[Cryptography] Uniform Data Fingerprint

[Arnold Reinhold]

> On Jun 1, 2015, at 1:34 PM, Phillip Hallam-Baker <phill at hallambaker.com> wrote:
> 
> 
> 
> On Mon, Jun 1, 2015 at 10:27 AM, Arnold Reinhold <agr at me.com <mailto:agr at me.com>> wrote:
> A small suggestion: instead of slavishly following RFC4648 Base-32 format, I would suggest a modified Base-32 where the letters I, O, and L  are excluded from the encoding alphabet as they are too easily confused with the digits 0 and 1 (L being confusing in lower case). Maybe exclude S as well for potential confusion with 5. Only 32 of the possible 36 characters are needed, so excluding three or four characters poses no technical problem. Here is a worst-case example I concocted: 0lSS0-l1151-I5S1O-I05S1 Of course the potential for confusion is font dependent but I don’t think it is practical to specify fonts in such a standard.
> 
> Since the fingerprint strings are intended for human recognition, we should do everything possible to minimize confusion.  At the standards writing stage, this is a cheap improvement.
> 
> Base32 does the same thing by excluding the corresponding numbers. Of course, both could be less confusing. The coding was originally proposed by Phil Zimmerman.
> 
> Maybe an explanatory note… 

I should have read the RFC more carefully. There is a second "Base 32 Encoding with Extended Hex Alphabet” that confused me (section 7).  “Base-32 Encoding” (section 6) as you say discards the digits 0 and 1, but not 5. Close enough. 

It’s good to know that they and you got it right the first time.

Arnold Reinhold

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20150601/176a8ed8/attachment.html>