It is worth noting that Zotero does not duplicate items when you drag them into a collection. There is still only one item in your library it just appears in multiple locations.

They're not "folders"; they're "collections." And they work as these things increasingly do in contemporary applications (Gmail and iTunes come immediately to mind, but there are many others).

Again: this is a powerful feature; not a bug. So throw out your assumptions and approach this with fresh eyes. Forget about the physical metaphor!

the reason this is annoying is that I have hundreds of unsorted items and I have no easy way to see which ones still need to be sorted.

Do you mean "tagged"? If yes, then instead of just complaining, why not ask if there's an easy way to do find all untagged items? I don't know if there is (since I've never cared about this), but it seems a reasonable question.

alternatively there was talk about a feature that finds "orphaned" items that are in no collection apart from "my library" - that should do the job, too, right? Has this been implemented?

As an workaround for now you could add new items to an "unsorted" collection and remove them from this when they are added to another collection (achieved relatively easily with one hand on the mouse and the other hovering over the delete key).

One option would be to have a (hidden) pref for moving rather than "copying" to be the default behaviour on click and drag, with a modifier key to trigger the non-default behaviour. This would enable those who did want to stick to the "items in one collection only" model to do so easily, without sacrificing the power of the current system for other users.

Personally I find it essential to have items in multiple collections. I have collections for each of my fields of interest as well as collections for each current project (typically essays as I am an undergraduate student). Any particular item might be relevant to more than one field of interest and be used in any number of projects.

I like collections and I like the way they work (though I'm still trying to figure out where they are and are not redundant with tags).

But I think the problem here is that the collection icon is a folder. So (especially new) users naturally expect collections to behave like folders. We all rely, for better and worse, on programmers to use these metaphors consistently. Call 'em collections all you want, but if they look like folders, there'll be a steady stream of new users wondering why the heck they don't act like folders given that almost all instances of folders in the world of computer UIs behave a certain way. The most commonly invoked comparison in these forums is iTunes' playlists, which have a very clear "not-a-folder" icon.

Unfiled and Untagged search options have been discussed and a ticket exists, but it's been awhile and the ticket seems to be low priority. I think given the collaboration functions that are rolling out, with multiple users adding items to a shared library, it will become increasingly important for users to keep track of what items have been processed and what not. Any chance ticket #745 could get a priority bump?

Hi everyone
I'm sure there are many discussions about this

http://forums.zotero.org/discussion/700

but anyhow ...

Just been spending some days tagging and indexing what sure looks now to be like lots of sources and it seems to me that the whole tag cloud in a single collection (or very few collections, if you're a very ecclectic researcher) is a very efficient way to go about classifying your library (ies). Since contrary to some people

http://forums.zotero.org/discussion/4051/remove-all-tags/#Item_0

i wasn't plagued with a mass of irrelevant automatic tags, i started everything from scratch, daming google scholar for not providing me with a tags set or, better yet, a tag choice. Anyhow, It seems to me now from the playing with the partial results i have that the tag search comes closer to what i wanted to get from the outset as a citation manager/research assitant. If at first i tried to have multiple collections, i quickly discarded this organizing mode as being impracticable. Then again, i'm kind of new here and i could yet learn otherwise. My technique was to put a lot of tags. At first you'll notice that the number of tags grows exponentially with every new addition, but at a certain point, there is a marginal decrease of tag augmentation. I suspect that with volume and usage, tag merging and conceptual integration of one's library take place and one's conceptual organization settles in. What i mean to say is that perhaps you'll have 200 tags after 15 texts, but you will probably have something like 350 tags after 100 texts. Since the whole purpose is about regrouping/organizing/linking, your tag cloud will probably shrink slightly with usage, once you've completed the indexing of your collection(s). Hope this helps some people getting started. Feel free also to tell about your efficient organization.

Note to the ones who do everything from scratch: it is a bummer but does give you a hell of a review for your doctoral examination ;).

hey!! Thx there Gstar!

For folks looking for it, the text can be found here

http://www3.interscience.wiley.com/journal/122420507/abstract

The title is
Better to Organize Personal Information by Folders Or by Tags?: The Devil is in the Details

Luis!!!
You just saved me a big BIG headache! The mass tagging feature (basically drag and drop onto tags) and the saved searches feature are my two most recent favorite "hidden" tools in Zotero... thanks for the explanation and hint. Those things should be publicized way more...

Cheers.

Hi Ian

concerning the limited semantics of tags, i believe the problem can be somewhat alleviated by longer tags which would be mindful of a carefully chosen meta-classification scheme. For example, if you were a researcher with a philosophy of science positions of the ilk of scientific/critical realism, you would have such a meta-language for sorting out used ontologies of various discipline on a number of social entities (lets say your a social scientist). For exemple, as with the concept of causation, you could have as tags "causation (category)" refering to all of the possible classes (or extensions) of determination ― that would be the hierarchically superior tag which would be attached to all extensions of the category "causation"; then as a different "level" of tag you would have a further classifications of forms of the causation category, for example "causation (form:constituting)", "causation (form:conjunction_of_events)", "causation (form:cataloguing_powers_propensities)", "causation (form:etc)" refering to specific form of the determinate relationship within the causal category; and you could have also a third "level:, "causation (doctrine:humean)", "causation (doctrine:etc) refering to the specific forms of the category causation that authors have deemed to be non-problematic, essential, unique, etc .... given their own ontological preference. This, by the way, is Bunge's (1961) classification of the causal.

In such a scheme, you would have to tag a lot in order to alleviate the ambiguity of language, and it would definitaly require possession of a prior theoretical understanding of how, in the abstract, classes of significations/existence may be classified appropriately: the problem of a meta-language for ontology (in essence, a layered ontology of ontologies) so it can be somewhat problematic. It implies that the combination of tools like zotero, VUE, protege, seasr, etc might help to inductively think of an appropriate scheme, but there will necessarily have to be a theoretical basis and awareness present in the tagger for him to use his newly presented data.

That said, i could not agree more on the validity of adding ontological membership as a supplement to tags/collections. It would be so sweet also to be able to do this ontological tagging on a piece meal basis, rather than importing ready-made ontological files. That's my biggest problem with VUE right now, as i have the theoretical knowledge to build ontological membership from nodes but i can't seem to find a suitable set fo ready-made relations to go with it. Like you say, "same as", "is subset of", etc... begs many question of the classes of determination (hence my breaking down of causality here), which is ill suited for complex conceptual mapping.

On the collaboration front, i think that tagging IS a major problem, and if i would have some users tagging with me the part of my zotero collection that is public (sadly, no one has yet showed an interest), then i believe that a coding methodology would be in order. Tags used normally are ambiguous and maybe that's a good thing. Maybe it is inescapable. But for collaboration, amibiguity will have to be shared to a certain extent (a shared ambiguity is at least a little less ambiguous, albeit meanings shared are not necessarily true).

So more tags fonctionnalities of the type parent/child would make a difference in the way i can use my zotero data in my opinion. Subjective input, though, will have to be done the old fashion way i'm affraid.

More comments on this subject would be much appreciated. The order of magnitude of my deficiencies is only matched in this regard by that of my interest.

Thanks alexuw - good to get these discussions linked up (:

Nick: I kind of continued this discussion in the second link above ('hierarchical-tags') as there was some reference to semantic web etc.. already there. I think it's important to 'keep it stupidly simple' (to quote the famous engineering slogan KISS). OTOH the utility of richer semantics is also clear, and perhaps even necessary for some. Have a look at my post on the other discussion.

Thank you all for this helpful discussion. I, too, find it useful to have similar discussions linked.

I am not a programmer. Many of the discussions here get way more technical than I need. So, may I ask a couple of really stupid questions?

1. If I have an item listed three times in the main library and I delete two, what does it mean?
1.a. Were there really three separate complete citation entries, each with different notes, attachments, etc?
1. b. If so, does this mean I need to manually review each entry to be sure I am deleting the least useful (to me)?

2. If a duplicate (triplicate) citation is listed in several collections and I delete two of three from the main library, will a citation disappear from a single or all collections?

3. When I import a library from Endnote, should I import it to a collection or the main library?

Many thanks.

Thank you for your response.

Are the duplications in the main library the result of the item being in multiple Endnote libraries? (Ah- is *this* what people have been talking about in some of the other threads about duplications? And there is no way for Zotero to automatically flag the duplications?)

Thank you so much for your response! *Finally* I feel like I understand this well enough to clean up my overpopulated library. Just in time for comps.... ;-)