Date:         Mon, 13 Dec 2004 06:28:06 -0800
Reply-To:     Metadata Object Description Schema List <[log in to unmask]>
Sender:       Metadata Object Description Schema List <[log in to unmask]>
From:         Bruce D'Arcus <[log in to unmask]>
Subject:      Re: PDF Metadata
Comments: To: Metadata Object Description Schema List <[log in to unmask]>
In-Reply-To:  <[log in to unmask]>
Content-Disposition: inline
Content-Type: text/plain; charset="iso-8859-1"

On Mon, 13 Dec 2004 09:51:29 -0000, "Hammond, Tony"
<[log in to unmask]> said:

> I too would be very interested in this. It seems at first glance that NPG
> (and other science publisher) RSS 1.0 feeds (being RDF) could be mapped
> to MODS and then to MARC XML.

Why MARC XML?

In any case, I just looked at PRISM again and there's one problem from
my standpoint as someone interested in metadata rich enough to be
suitable for citation formatting.  For that, it's essential to be able
to clearly identify and extract different name parts (family, given,
initials, etc) and titles/subtitles.

Titles aren't too difficult:

<xsl:template match="dc:title">
  <titleInfo>
    <xsl:choose>
      <xsl:when test="contains(., ':')">
        <title>
          <xsl:value-of select="substring-before(., ': ')"/>
        </title>
        <subTitle>
          <xsl:value-of select="substring-after(., ': ')"/>
        </subTitle>
      </xsl:when>
      <xsl:otherwise>
        <title>
          <xsl:value-of select="."/>
        </title>
      </xsl:otherwise>
    </xsl:choose>
  </titleInfo>
</xsl:template>

However, I don't know how to reliably parse this essential piece of
information:

    <dc:creator>Naoyuki Nakagawa</dc:creator>
    <dc:creator>Akira Ohtomo</dc:creator>
    <dc:creator>John L Grazul</dc:creator>
    <dc:creator>Harold Y Hwang</dc:creator>

Tony, I see your name on the PRISM spec.  Is there not something you
guys can do to address the name issue?  PRISM is 95% of the way towards
being suitable for citation processing.  Why not make it 100%?

(Aside: yes, I know this gets back to the data problem, but if we don't
provide a framework for improved metadata, users will endure perpetual
pain.)

Also, I think it's important to be able to include classification
information in the records; what you call prism:category.  Again, it's
essential for citation processing.  I'd almost say it should be
required.

In my XSLT stylesheets, I've taken to using a multi-level classification
scheme based both on the structure, as well as the genre terms within
that structure.  So, I have classes like "part-InSerial", and then
within that get reference types like "article-academic_journal" and
"article-newspaper".

Bruce