Draconian: Re: Determination of Encoding

 
> From: Gavin Nicol <gtn@eps.inso.com>
 
> Precisely why I way that we must rely on HTTP header. I'm starting to
> think that Rick's proposal of requiring servers to remove the PI
> is a good idea.

The idea was that there should only be one, but always one, header with the encoding data.

So a server uses and strips the encoding psuedo-attribute from the PI when reading and XML file from filesystem, but generates the
MIME header.  A client uses (and strips) the MIME header, and generates the PI when writing to a file.


I don't care.  As long as data is adequately labelled at everystage, that is the main thing. Given the Draconian thrust of XML so
far, I guess another proposal of mine (that someone agreed with) to use a header mismatch as an indication of a system failure
makes sense.

Perhaps we need another psuedo-attribute in the PI:

<?XML encoding="iso8859-1"  error-policy="draconian">
or 
<?XML encoding="iso8859-1"  error-policy="slack">

where the default is "draconian".  Draconian would be appropriate for CDF. Slack would be appropriate for debugging, and for
documents whose owners were happy if they went through dumb proxies that might disturb things.   


Rick Jelliffe

Received on Friday, 20 June 1997 02:00:49 UTC