Re: sXBL: Extensibility and Versioning

On Saturday, October 23, 2004, 6:39:50 PM, Ian wrote:


IH> On Fri, 22 Oct 2004, Bjoern Hoehrmann wrote:
>> 
>> It seems the latest sXBL Working Draft does not discuss extensibility or
>> versioning of the format. It seems reasonable to expect that the 
>> specification that replaces sXBL makes changes to the language that 
>> would cause sXBL fragments to behave in unexpected ways or fail in old
>> processors that expect such fragments to be sXBL fragments.

IH> sXBL will be an exact subset of XBL2. We've been very careful about that.

Yes. or rather, XBL2 will be a true superset of sXBL, and we have bee
very careful about that. But I agree that this needs to be said
explicitly in the spec, as well as being carefully designed that way.

IH> In addition, for sXBL the SVG error handling rules are used, so 
IH> unrecognised elements cause the processor to abort. So this won't be a
IH> problem. (We may want to also add to sXBL a rule saying that unexpected
IH> attributes are in error too, in case we expect to add attributes that
IH> change the processing model with no new elements.)





-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 Member, W3C Technical Architecture Group

Received on Saturday, 23 October 2004 16:53:27 UTC