Re: character sets in HTTP: translation tables

There are several places where character set issues do impact the design of 
HTTP; One example of this is in human readable meta data such as titles and
author names. HTTP-NG defines all human readable headers to be variable length
strings of bytes - I've argued for making all such text be UTF-* encoded 
UCS-2 or UCS-4. 

You're right in that the bitstream sent over the data channels is not
an HTTP issue; however, using UCS-* as the preferred charset for HTML would
probably be a really good idea.

Simon

Received on Thursday, 29 December 1994 14:14:54 UTC