[Standards] XEP-0300 and a future without reading tea leaves

Kevin Smith kevin.smith at isode.com
Fri Jan 13 10:59:27 UTC 2017

On 12 Jan 2017, at 09:48, Tobias Markmann <tmarkmann at googlemail.com> wrote:
> This can be seen as a continuation of my classic thread on this very channel from 2014 titled "File Hashes in XEP-0234". It was so popular, that even I did lost interest in it. :D
> Back to the future, today. XEP-0300 doesn't specify an encoding, the examples clearly use Base64, specifically Base64 as defined in https://www.w3.org/TR/xmlschema-2/#base64Binary and used by many other XML schema. Originally XEP-0234 used to use hex encoding, and still does in its examples. This lead to, for example, Gajim using hex encoding in its code for XEP-0300 representation and Swiften using Base64. What a wonderful world.
> Some people might suggest using heuristics, reading tea leaves, use an algorithm depending on the moon phase, or similar approaches. That does not sound like an appealing solution to a standard.
> My suggestion, as already explained in https://github.com/xsf/xeps/issues/349 , is pumping the version of the namespace, and explicitly stating that the encoding MUST be Base64 as described in 
> https://www.w3.org/TR/xmlschema-2/#base64Binary .
> What do you folks think about this solution? Any suggestions or problems you see with this approach?

It sounds like a sensible enough approach to me, unless we think Gajim is the only deployment using hex.


More information about the Standards mailing list