Re: Determination of Encoding (Re: Invasion of the pseudo-people ...)

>> Internet experts at Keio University (W3C host) and SGML experts 
>> in Japan have discussed the encoding detection issue.  Here is 
>> my (personal) summary of our agreement.
>
>Looks good.  Simple.  Checking heuristics and prefering MIME to PIs 
>in case of disagreement become (formally)  implementors' and applications'
>error handling strategies, and outside the realm of valid XML.

The proposal mentioned nothing at all about error recovery. 
 
>> Other information or huristics such as "Metadata provided 
>> by the native OS file system or by document management 
>> software" (4.3.3, Part 1) should not be used.   
>
>If you say MIME and PI must agree, else it indicates a system error,
>then surely you should also require  MIME and OS metadata must
>agree?

Not necessarily. The file could be transcoded before being transmitted
(quite common in DynaWeb, for example).

Received on Monday, 16 June 1997 08:19:29 UTC