DR-08-0012 Namespace Mapping Table v2

[Rick Jelliffe]

I am probably too late, and I am sure this has been well thrashed out.

Changing the namespaces means that it would be inconsistent for us to 
also make any other changes in order to cover existing OOXML documents' 
syntax better.  In particular, unrestricting the value of  booleans to 
also allow yes|no, one of the changes made at the BRM that effectively 
ruined OOXML Transitional, would be impossible to justify, since that 
change would not make any existing documents valid: they would be 
invalid because of using the old namespace.  In fact, it would mean that 
all discussions on OOXML Transitional could not include discussion of  
validity of existing documents  (i.e. discussion of how to make existing 
Office 2007 documents accord to the standard), but only to describing 
shared functionality.   

Already IS29500 was criticized because it was not a description of any 
significant collection of documents, such as the binary formats, and 
that it was aimed at representing the information in the legacy formats 
not their notation. Changing the namespace of OOXML does exactly the 
same thing to the now large collection of Office 2007 documents.

Changing the namespace has other bad effects:

  * All previous examples published become erroneous
  * Software needs to be rewritten so that every use of the namespace 
must be altered, not just places where there are changes relating to 
different functional requirements.

Tidyness in using a PURL is not an adequate reason, in itself, to cause 
this kind of disruption. No other standard has  changed its namespace 
after being published. But I understand if this is a decision already made.

If the idea is to make a clean break, then the first thing WG4 should do 
is to put out a warning for people not to implement IS28500:2008 since 
their implementation will immediately be obsolete. Or, at least, to 
detail the expected change.

Cheers
Rick Jelliffe