RW: Re: support for character entities too expensive for low-cost printers (PR#6776)

Forwarded response for issue 6776:

Subject: Re: support for character entities too expensive for low-cost
printers (PR#6776)
Date: Tue Oct  7 15:23:30 2003

Henri Sivonen wrote:
> The problem is that implementing such data set without reading the DTD 
> would mean that the parser would not be a XML processor as defined in 
> the XML spec. Using a modified parser would break one of XML's 
> benefits: the ability to use a ready-made off-the-shelf parser whose 
> functionality is well defined. 

An XHTML-Print processor is only required to deal with XHTML-Print documents

> Also, having such almost-XML processors 
> around could cause interoperability problems, since different parsers 
> would have different idea of what the pre-defined entities were and, 
> therefore, what entity references rendered a document not well-formed.
> 
The pre-defined entities that an XHTML-Print processor must support is
well-defined. These entities are specified in the XHTML-Print specification
in
[1]. No other entities are part of XHTML-Print and users do not have a means
to
create new entities. Therefore, a confroming printer need only implement
means
to recognize the set of pre-defined entities and replace them with required
Unicode code points. It is then up to the implementation of a conforming
printer
on how best to process the pre-defined set of entities.  

Some implementations have done this via a data table that is compiled into
the
code, thereby relieving the printer of the need to redundently access the
same
information from the DTD for each XHTML-Print document.

However, the specification does not constrain how a confroming printer
should
provide support for the set of pre-defined entities.


Jim Bigelow
Editor

Received on Wednesday, 8 October 2003 10:34:32 UTC