RE: [xml-dev] "Introducing MicroXML, Part 1: Explore the basicprinciples

[David Lee <dlee@calldei.com>]

> -----Original Message-----
> From: David Carlisle [mailto:davidc@nag.co.uk]
> Sent: Monday, July 16, 2012 8:31 AM
> To: David Lee
> Cc: xml-dev@lists.xml.org
> Subject: Re: [xml-dev] "Introducing MicroXML, Part 1: Explore the basic
> principles of ...
> 
> On 16/07/2012 13:19, David Lee wrote:
> > So far I have tried to define the syntax without actually choosing the
> separator as it is largely arbitrary
> > *as long as it is impossible to show up in the plain text of any serialized
> XDM value*
> >
> > Unfortunately those pesky CDATA sections make the set of charactors fairly
> small.
> 
> But isn't that exactly what the control characters are for, eg
> 
> U+000C FORM FEED
> or
> U+001F INFORMATION SEPARATOR ONE
> 
> But you still have to know the text encoding (unless you fix that it is
> always (say) utf-8). Otherwise any byte could appear anywhere.
> 
> 
> David
> 

I am not sure if that is what control characters are "for" --- they are an ASCII invention and were "for" control of actual hardware and in a very real sense make no sense in an XML world.   But , they can be *used* for this ... and probably best so.    Or maybe the creators of Unicode really did imagine them in a new role in the Unicode world to outlive their old context or why would they bother adding them to Unicode (and why did XML Exclude them ?) They dont have glyphs  which is interesting for Unicode.
I keep looking for some magic string that is more easy to insert in a plain text editor (try getting a 0x1F into a text document with VI or Notepad).
But some things have to give.    I have yet to find ANY character or string which could be in a plain XML document at the beginning or the end and not cause an XML processor to crash (except BOM) so my fantasy of having 
A) sequence of one document represented the same as a single document , AND
B) concatenation of single documents producing sequences  AND 
C) a single document being readable as XML 

is not viable.   But sometimes not all your wishes can come true. I have only found a way of having any 2 of the above but not 3.
( Note: substitute XML with XDM above I am trying to solve sequences of arbitrary XDM as well as XML documents - not fully but at least enough to 
represent their values and types - which means I need to distinguish between documents and elements, strings and numbers etc,
but the concept is still valid with XML documents).


----------------------------------------
David A. Lee
dlee@calldei.com
http://www.xmlsh.org