DR 09-0172 ? WML: Filename-to-IRI mapping

MURATA Makoto (FAMILY Given) eb2m-mrt at asahi-net.or.jp
Tue Nov 16 05:53:08 CET 2010


Chris,

> This is one we were almost done with - it covers the napping of filenames to IRIs. 
>We were agreed on the solution to the core DR, but we noticed in Tokyo
>that there was no normative definition anywhere for "DOS filenames". I
>was tasked with investigating.
> 
> I have investigated, and there appears not to be a definition anywhere in Part 4 
>(or Part 1). I propose we fix this by adding the term to "Terms and
>Definitions" in Part 4 (it is used twice in the part). I'm defining it
>using an EBNF.

Your solution looks great to me.  But I have to point out that a
different solution, namely an undefined term "system-specific file path
notation", has been used for three attributes in Part 1.  For example:

	18.13.1  connection (Connection) 
	....
	sourceFile (Source File Name) 

	Path to the text file to use to import external data. Can be
	expressed in URI or system-specific file path notation. 
	 
	[Note: Applications can decide what forms of URI they support,
	and whether system-specific file path notations are supported.
	end note] 
	   
	The possible values for this attribute are defined by the
	ST_Xstring simple type 
	(§22.9.2.19). 

The other attributes are:

@odcFile (Connection File) in 18.13.1  connection (Connection) 
@sourceFile (Source File Name) in 18.13.12  textPr (Text Import Settings)

I do believe that your proposed solution is much nicer.  Why don't we
adopt the same solution for these three attributes?

Hmm, it appears that Part 1 has some examples of the DOS path notation. 
Should we do the same thing for them?

@connection of dbPr 

@w:val of w:query 

@destinationFile of webPublishItem 

@localConnection of olapPr 

linkLocation of 18.17.7.145  HYPERLINK 

@schemaLocation of schema

@odcFile of connection

@sourceFile of textPr

@Target of Relationship

Cheers,
Makoto


More information about the sc34wg4 mailing list