Two different styles in one document?

lmb

Hi,

I have fallen positively in love with Zotero, using it on Linux in combination with OpenOffice. The University of Liverpool insists on an almost-like-harvard style, and I have been able to make the tweaks I wanted easily enough.

Almost.

I can't figure out to use two different styles easily - i.e., some instances I have text like "This has been shown to be wonderful (God, 300: 4)", and in other cases, I want "This has been proven correct, but not tested by Knuth (2009: 3)".

In LaTeX, there's two different cite commands for this; in Zotero's OpenOffice plugin, I can use the editor of course, but then have to redo this for some modifications.

Possible, but it'd be more elegant if the styles allowed multiple citation layouts to chose from - then they'd also propagate when I change the template.

Can this be done?

Kind regards,
Lars

noksagt

Have you tried using "suppress author" in the add/edit citation dialog?

lmb

Sure, but then I have to add the author manually, which is not quite what I want either. The parenthesized version is good as the "passive" form, but for active references it doesn't quite work.

Like I said, LaTeX/BiBTeX allow the invocation of \citet and \citep, respectively, which produce one or the other; it'd be nice if Zotero could do the same, but right now, I'm not sure whether CSL even would allow several citation layouts to be described ...?

noksagt

Suppressing authors enables more phrases, though. See
http://forums.zotero.org/discussion/5282

You can enter phrases like "Knuth's research (2008) shows that...."

lmb

Well, yes, thanks for the pointer. I think I'm hitting exactly those limitations and it seems that I'm basically stuck with exactly the same problems as the other LaTeX-converts in that thread ;-)

I think the argument can be summed up as either making it easy to switch styles for a document (which for me is rare), or making it easy to cite in several formats via one style (which for me is common).

I think having to type the names manually is a major hassle.

How I imagine this would work would be a drop-down box in the "Insert citation" dialogue allowing me to choose between the various "citation" definitions in the CSL, which would have an additional name attribute:

<citation name="Harvard with parentheses" ...

<citation name="Harvard without parentheses" ...

It'd default to the most common form, possibly as defined by the CSL too.

If one wanted to get fancy, they could have common identifiers (like "citet" or "citep" ;-) which would ease switching between styles, possibly, as they could then apply the "proper" formatting from the new style.

Suppressing authors strikes me as an awkward work-around.

I understand this may not be priority for the project, and I have no idea how to code firefox extensions or OpenOffice ones. But if someone implements this and points me to their Amazon wishlist, I shall be grateful.

bdarcus

I have no intention of supporting this (though not to say I'll never change my mind), and frankly get a bit tired of the "do it like BibTeX" arguments. I use author-date styles ALL the time, and find the current approach more flexible, and the alternative to offer mostly trivial benefits.

fbennett

Imb,

What you are asking for (masquerading part of a citation as being part of the main text) is possible, but the fact that something is possible doesn't necessarily make it a good idea; the benefits have to outweigh the costs.

Speaking as someone who has recently spent several hundred hours on implementing _part_ of the _existing_ functionality of CSL, it's kind of hard for me to see how typing names into the text of a document you are writing is "a major hassle" that needs to be eliminated. It just doesn't seem like a reasonable priority for a project this complex.

Frank Bennett

CB

fbennett: I haven't followed this whole thread, but writing as someone who uses LaTeX/BibTex for producing citations, I'd say that typing names in would for me not be so much a 'hassle', as error-prone. It seems like just the kind of process that computers are good at automating.

I'd actually like to use something like Zotero for this. I'm a very reluctant user of LaTeX; I'd much prefer to use GUI tools if there was such a thing as a decent structured word processor (unfortunately there isn't that I've found). But I'd be annoyed at having to manually type in-text citations (there are some scarily-named authors out there!). I don't think, contra Bruce's objection, that this is a just a matter of wanting Zotero to do things in the way that existing tools do, so much as wanting the computer to make routine things gradually easier and more accurate.

I'm not dismissing your objections regarding the difficulty; just pointing out that the desire for BibTeX-like abilities isn't as trivial as you're suggesting.

fbennett

CB: That's fair. I didn't mean to be casually dismissive, and I apologize if I came across that way. It's just that the core functionality of CSL (or any other citation engine) is aimed at routine tasks that are more time-consuming and more prone to error. Correcting the spelling of a name is something you would pick up in proofreading, and it can be fixed by an edit at one location. Addition or removal of a single citation can change year-suffixes and the number of printed names for multiple references scattered throughout the manuscript. It's a harder problem to solve, but also provides a greater lift to the editorial chain (to mix metaphores).

With time, this kind of facility is likely to work its way into Zotero and similar tools. There are just other tasks at the moment, some not readily visible on the surface, and some not relevant to all styles of citation, that seem more pressing.

CB

fbennett: no offense taken, and I do see the point you're making. It does all seem quite nightmarishly complex: I'm glad I'm just a user! Now back to finishing typing Czicksentmihalyi ...

lmb

Hi Frank,

as I said, I understand that it may not be a very pressing need from the perspective of the overall project. Some of the LaTeX books do contain a long discussion of the different citation styles, and I definitely see your point about this being a very complex topic!

However, I also think that typing names manually - basically implying the need to cut&paste from the bibliography - is not just error prone but one of those tasks which computers should automate (eventually). It would make the writing process easier for me at least, and the stories from other users who have used BiBTeX (and other tools) seem to lend this some credibility. In the end, I think this is similar to automatic bibliography generation: for sure a missing or badly formatted entry would be caught during the proof-reading process, yet we seem to want to automate these.

I see that it is not high on your list of priorities; that is fair, and I trust there are more urgent and important issues.

But my offer to sponsor someone to implement such a feature via their Amazon wishlist seem fair too ;-)

Thanks for the discussion, at least for now I see that what I want/need is not currently possible, and that I must use some other mechanisms for the time being.

Regards -

bdarcus

@lmh: you say you understand this is a complex topic, but it's clear you do not, nor that you've carefully read the earlier discussion. You simply keep repeating yourself.

bdarcus

@CB: when I use the word "trivial" here, I mean relative to implementation complexity and other tradeoffs. Nobody who wants this has explained how it would work in a dynamic GUI, where users can seamlessly switch between author-date and other (footnote, number, etc) styles.

lmb

@bdarcus: I apologize if you feel this way.

I thought I addressed this point, namely by saying that for my use case, switching seamlessly between so vastly different citation styles simply is not important - getting the one I need to use "just right" (without manually typing/c&p of author names) however is. Options for how this might be implemented in the UI have also been mentioned.

bdarcus

@lmb: On your first point, then that definitely means we won't support it. It is a fundamental design goal of CSL that it be general and flexible enough to support this requirement.

Rather than just discuss this in the abstract, I've created a ticket for this issue over at the xbib project (which develops and maintains CSL). I'm too busy to think in depth about this now, so I need others to do that work. I would really appreciate if everybody that has contributed to discussions on this issue can find time to think hard about how this might actually work, and post their thoughts as a comment on that ticket.

Note: I do NOT want "this is how BIbTeX works" comments; I know how it works. I also don't want "it would be nice if x" comments. Rather, I want specific solutions for CSL and CSL implementers (which are not just Zotero, but include the markdown-based pandoc/citeproc-hs, Mendeley, etc).

lmb

Okay, so the official path for users to make suggestions is to define the implementation, instead of describing their use case? You know, as a software engineer, that strikes me as kind of the wrong way around ;-)

But yes, I got your message loud and clear. Thank you for your time.

bdarcus

No, the use case is known; it's been stated over and over. But a) I am adding to that the requirement to be able to change styles radically without having to modify the in-document citations, and b) I want to think about how to actually do this (rather than just talk about it) ;-)

For background: BibTeX is largely targeted at the sciences. Zotero is largely targeted at the humanities. CSL tries to cover both, which is not always easy.

Also, there is no "official path," but the situation here relates to CSL, and so is (partially) independent of Zotero. If we figure out how to do it in CSL, then I'm sure we can get fbennett to implement it it Zotero ;-)

Simon

I think it would be wrong to characterize Zotero as "largely targeted at the humanities." While Zotero offers a few features that are more useful to those in the humanities, it also offers all or almost all of the features of competing products that are "largely targeted toward the sciences." I am both a neuroscience student and a contributor to Zotero, and it suits my needs better than anything else out there. This particular situation has never troubled me; I've just typed the author's name and then inserted a citation with the suppress author box checked. As such, I don't have much of an opinion. However, if support for a second citation style gets added to CSL, I'd be happy to add it to Zotero.

bdarcus

Yeah, I was in a hurry. "Tagetted at" is the wrong verb. I just meant that different areas have different priorities, and that's hard to account for in code. For example, in my area I often see (and do) stuff like:

Jones argues blah, blah, blah (2004; see also Doe, 2002).

erazlogo

I think it would be wrong to characterize Zotero as "largely targeted at the humanities." While Zotero offers a few features that are more useful to those in the humanities, it also offers all or almost all of the features of competing products that are "largely targeted toward the sciences."

I concur--some changes that would be useful for humanities are not a priority currently. It's still impossible to view research notes organized by tags in Zotero for example (a standard way of organizing archival research), while bibliographic features and Word integration have priority--it is my understanding that those are especially important for the sciences. I would say the sciences are being served pretty well overall.