Upgrade StorageSearch Ranking / alternative search tools
I'm wondering what (if anything) people are using to search their Zotero database, beyond the built-in search functionality.
So far, I've found Zotero to be a great too for getting stuff into a database; it allows me to tag, catalog, and note up documents very easily and consistently, particularly for documents which Zotero can extract info from, but even for those documents where information has to be added manually, Zotero is an easy interface to use.
The search feature, however, is pretty basic. First, it doesn't rank documents, just lists them in alphabetical order. This may be OK if there is only one or two documents on a subject, but once you get to the level of hundreds of documents similarly tagged, a search engine without ranking loses its usefulness. I'm guessing this may be more the case in my world (law) than in the academic realm.
I'm guessing this is the kind of functionality that would be achieved by a full text search tool that could read and rank documents based on certain factor, e.g., if search term is in title, tag, author, or abstract, document gets higher ranking than those where the term is only in the full text/attachment.
I'm not familiar with what kinds of tools might be able to interact with a Zotero database to get that kind of functionality. Just wondering if anyone has any thoughts on this. I recognize this is not what Zotero is designed to be, but it is such a good tool to get stuff in, it would be great if there was a similarly well developed tool to get stuff out again.
So far, I've found Zotero to be a great too for getting stuff into a database; it allows me to tag, catalog, and note up documents very easily and consistently, particularly for documents which Zotero can extract info from, but even for those documents where information has to be added manually, Zotero is an easy interface to use.
The search feature, however, is pretty basic. First, it doesn't rank documents, just lists them in alphabetical order. This may be OK if there is only one or two documents on a subject, but once you get to the level of hundreds of documents similarly tagged, a search engine without ranking loses its usefulness. I'm guessing this may be more the case in my world (law) than in the academic realm.
I'm guessing this is the kind of functionality that would be achieved by a full text search tool that could read and rank documents based on certain factor, e.g., if search term is in title, tag, author, or abstract, document gets higher ranking than those where the term is only in the full text/attachment.
I'm not familiar with what kinds of tools might be able to interact with a Zotero database to get that kind of functionality. Just wondering if anyone has any thoughts on this. I recognize this is not what Zotero is designed to be, but it is such a good tool to get stuff in, it would be great if there was a similarly well developed tool to get stuff out again.
Although I haven't reached the number of items where it is an issue I agree that a ranked search algorithm would be a good idea. Brucebest's suggestion in ranking by the field in which the search term appears seems like the simplest to implement, and there could even be a hidden pref to set the rankings if people felt that would be useful.
What I'm looking at is more on the group libraries / shared research side. In the law world, what this means is that if I go off and do some research on a subject, e.g., "What is the legal test before the Human Rights Tribunal of Ontario to grant an interim remedy?", I may start at the Tribunal's decision database, do some searches, and start saving cases into Zotero, then adding notes/tags as I go. Another 30 people in my office may be doing the same with their own research for their own clients. Once there is a certain body of research in there, I would start in Zotero first to see what my colleagues have written on the subject, before I go off to the Tribunal database to reinvent the wheel.
This is where a decent search ranking algorithm becomes really useful; being able to re-use the research already done, when dealing with a large number (e.g., 1000s) of similar documents.
Having a pref to be able to tweak the rankings is a great idea. Other items become important in ranking as well (e.g., date, frequency of search terms). I'm not sure how simple it is, though - from what I understand, there is a certain amount of magic involved in getting a good search algorithm (or so Google would lead us to believe).