[uf-discuss] Citation format straw proposal on the wiki

Michael McCracken michael.mccracken at gmail.com
Wed Mar 29 10:29:41 PST 2006

On 3/29/06, Alf Eaton <lists at hubmed.org> wrote:
> On 28 Mar 2006, at 17:27, Michael McCracken wrote:
> > Hi, so since we've got a large number of examples on the examples
> > page, I think it's time to float an initial version of a citation
> > microformat.
> >
> > I've put a suggestion up on the wiki at
> > http://microformats.org/wiki/citation-brainstorming - I don't claim
> > that it's complete - help fill it in, but keep in mind that we should
> > start out simple. I think I've captured nearly enough to satisfy the
> > 80/20 rule based on the examples we've got, and I think a majority of
> > BibDesk users would be delighted with 50/50!
> >
> > I tried hard to think of it as a microformat first and resist the
> > temptation to include everything I've seen in other bibliographic data
> > formats. I think there's a real danger here of making it complicated
> > enough that it's hard to adopt, possibly more so than with other
> > microformats.
> >
> > I also don't claim to be any good at designing microformats - I just
> > really want to see one, and a bad strawman has to be better than
> > nothing at all. I'm eager to hear where I diverged from the design
> > patterns, exhibited ignorance, or any other criticism.
> Thanks for setting that up Mike, I'd like to see something finalised
> too. I think the main difficulty is defining the aims and uses for a
> citation microformat (and I agree with Ross that linking is probably
> the most important). Is the aim to be able to convert losslessly
> between a full citation format like MODS and microformatted XHTML, or
> does there just need to be enough information to identify the item?

In my opinion, converting losslessly is not the aim. Given a
microformatted citation, I should be able to take one more step and
find enough information to populate a full format, but I don't think
embedding a full bibliograhpic record at a citation point really
follows existing practice. And it might result in a format that is
huge and hard to understand, inhibiting adoption.

Now, that's not to say that we shouldn't support enough information
for almost full bibliographic information for a large portion of
references, just that there needs to be a balance, and I think that
decisions about what info stays and what info could be left out*
should be made based on use cases that are written down for everyone
to mull over.

(* left out or rolled into a more detailed sub format like work-of-art
and cited as 'item', like my suggestion from Breton's comments)

> Either way, here's the kind of thing I had in mind, just for a
> comparison to your straw man...
> <div class="bibliography">
>      <ol class="reference-list">
>         <li class="citation reference book">

Seeing 'book' here, does this imply that every reference type has a
separate specific class?
I think this might be a slippery slope into a huge format, judging by
the multitude of types out there in formats like MARC. If we want to
explicitly encode type, why not something like a type class whose
content can be whatever type you like? eg:
<span class="type" style="display:none">book</span>

Condsidering existing practice, citations don't generally say what
type they are, do they? It's usually implied by the format or kind of
information they provide. It supports my favorite use case, since I
want to know what kind of record to create when harvesting references,
but is it really important, or just an artifact of us being used to
looking at bibtex etc...?

>                 <span class="title">Title of the Article</span>.
>                 <span class="creator name">Doe J</span> and
>                 <span class="creator name">Bloggs F</span>.
>                 <span class="container"> In
>                         <a class="uri" href="urn:isbn:45346327"/>
>                         <span class="title">Title of the Book</span>
>                         (<span class="year">1988</span>);
>                         <span class="editors">
>                                 <span class="editor">John Green</span> and
>                                 <span class="editor">Simon Brown</span> (Eds);
>                         </span>
>                         <span class="publisher">
>                                 <span class="name">ABC Books</span>,
>                                 <span class="location">New York</span>.
>                         </span>
>                 </span>. pp
>                 <span class="pages">
>                         <span class="start">33</span>-
>                         <span class="end">56</span>.
>                 </span>
>         </li>
>      </ol>
> </div>

for editors and pages, you have a container span, but for authors you
don't - what are the containers for?

> I guess there are some things in there that could become vcards, if
> you wanted.

Yeah, it seems like a principle of microformats is to reuse other
microformats when possible, so publisher and any people ought to be
vcards, IMHO.


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

More information about the microformats-discuss mailing list