I18N-ISSUE-373: Changes of deployment [encoding]

I18N-ISSUE-373: Changes of deployment [encoding]


Raised by: Addison Phillips
On product: encoding


Comment from Martin Duerst:

This is a Last Call comment on 

I wonder whether the WG has considered the chances of deployment in 
particular for the actual encoding conversion definitions.

Based on about 20 years of experience, my understanding of why e.g. the 
IANA charset registry does such a bad job, and why e.g. the Unicode 
Consortium never published 'normative' conversion tables, is that 
implementers from libraries to programming languages to databases to 
networking and communication software are extremely conservative with 
changing definitions of encoding conversions. The reason for this is 
that nobody wants things to break that worked before, for whatever 
definition of 'work'.

There is a strong possibility that this extends to browsers, the major 
targets of this spec. Some browser makers also have their encoding 
conversion libraries tied to uses in other software and will therefore 
be even more conservative.

The WG should be well advised to consider the above points, and how they 
affect the chance to pass CR.

Received on Thursday, 10 July 2014 04:04:41 UTC