Re: [dita] Types of links

[Don Day <donday@donrday.com>]

Eliot is so right about the terminological challenges of maps. They can 
be whatever you want them to be; your definition will depend upon the 
context in which you apply them. In fact, in the original article about 
DITA (http://www.ibm.com/developerworks/library/x-dita1/) we called them 
"delivery contexts" in an effort to not hinder the understanding of 
their use with any single normative example.

My web applications use maps to describe headers for web sites, bookmark 
files, search results, RESTful collections, and more. Excepting the 
processing-specific metadata in the map schema, topicrefs are basically 
a serialization of a directed graph. In that regard, the order and 
nesting of topicrefs can say a great deal about the relationships of the 
resources they cite, particularly when specialization applies patterns 
of semantics and use to the resources (contrast the role of subject 
scheme map with bookmap or a translation manifest map, for example). 
Some of these uses may represent Views (publishing renditions) while 
others may represent Models (subject schemes or site maps) or 
Collections (lists of things selected by query).

The most pure thing you can say about maps is that they represent a data 
model for resource relationships; everything else is a skin applied to 
that model.
--
Don

On 12/3/2015 8:50 AM, Eliot Kimber wrote:
> I think of normal-role topicrefs not in reltables as "organizational" or
> "structural" (to the degree that I've thought about a single term at all).
>
> I've used "navigational" to refer to links within content that take you
> somewhere else (xrefs, reltable-defined links rendered as navigational
> links).
>
> Conrefs are clearly transclusion links in the sense meant by Ted Nelson.
>
> One of the terminological challenges here is that DITA does not limit the
> use of maps to the description of publication structures, even though
> that's the most common case. So we can't make a statement like
> "normal-role topicrefs not in reltables establish the organizational
> structure of a publication" because that's not the *only* thing they can
> do, unless you take "publication" to mean "any set of resources organized
> into a hierarchy for any purpose", which is kind of a pointless definition.
>
> One thing to consider for DITA 2.0 is to enable this codification of map
> == publication by definition a specialization of map that explicitly means
> "publication" in the sense given above. It would be structurally identical
> to base map (that is, it wouldn't impose any additional constraints) but
> it would allow us to then say "in the context of a publication map
> normal-role topicrefs not in relationship tables establish the initial
> organizational structure of the publication. The effective organizational
> structure may be determined by additional processing that reorders or
> augments the base structure, such as generating a glossary based on terms
> referenced from topics or sorting reference entries based on
> locale-specific collation rules."
>
> Cheers,
>
> Eliot
> ----
> Eliot Kimber, Owner
> Contrext, LLC
> http://contrext.com
>
>
>
>
> On 12/2/15, 6:28 PM, "Tom Magliery" <dita@lists.oasis-open.org on behalf
> of tom.magliery@justsystems.com> wrote:
>
> > A brief distraction while we await The Vote.
> >
> > I've never been entirely happy about how I characterize topicrefs when
> > I'm teaching people about link types. I'm happy with "navigational" links
> > (related, xref, et al) and "transclusional" links (image, conref), but
> > not so much with "structural", as I've called topicrefs. A colleague just
> > suggested "compositional", which sounds better right this minute, but
> > that might be only because it's a new name to consider, and because of
> > the neat "...ional" naming parallelism.
> >
> > Do you have another word for them?
> >
> > Or do you consider topicrefs to be in the same category as conref and
> > image links? And in that case, what do you call that category?
> > Transclusional doesn't seem right.
> >
> > mag
>
>
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  Follow this link to all your TCs in OASIS at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php


---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus