Problem with Harvard ambiguous entires

lmb

Hi,

I'm using 2.0b3 with OpenOffice and the harvard styles. I'm not sure whether I'm facing a problem with Zotero or with the style, but from the symptoms, I think it might be the core engine:

Here's a sample bibliography generated from my 3 entries:

Flick, U. (2008) Design und Prozess qualitativer Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 252-264.

Flick, U. (2009) Qualitative Sozialforschung 2nd ed., Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag.

Flick, U. (2008) Triangulation in der qualitativen Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 309-318.


Note how Uwe Flick's two entries for 2008 are separated by the 2009 one! For some reason, Zotero doesn't merge them, but seems to count the third one as quite different - sorting it as if it was by a separate author. Maybe the three editors are causing this?

If I change the book sections to "Flick, U." instead - i.e., shortening the first name only, no other change, I get this:

Flick, U. (2008a) Design und Prozess qualitativer Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 252-264.

Flick, U. (2008b) Triangulation in der qualitativen Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 309-318.

Flick, U. (2009) Qualitative Sozialforschung 2nd ed., Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag.


Changing the name of the 2009 author has the same effect, but results in that one being sorted first.) Because the style shortens the first name anyway, this change is invisible in the final output.

So I have a work-around, but I'm not quite sure what is causing this. I did export the database and validate that everything is identical and spelled properly, but couldn't find anything wrong.

Please let me know if you need anything further to debug this.

Regards,
Lars

fbennett

I'm sorry that I can't help with the sorting machinery, but I'm curious about one thing that might be relevant. Which of the Harvard styles produced this result? It looks like it might be Harvard 7de, but that's coded to use full names (not initials), in which case it would be odd that it's initializing the names in the first place.

lmb

Sorry for not being precise: I use "Harvard Reference Format 3 dev". (I modified it to include numbers around the year as required by my university (U.o. Liverpool), but nothing else).

The original style produces:

Flick, U 2008, “Design und Prozess qualitativer Forschung,” in U Flick, E von Kardorff, & I Steinke (eds), Qualitative Forschung, Rowohlt Taschenbuch Verlag, Reinbek bei Hamburg, Germany, pp. 252-264.

Flick, U 2009, Qualitative Sozialforschung 2nd ed., Rowohlt Taschenbuch Verlag, Reinbek bei Hamburg, Germany.

Flick, U 2008, “Triangulation in der qualitativen Forschung,” in U Flick, E von Kardorff, & I Steinke (eds), Qualitative Forschung, Rowohlt Taschenbuch Verlag, Reinbek bei Hamburg, Germany, pp. 309-318.


The style you suggested produces:

Flick, Uwe; Flick, Uwe; von Kardorff, Ernst; Steinke, Ines (eds.) (2008): „Design und Prozess qualitativer Forschung“. In: Qualitative Forschung. 6. Aufl. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag pp. 252-264.

Flick, Uwe (2009): Qualitative Sozialforschung. 2. Aufl. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag.

Flick, Uwe; Flick, Uwe; von Kardorff, Ernst; Steinke, Ines (eds.) (2008): „Triangulation in der qualitativen Forschung“. In: Qualitative Forschung. 6. Aufl. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag pp. 309-318.

So exhibits the same issue.


Thanks,
Lars

fbennett

Ah, I see what the deal is. The year is not included in the sort keys for either style; it's sorting on the author's name and initial (last name first) as primary key, and the title as secondary. If you want to change the sort, you should be able to modify the style. If you search down to the <bibliography> tag in the style file, then find the <sort> tag, there will be two key elements inside the sort span. Add the year with <key macro="year-date"/> in the position where you want it to take effect.

lmb

Not quite. If I add the line you suggest into the middle (as I'd expect - sort by author, than by year, than by title), the result is exactly the same.

However, if I remove the title reference, I get this:

Flick, U. (2008a) Triangulation in der qualitativen Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 309-318.

Flick, U. (2008b) Design und Prozess qualitativer Forschung. In U. Flick, E. von Kardorff, & I. Steinke, eds. Qualitative Forschung. Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag, pp. 252-264.

Flick, U. (2009) Qualitative Sozialforschung 2nd ed., Reinbek bei Hamburg, Germany: Rowohlt Taschenbuch Verlag.

Now that I can live with, but in theory wouldn't mind sorting by title too (ie, I'd expect 2008a and 2008b to be in the opposite order).

I couldn't find any of the standard styles using 3 sort keys, so maybe Zotero doesn't support >2 keys?

But thanks, this seems to address my issue sufficiently well!

lmb

Oh: and in the OO extension, the sorting is the opposite way around - 2009 comes before 2008a and 2008b :-/ Is that a problem with Zotero itself or the OpenOffice extension?

Rintze

Don't know if it helps, but you can't actually sort by the year variable, as items generally have multiple dates (date of access, date of issue). You can use the date of issue variable:
<sort>
<key macro="author"/>
<key variable="issued"/>
</sort>

fbennett

@Imb,

What Rintze said -- I didn't catch that you can use the "issued" variable directly as a sort key. Your sample test might be doing roughly the right thing because the order of the references in the document is fortuitously more or less correct, so it's worth adding the code he suggests, to see if you can get a more solid result.

lmb

@Rintze: sorting by the year-date macro effectively does sort by "issued", as the macro fbennett suggested is just a wrapper around that. So that doesn't address this particular issue, but thanks for the pointer.

Seems like sorting by three keys doesn't quite work.