Re: [Refdb-users] refdbc (0.9.7-pre1) getref -t ris losing info?

[Markus H. <mar...@mh...>]

Doug du Boulay <dou...@gm...> was heard to say:

> Basically, if I feed the following into addref
>
> <ris>
>   <entry type="BOOK" citekey="RefDB_Man">
>     <part>
>       <title type="full">RefDB (0.9.5) Handbook</title>
>       <author>
>         <lastname>Hoenicka</lastname>
>         <firstname>Markus</firstname>
>       </author>
>     </part>
>     <publication>
>       <pubinfo>
>         <pubdate type="primary">
>           <date><year>2005</year></date>
>         </pubdate>
>         <issue>0.9.5</issue>
>         <address></address>
>         <link type="url">http://refdb.sourceforge.net/manual/</link>
>       </pubinfo>
>     </publication>
>   </entry>
> </ris>
>
> then what comes back out with getref, the resultant RISX,
> looks like this:
>
>   <entry type="BOOK" id="2" citekey="RefDB_Man">
>     <publication>
>       <author>
>         <lastname>Hoenicka</lastname>
>         <firstname>Markus</firstname>
>       </author>
>       <pubinfo>
>         <pubdate type="primary">
>           <date><year>2005</year></date>
>         </pubdate>
>         <issue>0.9.5</issue>
>         <link type="url">http://refdb.sourceforge.net/manual/</link>
>       </pubinfo>
>     </publication>
>   </entry>
>
> I seem to have lost the part and its title
> and the author has been shifted from the part, to the publication.
> My guess was that because <publication> is a container for secondary info,
> this implies that a publication/author is a book editor rather
> than the primary book author that I would expect. No?
>
>

risx distinguishes three levels of bibliographic information (just like
librarians do): analytic, monographic, series. The monographic information is
encoded in the publication element. A publication is something that you find as
such on the shelves of a library, e.g. a book. If your work is a chapter in a
book, then you're looking at analytical information (the chapter) along with
monographic information (the book that contains the chapter). A book that
appeared as part of a limited series would contain series information too. In
your example, you'd have to put both the author and the book title into the
publication element, as you have only monographic information available.

The risx DTD unfortunately allows too much freedom to do things the wrong way.
I've started writing a Relax NG schema that is supposed to avoid these
problems. I'll announce it for discussion once I'm done with it.

>
> Also for a particular journal citation/ bibliography formatting style
> that I use (I think it is a Harvard variant), I need to cite myself
> as "du Boulay", but then bibliography entries should be sorted and
> formatted according to
>    Boulay, D., du (2001) ...
> another example
> Jacob Horatio Van Der Marel would be cited as "Van Der Marel",
> but bib formatted and ordered according to
>    Marel, J. H., Van Der (2001)...
>
> So I'm thinking that author lastname|firstname|middlename|suffix may not
> be expressive enough to capture this information properly.
> Or does/can RefDB do some fancy footwork internally with lastname?
>

I'm afraid there is no way to do this currently. We had some discussions in the
past to add a "displayname" element which would record the name as it appears
on the work (this is needed by some citation styles). To solve your problem,
we'd also need some mechanism to separate these honorific-like name parts from
the last name proper. This is doable, but it will add some complexity to the
author element. I don't plan to fix the risx DTD along these lines, but the
schema mentioned above might be a good place to experiment with this.

regards,
Markus

-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de