[xml-proc-profiles] - More minimal profiles

I'm encouraged by this spec, but was somewhat surprised at the base level
"minimum" profile.

I suggest there are significant use cases for much more basic profiles.

 

Use cases involve 

* Processor / memory restricted devices & programming languages

* Simple data description languages aimed at compatibility with bindings to
common programming languages

 

Providing a very low end profile could greatly benefit the mobile and
embedded device space, as well as enterprise data interchange between
divergent systems that can't or don't want to implement even the "minimum"
XML parser proposed.

I suggest a low end profile could be implemented with an order of magnitude
less code and still fulfill a large class of needs,

while still being compatible with full  processors.

 

For these I propose much more minimum profiles defined.

 

For example:

 

Bare Bones Profile

 

* No internal DTD subset processing

* No Attributes

* No mixed content  (only blank space allowed as non-leaf element
contestant.

 

 

No Namespace Aware Profile

 

* Namespace declarations entirely ignored, presented simply as attributes

* No xml:id or special treatment of xml: attributes

 

 

If this is of interest to the group, I would be willing to work out more
detailed proposals.  

If the entire concept is not of interest, then I suggest this spec has much
more limited scope then possible (or intended).  There are a large class of
XML documents, both current and potential, which don't need *most* of the
features of the "minimal" profile.   This spec has the chance to open those
uses to the "XML World", where they would otherwise lean towards
non-standards or simpler (but less supported) standards such as JSON.

 

 

 

----------------------------------------

David A. Lee

 <mailto:dlee@calldei.com> dlee@calldei.com

 <http://www.xmlsh.org> http://www.xmlsh.org