[referencer] some author related issues, remarks, and ideas.

John Spray jcspray at icculus.org
Mon Mar 17 18:54:35 EDT 2008


On Sat, 2008-03-15 at 13:56 +0100, Christian Glahn wrote:
> One (the first) thing that feels strange for my personal scientific
> background (which is educational technology research) is that the first
> names of the authors are shortened in a strange way (only first letters,
> not dots between the names). 

The shortening is entirely a function of whatever database referencer's
getting the information from -- referencer just takes what it is given.

> In my department we use APA style for referencing, which requires the
> dots after each initial of an author. This is not so much of a problem,
> but SOME of the journals in which we are publishing, even like to have
> the complete first name of the authors in the references.

Clearly, the ideal thing would be to have the full first name in the
database, and shorten it when formatting the document.  Unfortunately,
many databases do only store initials.

> I double checked and I found that Crossref for example provides the full
> name of a contributor. So when I resolve the metadata via Crossref
> referencer removes some information. I wonder why this is done?

Referencer doesn't purposely remove any information, although the
crossref interface has a tendency to leave things out.  Here's an
example of crossref output:
        <contributor first-author="true">
                <given_name>J.</given_name>
                <surname>Keller</surname>
        </contributor>

So it wasn't me, I swear!

> A second thing that I came across was that when I resolved metadata via
> Crossref, referencer overwrites manually entered information. This is
> was a bit frustrating because for me because I entered some complete
> author records in 1.0.4 and when I toyed around with updating the
> records via Crossref in 1.1.1, I suddenly had fewer authors in the
> references. This happens because Crossref provides only incomplete
> datasets for the contributors for some journals. 
>
>
> In such cases it would be cool, if referencer would detect if there is
> already some metadata for a document, and smartly extends the meta data.
> Such smart expansion could also be very useful in the future because
> many journals start having "online first" publications, where the
> meta-data remains incomplete for the online only period (sometimes for
> more than two years). 
> 
> I think in these cases reference should update the records instead of
> replacing the complete meta data.

In general I simply wouldn't recommend doing "Get metadata" on documents
you already have the metadata for.  Although it would perhaps be nice to
have a special mode for going over the library filling out only empty
fields.

https://bugs.launchpad.net/referencer/+bug/192175

A boolean "protect" state for each field would probably do the job,
where "protect" got set when the user edited the field.

> The third thing that I was missing a way of accessing/filtering the
> documents not only via tags but also via author names. This would be a
> really useful feature for people who work like I do: "oh yeah, there was
> this article published by Donald Norman. What was its exact title
> again ...". 

Personally, I tend to do this kind of thing by putting "Norman" in the
search box.  If I can't remember how to spell the author's name then I
would sort the documents by author.

A more advanced UI to browse by values of fields as well as by tags
would be nice.  This is one of those awkward "patches are welcome"
moments.

John




More information about the referencer mailing list