Re: spec-prod, xmlspec, docbook and Co.

This is all a bit theoretical at the moment, and depends on editorial style 
and available markup.  Arc roles are more flexible precisely because you 
don't have to keep inventing new elements, and yet they give you lots of 
expressiveness.  (But both allow a human to make and apply a judgment, 
which a raw <ref> element doesn't give you.)

E.g., the current style is to use bibrefs to link to bibls, but there's 
been some talk about also/instead linking directly to the cited work.  If 
we adopted a style where you link directly to the other document, and you 
wanted to indicate whether the other document is normatively or 
non-normatively cited, you could use an arc role to indicate this.  In 
fact, using XLink extended links, you could handily link the citation both 
to the bibliography entry and to the external document.

         Eve

At 11:33 AM 10/22/01 -0400, Norman Walsh wrote:
>/ "Eve L. Maler" <eve.maler@sun.com> was heard to say:
>| At 12:05 PM 10/21/01 -0400, Charles McCathieNevile wrote:
>| >well, if we added some RDF into the document...
>|
>| Right -- either RDF, or XLink arc roles that are harvestable as RDF if
>| you wish...  My point was that you can't tell by inspection what the
>| semantic is; a human has to decide and record it.
>
>Well, that may be true in the general case, but an xref to a bibl is
>probably a bibref and an xref to a div is probably a specref, etc.
>
>Individual elements work fine when there are only a few, but how many do 
>you want?
>
>   specref
>   bibref
>   glossref
>   tableref
>   figureref
>   prodref
>   protoref
>   funcref
>   ...

--
Eve Maler                                    +1 781 442 3190
Sun Microsystems XML Technology Center   eve.maler @ sun.com

Received on Monday, 22 October 2001 11:54:00 UTC