At 09:20 AM 7/3/2006, it was written:
On 7/3/06, Chad Chelius <cchelius@xxxxxxxxxxxxxxx> wrote:
David,
I have a similar situation. The people that I am sending XML files
to, require that entities be used for certain characters.

It shouldn't matter to them whether the characters are represented on
disk as entity references or as actual characters - once the XML has
been parsed the result is the same.

Two occasions when someone might specify this is when they are parsing
the XML with some kind of custom parser (or perhaps a custom entity
resolver), or they are reading the file in the incorrect encoding and
seeing odd characters in the output - using entities would bypass the
encoding issue.

The "custom parser" might indeed be wetware engines, some of them 
running very old operating systems, difficult to maintain and even 
more difficult to upgrade. For whatever reason, these components 
frequently expect that data be expressed in idiosyncratic forms, such 
as "& copy;" for "&# 162;", more amenable to their peculiar 
processing algorithms. Yet because these "human users" are often 
capable of performing operations beyond the capabilities of even the 
most advanced digital applications, they are often regarded as 
indispensable parts of the system, and interfaces are designed around them.



If the target system doesn't present this complication then there's a 
lot less reason to output entities. And even if it does ... Unicode 
and good editing software were supposed to do away with such quaint 
ideas as having people inspect and modify source code in raw form. Ha.



Regards,
Wendell