Re: Schema refactoring

Jirka,

What you suggest makes a lot of sense to me and I'm entirely in favor of this approach. But I suspect Yves and Felix are the ones who will need to approve.

Best,

-Arle

On Sep 7, 2012, at 2:38 PM, Jirka Kosek <jirka@kosek.cz> wrote:

> Hi,
> 
> while I was inspecting datatypes used in schemas I have noticed that
> current approach for schema design taken from ITS 1.0 contains a lot of
> redundancy. For example for WithinText data category we have 5 very
> similar definitions of corresponding attribute:
> 
> [19] 	att.local.no-ns.attribute.withinText	 ::= 	 attribute withinText {
> "yes" | "no" | "nested" }?
> 
> [29] 	att.local.with-ns.attribute.withinText	 ::= 	 attribute
> its:withinText { "yes" | "no" | "nested" }?
> 
> [99] 	withinTextRule.attributes	 ::= 	att.selector.attributes, attribute
> withinText { "yes" | "no" | "nested" }
> 
> [101] 	att.withinText.attribute.withinText	 ::= 	 attribute
> its:withinText { "yes" | "no" | "nested" }?
> 
> [103] 	att.withinText.html5.attribute.its-within-text	 ::= 	 attribute
> its-within-text { "yes" | "no" | "nested" }?
> 
> This is hard to maintain and keep synchronized. I propose to redesign
> whole schema in the following way:
> 
> 1. For each data category create pattern holding datatype which can be
> reused:
> 
> withinText.type = "yes" | "no" | "nested"
> 
> 2. Create pattern holding attribute definition in no namespace (for use
> on rules/span/ruby):
> 
> attribute.withinText.nons = attribute withinText { withinText.type }
> 
> 3. Create pattern holding attribute definition in ITS namespace (for use
> on elements in a host language):
> 
> attribute.withinText = attribute its:withinText { withinText.type }
> 
> 4. Create pattern hodling attribute for use in HTML5:
> 
> attribute.withinText.html = attribute its-within-text { withinText.type }
> 
> 5. Update all references to use new patterns above
> 
> 6. Make sure that resulting generated schema uses its- prefix on all
> patterns in order to prevent clashes with other schemas.
> 
> The main advantages of new approach are:
> 
> - single definition of datatype -- this is important especially for
> newly introduced categories containing long lists of enumerated values
> 
> - shorter schema and easier to maintain schema
> 
> All comments welcomed.
> 
> 					Jirka
> 
> -- 
> ------------------------------------------------------------------
>  Jirka Kosek      e-mail: jirka@kosek.cz      http://xmlguru.cz
> ------------------------------------------------------------------
>       Professional XML consulting and training services
>  DocBook customization, custom XSLT/XSL-FO document processing
> ------------------------------------------------------------------
> OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member
> ------------------------------------------------------------------
>

Received on Friday, 7 September 2012 12:48:18 UTC