Re: [xml-dev] Proposed requirements on solutions that convertXML-illegal

[Rick Jelliffe <rjelliffe@allette.com.au>]

I think the issue springs from

1. XML is a textual data transmission format that uses Unicode internally and standard encodings externally.

2. In the standard external encodings, those control characters like NUL have reserved meanings and must be stripped out before parsing: flow control and print head control. These are archaic, but thats the standards.  In Unicode internally, the characters are not appropriate because there is no flow control on a string in memory; and because C ALI's use 0x00 as a string terminator.

The way out if it would be for XML to define named character refences for all the control characters. How and whether an implementation handles them is then their problem. Like a limited SDATA...  (XML should also build in all the standard character entities, so they don't need to be loaded each time and so we can say goodbye to DTDs.

Rick

On 28 Apr 2017 06:57, "David Carlisle" <d.p.carlisle@gmail.com> wrote:

On 27 April 2017 at 21:06, Timothy Cook <timothywayne.cook@gmail.com> wrote:

Ah, I see. I thought we were actually trying to solve a problem.  

That's not the way Roger's threads work:-)

David