harvard1, document with multiple authors

[Carles Pina]

Hello,

Find attached unit tests.

bugreports_MultipleAuthors4.txt: is passing correctly. I think that
the other ones should generate the same output than
bugreports_MultipleAuthors4.txt but they don't do (et al. is missing,
etc.)

Tested with citeproc-js 1.0.94 and 1.0.87.

Am I correct with the unit tests expected result? I don't see why not.

Regards,

--
Carles Pina | Software Engineer
http://www.mendeley.com/profiles/Carles-Pina/

Mendeley Limited | London, UK | www.mendeley.com
Registered in England and Wales | Company Number 6419015

[Frank Bennett]

On Jan 11, 8:35 pm, Carles Pina <carles.p...@mendeley.com> wrote:
> Hello,
>
> Find attached unit tests.
>
> bugreports_MultipleAuthors4.txt: is passing correctly. I think that
> the other ones should generate the same output than
> bugreports_MultipleAuthors4.txt but they don't do (et al. is missing,
> etc.)
>
> Tested with citeproc-js 1.0.94 and 1.0.87.
>
> Am I correct with the unit tests expected result? I don't see why not.

The results are intended, although views may differ on whether they
are desirable. Names that have no given name set are interpreted as
institutional names, and trailing personal names (i.e. one or more
names with a given name set, following the last institutional name)
are taken as unaffiliated authors.

MultipleAuthors4 is a normal set of three personal names, so you get
family01 et al.

MultipleAuthors3 is an unaffiliated personal author, with two separate
institutional co-authors, so you get family03 et al. This needs to be
documented, but it is intended; within the limitations of a simple
list of two-field name objects as input, this seems to be the only way
to express the exceptional case of unaffiliated authors. It's a
sufficiently rare case that this slightly odd way of handling it may
be acceptable. That's the thinking, anyway.

MultipleAuthors2 is two personal authors, both co-authoring with an
institution, so you get family01 & family02. The institutional co-
author is left out of the in-text form of the reference. Possibly
this should use et al. (?)

MultipleAuthors1 is three institutional authors. The processor uses
only the first name in the in-text form of the reference, without et
al. Again possibly this should use et al. (?)

The thinking behind ignoring the institutional author in the in-text
form in MultipleAuthors2 and MultipleAuthors1 in in-text references
was to keep things concise. But looking at the tests, it does seem
like including et al. might be a better policy. Shall we go that
route?



>
> Regards,
>
> --
> Carles Pina | Software Engineerhttp://www.mendeley.com/profiles/Carles-Pina/

>
> Mendeley Limited | London, UK |www.mendeley.com
> Registered in England and Wales | Company Number 6419015
>

>  bugreports_MultipleAuthors1.txt
> < 1KViewDownload
>
>  bugreports_MultipleAuthors2.txt
> < 1KViewDownload
>
>  bugreports_MultipleAuthors3.txt
> < 1KViewDownload
>
>  bugreports_MultipleAuthors4.txt
> < 1KViewDownload

[Frank Bennett]

On second thought ...

For MultipleAuthors2, the names would format, in full form, as:
family01 & family02, family03
In a typical cite, this would look like: Smith & Jones, International
Business Machines
So in this case, the current treatment for in-text references is
probably appropriate.

For MultpleAuthors1, the names would format, in full form, as:
family01, family02 and family03
In a typical cite, this might look like: IBM, WTO & UNHCR
So in this particular case, et al. might indeed be appropriate.

[Frank Bennett]

Another second thought ...

Where institutional authors beyond the first are used in the in-text
form, you could get things like this:

(UNHCR and Jones & Smith, WHO, 2000)

These are all edge cases, since institutional authors are relatively
uncommon, and mixed authorship by multiple institutuions is a rarity.
I'm certainly open to suggestions for how to handle in-text references
for this category of material. Current behavior leans heavily toward
brevity; the other extreme would be to stop doing special truncation
of content in these cases, and let everything through on the standard
et al. truncation rules. I'm not sure which form publishers would
generally prefer.