[uf-discuss] Citation: next steps?

[Michael McCracken]

On 9/22/06, Bruce D'Arcus <bdarcus.lists at gmail.com> wrote:
> On 9/21/06, Michael McCracken <michael.mccracken at gmail.com> wrote:
>
> > I think the ROLE attribute of vcard means something different than
> > what you guys were describing with a role.
> >
> > You were saying role = "what this person's relationship to the cited work is"
>
> Right. The vCard role specifies a relation between an agent and an organization.
>
> Contribution roles specify a relation between an agent and a cited work.
>
> So:
>
> book --> creator --> Agent --> role ---> Organization
>
> Let's put it this way:
>
> In a relational database, you'd have three tables: resources,
> contributions, agents. Role is a column in the second.
>
> But I think for a uF, the easiest thing to do is something like:
>
> <span class="contributor">
>    <span class="fn vcard">Jane Doe</span>
>    <span class="role">translator</span>
> </span>
>
> Alternately:
>
> <span class="contributor">
>    <span class="vcard">
>       <span class="fn">Jane Doe</span>
>    </span>
>    <span class="role">translator</span>
> </span>

That is the most straightforward way, yes. The problem I have with it
is the repeated role term will be displayed for every contributor, and
will likely end up being more hidden data. For instance, nobody wants
display like this:

George  Almasi (author), Gyan Bhanot (author), Dong Chen (author),
Maria Eleftheriou (author), Blake Fitch (author), Alan Gara (author),
Robert Germain (author), John Gunnels (author), Manish Gupta (author),
Philip Heidelberger (author), Michael Pitman (author), Aleksandr
Rayshubskiy (author), James Sexton (author), Frank Suits (author),
Pavlos Vranas (author), Robert Walkup (author), Chris Ward (author),
Yuriy Zhestkov (author), Alessandro Curioni (author), Wanda Andreoni
(author), Charles Archer (author), Jose Moreira (author), Richard Loft
(author), Henry Tufo (author), Theron Voran (author), and Katherine
Riley. Early experience with scientific applications on the Blue
Gene/L supercomputer. In Jose C. Cunha (editor) and Pedro D. Medeiros
(editor), Euro-Par 2005 Parallel Processing: 11th International
Euro-Par Conference, Lisbon, Portugal, August 30 - September 2, 2005.
Proceedings, volume 3648 of Lecture Notes in Computer Science, page
560. Springer-Verlag, 2005. (doi:10.1007/11549468_63)

It's a design guideline that I've heard often on this list to avoid
hiding information, since we're desiging for humans first. However,
this exercise has given me another idea.

What about an extra container around all contributors with a similar
role, like this:

<span class="citation">
  <span class="contributors">
    <span class="role">author</span>
    <span class="vcard">Almasi</span>...etc
  </span>
  <span class="title">Early experience...</span>
  <span class="citation container">
    <span class="contributors">
      <span class="role">editor</span>
      <span class="vcard">Cunha</span>...
    </span>
    <span class="title">Euro-Par...</span>
  </span>
</span>

It's a little more complicated but repeats less, and I think
displaying "author" once is reasonable. Hiding is always an option,
but IMO, we should try to make it possible to avoid.

I haven't thought this totally through, but I'd be interested to hear pros/cons.

-mike

-- 
Michael McCracken
UCSD CSE PhD Candidate
research: http://www.cse.ucsd.edu/~mmccrack/
misc: http://michael-mccracken.net/wp/