2019:
    Which word are you referring to? tunduk, tunggal, tuntutan, turbin? All of them? What is the basis? (See my comment on the upeti word above).
  24. in bip-0039/indonesian.txt:1935 in 4fcfed70d2 
    1930+topik
1931+tradisional
1932+tragedi
1933+trailer
1934+transportasi
1935+trek
    heri16 commented at 3:57 am on July 3, 2019:
    This word is unfamiliar even to native speakers.
    perlancar commented at 8:29 am on July 3, 2019:
    Which word are you referring to? trek? I tend to agree.
  25. in bip-0039/indonesian.txt:1915 in 4fcfed70d2 
    1910+tiga
1911+tiket
1912+tikus
1913+timah
1914+timbul
1915+timnya
    heri16 commented at 3:58 am on July 3, 2019:
    This word is a conjugation.
  26. in bip-0039/indonesian.txt:1901 in 4fcfed70d2 
    1896+terowongan
1897+terpilih
1898+tersebut
1899+tertentu
1900+terutama
1901+terwujud
    heri16 commented at 3:58 am on July 3, 2019:
    This word is a conjugation.
    perlancar commented at 8:29 am on July 3, 2019:
    I agree that conjugated words should be avoided.
  27. in bip-0039/indonesian.txt:1900 in 4fcfed70d2 
    1895+ternyata
1896+terowongan
1897+terpilih
1898+tersebut
1899+tertentu
1900+terutama
    heri16 commented at 3:58 am on July 3, 2019:
    This word is a conjugation.
  28. in bip-0039/indonesian.txt:1854 in 4fcfed70d2 
    1849+tantangan
1850+tanya
1851+tapi
1852+tari
1853+tata
1854+tatkala
    heri16 commented at 3:59 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  29. in bip-0039/indonesian.txt:1834 in 4fcfed70d2 
    1829+tadi
1830+tahap
1831+tahta
1832+tahun
1833+tajam
1834+takhta
    heri16 commented at 4:00 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  30. in bip-0039/indonesian.txt:1821 in 4fcfed70d2 
    1816+surya
1817+susah
1818+susu
1819+sutradara
1820+swasta
1821+syair
    heri16 commented at 4:00 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  31. in bip-0039/indonesian.txt:1812 in 4fcfed70d2 
    1807+sungai
1808+suntingan
1809+supaya
1810+support
1811+surat
1812+surel
    heri16 commented at 4:01 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  32. in bip-0039/indonesian.txt:1791 in 4fcfed70d2 
    1786+suaminya
1787+suara
1788+suasana
1789+suatu
1790+subjek
1791+subspesies
    heri16 commented at 4:01 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  33. in bip-0039/indonesian.txt:1783 in 4fcfed70d2 
    1778+stasiun
1779+status
1780+stop
1781+strategis
1782+string
1783+stroke
    heri16 commented at 4:01 am on July 3, 2019:
    This word might be unfamiliar to some native speakers.
  34. in bip-0039/indonesian.txt:1782 in 4fcfed70d2 
    1777+star
1778+stasiun
1779+status
1780+stop
1781+strategis
1782+string
    heri16 commented at 4:02 am on July 3, 2019:
    This imported word is unfamiliar to almost all native speakers.
  35. in bip-0039/indonesian.txt:1777 in 4fcfed70d2 
    1772+spons
1773+stabil
1774+stadion
1775+staf
1776+standar
1777+star
    heri16 commented at 4:02 am on July 3, 2019:
    This imported word is unfamiliar to almost all native speakers.
  36. in bip-0039/indonesian.txt:1745 in 4fcfed70d2 
    1740+silsilah
1741+siluman
1742+simbol
1743+sinar
1744+sinetron
1745+singel
    heri16 commented at 4:03 am on July 3, 2019:
    This imported word is has multiple spellings debated by native speakers.
    perlancar commented at 8:30 am on July 3, 2019:
    Agreed.
  37. in bip-0039/indonesian.txt:1739 in 4fcfed70d2 
    1734+signifikan
1735+sihir
1736+sikap
1737+siklus
1738+silat
1739+silinder
    heri16 commented at 4:05 am on July 3, 2019:
    Technical term for most native speakers.
    perlancar commented at 8:30 am on July 3, 2019:
    I can agree to this.
  38. in bip-0039/indonesian.txt:1705 in 4fcfed70d2 
    1700+sengaja
1701+seni
1702+senjata
1703+sensus
1704+sentral
1705+senyawa
    heri16 commented at 4:07 am on July 3, 2019:
    Is a Conjucation or Technical term for most native speakers.
    perlancar commented at 8:30 am on July 3, 2019:
    Agreed.
  39. in bip-0039/indonesian.txt:1694 in 4fcfed70d2 
    1689+selisih
1690+seluruh
1691+semakin
1692+sembilan
1693+sementara
1694+seminggu
    heri16 commented at 4:09 am on July 3, 2019:
    Conjugation
    perlancar commented at 8:30 am on July 3, 2019:
    Agreed.
  40. in bip-0039/indonesian.txt:1637 in 4fcfed70d2 
    1632+salju
1633+saluran
1634+sama
1635+sambil
1636+sampai
1637+samurai
    heri16 commented at 4:11 am on July 3, 2019:
    Imported word not familiar to most native speakers.
    perlancar commented at 8:31 am on July 3, 2019:
    Agreed on imported words, though I don’t share your opinion of “samurai” being not familiar with most native speakers.
  41. heri16 changes_requested
  42. heri16 commented at 4:47 am on July 3, 2019: none

    Native indonesian speaker here. Instead of the Wikipedia Corpus (which contain too many technical terminology), have we considered the “Sari Kata Bahasa Indonesia” book?

    This book is used nationwide in junior/primary schools. It would ensure anyone with a basic education would be able to understand our wordlist.

    This same book is also used as a reference to test if a foreigner can be permitted to work in Indonesian companies.

  43. perlancar commented at 3:33 pm on July 3, 2019: none

    I have created another wordlist. The source is still from Wikipedia Indonesia, but this time I manually curate the BIP wordlist using this criteria:

    1. Words from 4 to 10 letters long.
    2. Avoid conjugated words, choose only root words.
    3. Avoid prepositions, conjunctions, pronouns (e.g. dan, atau, jika, kamu, saya, …).
    4. Avoid technical words when possible.
    5. Avoid loan words when possible.
    6. Avoid words that have multiple competing spellings, if possible.
    7. The above criteria is balanced/countered by this: I want to make the BIP wordlist unique by its 4-letters. So even if user does not type or mistype some letters at the end of words, they can still be corrected.

    The resulting BIP wordlist is put here: https://github.com/perlancar/perl-WordList-ID-BIP39/blob/master/devdata/words-bip.txt. I haven’t made a change to this PR. Inputs/comments welcome.

    As you might see, some of the words are indeed not very popular to native speakers, because I also want to satisfy criteria number 7.

    Also useful is the larger wordlist from which I curated this BIP wordlist: https://github.com/perlancar/perl-WordList-ID-BIP39/blob/master/devdata/words.txt . I indented the words which I want to use in the BIP wordlist. You can suggest corrections or changes by submitting a PR which changes this file.

  45. luke-jr closed this on Jul 2, 2021


perlancar dabura667 nym-zone ubunteroz luke-jr heri16

Labels
Proposed BIP modification

github-metadata-mirror

This is a metadata mirror of the GitHub repository bitcoin/bips. This site is not affiliated with GitHub. Content is generated from a GitHub metadata backup.
generated: 2024-12-28 03:10 UTC
This site is hosted by @0xB10C
More mirrored repositories can be found on mirror.b10c.me