GnuPG patch: long fingerprints using PGP biometric word lists

tftp tftp at
Mon Mar 5 21:09:04 CET 2001

--- Detlef Lannert <lannert-gpgspam at> wrote:
> On Mon, Mar 05, 2001 at 09:14:16AM +0100, Matthias Urlichs wrote:
> > > Some of the words appear to be cumbersome to pronounce for non-native
> > > speakers. 
> > 
> > Probably. But that's not important, as the idea is that both partners
> > have the word list in front of them and are just verifying that the
> > words are the same. Whether I pronounce "chambermaid" correctly or
> > not doesn't matter.
> Many folks with a poor command of English won't know how "guidance"
> is spelled or how "breadline" or "scenic" should be pronounced.

Several related problems indeed:

a) This puts a dent into localization effort. 256 foreign words are
not something to be taken easily. The program may not qualify as
completely localized and managers will be wary of that. Nobody
wants to deploy a program that suddenly starts talking English!

b) Some languages are not based on latin characters at all. I'd say,
most languages are not using Latin alphabet! In Chinese, for instance,
there are separate native 10 glyphs for digits 0-9 (-, = etc.)

c) English words -may- be represented by the current font, but this
does not guarantee that the user knows how to read or spell them even
one by one! Some fonts may be wholly national and don't have Latin
in them.

d) Some languages have directionality issues. Arabic is -mostly-
written from right to left, but numbers are written left to right!
Japanese can also be read vertically. Arrays of latin characters
may be hard to grok without understanding words behind the letters,
and even then reading them as words and as numbers may produce
different results!

e) It is better to forget about correct pronounciation of English
words. Many native English speakers still argue how to say this or
that :-) Non-native speaker will not want to read unfamiliar, highly
unusual, mostly unknown words to another person who is also not good
in English! How often, for example, the word "chambermaid" is used

I'd say, leave fingerprints as hex strings. Localization will deal
with correct representation of 10 common digits and 6 extra characters.
Everything beyond that is overkill.

> As a compromise you could convert the fingerprint into a base-26
> number coded as letters and pronounced according to the "Alpha-Bravo-
> Charly" alphabet already mentioned in this thread.

Still assumes that Latin is the foundation of life on Earth. It isn't.

> The question still remains whether "Bravo Delta Lima ..." works
> much better than "Foxtrott Echo Echo Echo, Seven Delta Echo ...".

Foreign languages come with their own lists of words for this purpose.
FE7D could be "Fyodor Elena Sem' Denis", for example, in Russian
(where names are commonly used in such lists).


Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.

More information about the Gnupg-devel mailing list