Re: Proposed designation of OASIS pointer type 2

[Tony Graham <tkg@menteith.com>]

At 9 Jul 2001 13:03 -0400, G. Ken Holman wrote:
 > At 01/07/09 11:49 -0400, David_Marston@lotus.com wrote:
 > >I'm working on a better blend of the Catalog of Discretion and the
 > >test catalog, and I now find it necessary to cite the second type of
 > >document pointer. I'd like to call that "OASISptr2" for "OASIS Pointer
 > >Type 2" - any objections?
 > 
 > No objection from me.  Is the need lessened if we adopt the same 
 > specification referencing scheme in both the Discretionary document and the 
 > test cases?  Would the package be better accepted if we were more consistent?

Probably.

It's an accident of history that the two documents use different
identifiers.  I made the original catalog at around the same time that
Ken wrote his script for creating pointers into the spec.  The
catalog's fragment identifiers start at the root because that was easy
to automate.  The identifiers include the URL of the XML document
because I wanted to nail down which version of the spec was being
referenced, but that could be handled almost as easily by putting a
base URL on the catalog's document element.  (Leaving the base URL off
the identifier, on the other hand, makes it harder to reference both
the main spec and any errata document from within the same catalog.)
The identifiers in the catalog don't go down as far as "text()[1]",
etc., because some of the discretionary text spans multiple text nodes
that are separated by cross-reference elements.

The identifiers in the catalog could be re-written in the other style
(and vice-versa), but it's just something that I never had time to do
before I resigned from the committee. (And, no, I don't have time to
do it now.)

Regards,


Tony Graham.