[Refdb-users] xnotes

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

Bruce D'Arcus writes:
 > Has anyone out there looked at the new notes functionality that Markus 
 > has added?  If not please do.  I'm curious if anyone has any feedback.
 > 

I second that. I know it's just a prerelease, but I thought this
functionality might cause a few people to give it a try.

 > The idea came from me, and is tied to my interest in xml and 
 > bibliographic annotation.  So, Markus, has come up with the basic 
 > structure, while my focus has been more on the micro-markup.  We thus 
 > currently have two somewhat different formats; one written as a DTD, 
 > the other as a Relax NG schema.
 > 
 > Most of the differences are minor, but I still wonder what people think.
 > 
 > Here's the example in 0.9.4-pre2:
 > 
 > <xnoteset>
 > 	<xnote id="1" key="firstnote">
 >    		<title>myfirstnote</title>
 >    		<date><year>2003</year><month>10</month><day>12</day></date>
 >    	<note>the note proper</note>
 >    	<keyword>biochemistry</keyword>
 >    	<keyword>enzymes</keyword>
 >    	<link type="reference" target="WANG2002"/>
 >    	<link type="reference" target="Phadke1994"/>
 >    	<link type="author" target="Walsh,N."/>
 >    	<link type="journalabbrev" target="Biochem.Pharmacol."/>
 > 	</xnote>
 > </xnoteset>
 > 
 > I have argued a few things:
 > 
 > 1)  I don't like all the markup for dates, since this is just a 
 > datestamp that can be automated.
 > 

I've noted previously that the date element is a fallback that will
probably not be used very often. The current CVS version of RefDB will
insert the current date if no date is specified which is most likely
what you need anyway.

 > 2)  I prefer changing "author" to "name" to allow linking to non-author 
 > names.
 > 

I'm reluctant to change this now as "name" has a somewhat different
meaning in risx.dtd.

 > 3)  The bigger conceptual difference is that I was thinking of user 
 > information as tied to the note proper, and not the xnote.  Whether I 
 > am right or not depends entirely on how people will use the format.  In 
 > my vision, a group could define an xnote with a topic title, and then 
 > individuals could add their own notes within that.
 > 

I've told you previously that this is not going to work. An xnote
element contains exactly one note. The note is the text that you want
to attach to something in the database. The xnote element contains
this note plus the information that you need to manage this note.

 > The other issue is that whatever we settle on ought to be able to 
 > standalone, or to be embedded elsewhere.  In my case, I embed the 
 > content in MODS (in its "extension" element).  Here's an example:
 > 
 >      <extension>
 >        <xnote xmlns="http://refdb.sourceforge.net/xnotes-ns">
 > 	<note user="darcusb" date="2003-12-11">
 > 	   <p>The note content</p>
 > 	</note>
 >        </xnote>
 >      </extension>
 > 
 > Part of why I like the note-level user info is because it's easier to 
 > manage when embedding elsewhere (I find it awkward to a) include 
 > xnoteset in this context and b) declare the namespace in the same 
 > element as the user).
 > 

You don't need xnoteset in this context. The user information is
attached to the xnote element, not to the xnoteset element. The
xnoteset element is a wrapper to allow several xnote elements in a
single file that uses the xnote.dtd. If you embed the xnote
functionality in a different DTD/Schema, all you need to do is make
sure that more than one xnote is allowed.

 > The datestamp is validated with schema datatyping, and it along with 
 > the user is auto-inserted with the macro templates I've written.
 > 

This is great as long as you use Schema and a Schema-aware editor.

regards,
Markus

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