This topic is currently marked as "dormant"—the last message is more than 90 days old. You can revive it by posting a reply.
Of course, if you were talking about getting other people to use your hierarchical tag schema then that's a whole other ball game.
Top tags foo(100), foo:bar(10), foo:boo(30), ...
So hierarchy only matters for counts and may be searching (but there you can already use wild cards). What you lose is use of one char that separates subtags (: in the above example).
I see this situation as analogous to flat vs. hierarchical file names. LT may not want to do this any number of reasons but I want to make sure I convey my idea clearly.
If tags can be tagged, would you get what you want? Tagging tags is more flexible than hierarchical tags since now you can attach multiple tags to tags (instead of just one) and what's more, you can reorganize tags by renaming a meta tag.
It's a tradeoff that different users will value differently.
I am not unhappy with the way tags work now; I'm interested in a simpler way to find the tags I want to use. The tag list now is hard to read and organized either alphabetically or by frequency, so I occasionally end up inadvertently creating alternate tags because I don't take the time to check my tag list. Of course, I can always go back and edit, but it would be very handy to be able to organize my tag page in a way that would work for me.
Anyway, it got me to thinking. I'm not interested in heirarchy for tagging. I try to keep my tagging pretty clean and I use conventions I will remember. But it would be nice, say on the tags page, to pool the tags into categories. It wouldn't change the structure of tags, but instead would give me some flexibility for their use. I could make one category for genres, another for book awards, and a third for locations (where I keep them). Then when I go to my Tags page I can quickly find the tag I want and go to it. I definitely think this idea, which was given above, would give members a lot more flexibility and reliability with their tags.
Also, has there been any discussion on a drop down feature for tags? Sometimes I'd like to see what tags are available to me and a drop down menu would be very useful. I'm sure that would be difficult for some users with thousands of tags to use, but it might be a nice option you could turn on and off. Anyway, them's my two cents.
-- Biology (54), Physics (21), ...
Seaching on Science will get you all Biology and Physics books as well as any other books that are tagged directly or indirectly as Science. And Biology tag in turn can be used to tag other tags if you want to differentiate subcategories of Biology. And of course, you can categorize tags mulitple ways. The system has to track how tags are related and provide some display options but can allow users to tag anything they want.
To BVS & rebeccanyc:
Simply use a controlled vocabulary when deciding what tags to apply, and then make them hierarchical. A hierarchy can be built into a controlled vocabulary, even with tags.
There are broad categories as well as specific categories for this title. A few possible hierarchies are immediately apparent from the tags applied.
'leabharlann leictreonach' places items within a very broad category identifying all texts that I have in various electronic formats. It is a first tier tag in this hierarchy since most of my library is in electronic form. That makes the entire example idiosyncratic since e-texts appear to be the exception rather than the rule here at LT, but idiosyncratic organisation is part of tagging.
'academic journal' makes these entries part of a broad category within the first category, or it can be a first tier category in itself. Since 99% of my academic journal subscriptions are for the electronic forms of those journals, for this example 'academic journal' is a second tier tag in this hierarchy.
'non-fiction' and 'reference' are also in the second tier (logically they would be first tier, but not in this particular example), but allow for overlap in the categories of 'leabharlann leictreonach' --> 'non-fiction' and 'leabharlann leictreonach' --> 'reference' which may exclude academic journals in my collection.
'folkloristics' identifies the a shared scientific discipline in this example for all three second tier category specialisations, making it a third tier category, and the only one within this data set.
Finally, there are the tags that identify which specific subject areas are dealt with in each issue. 'folklore (proverbs)', 'folklore (Russian)', 'folklore (British)', and 'folklore (medicine)' are some examples. These are fourth tier categories.
All of the other tags are tangential from this hierarchy; they are useful, but outside of the main purpose of the tags identified here.
The beauty in tagging, at least for me, is that a nearly infinite number of these hierarchies can be implemented without much thought, and they can be tailored to individual use (like my own) rather than having to learn, say, the entire set of rules used by the Library of Congress when they make their monolithic hierarchies. It is just a matter of working with really broad tag categories at the beginning, then adding increasingly specialised tags as one's collection in LT develops shared traits.
For more information on controlled vocabularies, particularly within information science, the following are decent-to-good basic explanations along with some further reading. Remember to read Wikipedia articles with mild skepticism, due to the nature of wikis.
What do you gain from such herarchies as
that you dont get from taging with these indiividually.
location1 (x), genre1 (x), type1(x), location2 (x), genre2 (x), type2 (x), etc ...
You'd get these user defined groups on your tag page:
location1 (x), location2 (x), etc...
genre1 (x), genre2 (x), etc...
type1 (x), type2 (x), etc...
Also, to make this point clear, the Groups are NOT tags. They just help organize your tags into convenient groups. This can significantly decrease the amount of time it takes to find tags in your page and can reduce confusion that you can get with having only the alphabetical and frequency groups. What you get is another level of flexibility that the users can have to organize their library.
What I would like is something like what natantus #20 says -- a way to organize my own tag page for ease of use.
Note that tags are *not* predefined. You, the user, define your own tags and how they relate.
#20 talks about grouping on the tag page. If I understand it right, the idea seems to be that each group defines a set of a particular type of tags and that one is not likely to search on a group (such as give me *all* books with some genre tag) as the result is not particularly illuminating. I agree that such groups are different from tags. This is a perfectly useful thing to want, but that is not what I was asking for originally!
I guess if you made the groups searchable (ie like tags) then it would fulfill your initial request. I certainly see the utility of what you're trying to get at. I guess I'm afraid this whole thread will get ignored as it smells too much like the heirarchy which LT is trying to avoid. And I desperately want at least some subset of the features we discussed.
I agree I think some way to 'group' the tags in catagories would be great. The Hierarchy part I would use if available but I do just fine without it. Finding tags I have already used and trying to be consistant, can be difficult.
also designing WHAT the groups are should be user definable
An example :
Dragon Riders of Pern
The reason I strongly believe groups -- indeed the whole tag page -- should be user-definable is because we each have different ideas of what groups and tags would be useful for our books, our ways of organizing books, and our views of what's important to us. Thus, although I'm thinking along the lines of readafew, above, my groups would be different.
What is SO difficult about creating at text file (you know, that thing that Notepad opens) where you've put in lists of all the tags you typically use, organized in whatever way floats your boat, and cut-and-paste into L.T. as needed?
I have 3-4 such text files sitting on my Windows desktop, with everything from frequently used bits of HTML (you never know when you're going to want to talk about the BEARS ) to review notes of stuff I'm currently reading (or listening to).
Having a 6k .txt file running isn't going to cramp up anybody's computer, and it's every bit as handy as having it being maintained by L.T.!
By the same token I also have my entire book collection in a program I wrote myself on my home computer. One reason of many I use LibraryThing is I don't NEED to be at MY computer to use it and look things up.
The point of a database is to collect data and make it useful to those using it.
Also, who goes outside? Aren't there bears outside? (see Penny Arcade)
I'd like to be able to see horizontal levels at a glance; so I have a ton of Classics books; I'd like to be able to see at a glance my translations, which are currently split into Translations (Latin) and Translations (Greek). Because if I just tag them Translations, Latin then I get something else.