[pmwiki-users] Re: yet another documentation suggestion ...

[V.Krishn]

On Thursday 04 August 2005 09:15, John Rankin wrote:
> On Thursday, 4 August 2005 3:16 PM, V.Krishn <mistyfire at autograf.pl> wrote:
> >> As I understand it, Dublin Core defines /structure/ not content.
> >> I think we could map (:keywords:) to dc:subject -- again pmwiki
> >
> >mmmm.... this would be not quite fair as the scope of these are
> >different.
>
> Yeah, !categories seem more appropriate to me for dc:subject
<dc:subject> -- ignore, or spaced-out version of the title ("SomeWikiPage" -> 
"Some Wiki Page") 
<dc:title> -- the name of the wiki page, without the Wiki name or other stuff
I think here it means :
<dc:title> eg. Main.HomePage
<dc:subject> Either the spaced out version OR text from (:title:)
:-))
>
I think dc:description means (:description:)
coz if there exist a meta for a bookmark this would sound weird-
<dc:description> -- ignore, or use the first paragraph of the wiki page.

Check here:
http://www.w3.org/2003/07/Annotea/BookmarkSchema-20030707
dc:description
    The description of or notes about the bookmark.
> >I also see no harm in author making change to the meta content, remember
> > its "wiki" and authors best know how to describe the page if he/she has
> > full freedom to change content.
>
> Except that pmwiki already looks after some dc elements, like the date
> and the page's URI. I don't think we want writers to change these.
> Those dc elements that pmwiki looks after, the author shouldn't be
> able to change. Those that pmwiki doesn't look after, the author
> as you say knows best. Basically, we have to consider each dc
> element on its own merits and decide whether it's author managed or
> pmwiki managed.
>
> And the metadata don't need to be physically separate. Pmwiki
> already holds various attributes as part of the page; the dc-mes
> would just add to these and be generated as dc xml on demand, in the
> same way that pmwiki generates html of the contents on demand.
>
> >What would be status of these documents:
> >http://www.w3.org/TR/NOTE-MCF-XML-970624/
> >
> >This does not define <link rel="meta" ...> ??
> >http://www.w3.org/TR/REC-html40/types.html#type-links
>
> As I understand it, DC is a particular metadata model, whereas the
> MCF is a /framework/ for implementing metadata models. But I am
> rapidly getting out of my depth.
>
> I'd turn the question on its head: why /would't/ we use DC as a
Nope, I am not againt it. What I am trying to emphasise is that while 
designing it we keep in mind other rdf models too 
Eg:
http://www.w3.org/2003/07/Annotea/BookmarkSchema-20030707
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
         xmlns:dc="http://purl.org/dc/elements/1.1/#"
         xmlns:a="http://www.w3.org/2000/10/annotation-ns#"
         xmlns:b="http://www.w3.org/2002/01/bookmark#"
>
i.e other rdf models, dc, annotea, bookmarks.... hmmmm
--------
And the design should be flexible rather modular enough to adapt itself.
http://www.w3.org/RDF/
http://www.w3.org/2001/Annotea/

> basis for wiki page metadata? I haven't been able to think of
> a good reason. AFAIK, it's the most widely-used standard and
> still leaves the door open for other metadata models to be used
:-)
> if these are more appropriate in any given situation. Best of all,
> there is great documentation about how to encode dc, so it makes
> correct implementation very straightforward.
>
> See http://dublincore.org/documents/dcmes-xml/
>
> Linking to Dublin Core metadata in XML from HTML
>
> Dublin Core encoded in the method described here can be referred to from an
> HTML document and associated with it by means of the HTML <link> element.
> The recommended relation type for this purpose is rel="meta", used like
> this: <link rel="meta" href="mydoc.dcxml" />
>
> where mydoc.dcxml is the URI of the XML document being refered to.
I think its not far dedicated meta servers would be ubiquitous.
Shall we christen it DMS or MFS (meta file servers) from where meta files can 
be picked up or shared. :-)

V.Krishn