Sunday, January 31, 2010

HKLA 50th Anniversary Conference: Part 5

Hmm, I saved this as a Draft and forgot to publish it. Here's a belated part-5 from attending the Hong Kong Library Association 50th Anniversary Conference, Nov 2008.

[Continued from Part 4]

"Analysis of social tagging and book cataloging: a case study". Yi-Chen CHEN. Department of Library & Information Science, National Taiwan University.

Yi-Chen Chen - HKLA 50th Anniversary Conference

Her premise for the study: little research has been done to examine how social tagging has been applied to books.

So she looked at items tagged in

Research questions:
  • How can tags be organised to different function types?
  • What kind of tags are used?
  • How can it help the library?

Study involved a random sample of "most often tagged" Fiction & Non-fiction records in

Some findings
  • For Fiction titles, users tend to tag with "Bibliographic Information" (i.e. author, title, publisher)
  • For Non-Fiction titles, the tags tend to be "subjects"
When she did a comparison of the user-created tags and the Library of Congress Subject Headings (LCSH), she found:
  • 90% of tags were not reflected in LCSH (i.e. 90% of the tags were unique)
  • The overlap (between the user-tags and LCSH) was less than 12% overall
  • Tags tend to give more "genre" information, especially for Fiction works
  • In tagging, users tend to describe more character names from the books
  • Tags often had simpler and informal usage on person names, geographical names

My rambling thinking-aloud
I think when "social tagging" or "folksonomy" is mentioned, there will be some librarians who will inevitably pooh-pooh the former and start extolling the virtues of Authority Control exercised by librarians (i.e. LCSH).

That sort of argument -- of which is "better" -- is is irrelevant. It's like asking, "Is it better to search by author or by subject"?

The answer depends on what you prefer, and what you hope to find.

I noted that the areas of overlap between "social tags" and LCSH is only 12%, i.e. there tends to be different set of results retrieved, depending on whether you use tags or LCSH.

That suggests to me that it's really about different ways to find a book, according to users' preferences.

There are pros and cons: those who aren't familiar with LCSH can go by social tags. It is informal (i.e. not constrained by rules) and hence more easily remembered by those not familiar with LCSH. But there tends to be more misses in finding records, since the nature of social tagging there isn't any effective means of standardising the terms used (unlike LCSH).

I know some libraries are developing a sort of "hybrid systems" that marries -- or tries to bridge -- a structured and more consistent system like LCSHs, and a more user-oriented system like social tagging.

We'll see.


  1. LCSH is, so far one of the better systems

  2. hi, i came across your blog thru zarah's blog and was inspired by all librarians with their thoughts and experiences sharing it to the world. Finally, I have the courage to make my own and share my experiences in my own special way. Hope you don't mind me add you in my links.


  3. Not at all, alynn. And as someone told me when I first blogged, "welcome to the blogosphere" :)

  4. LCSH is good as a basis for tags when it is first inventorised in the library.

    For a small library in a school for example where there isn't a full-time librarian, nobody has the time to read through the new book and apply relevant tags. (Assuming tagging system like LCSH already intact)

    For social tagging to happen, it would most probably come from users who have read the book already and it would be their contribution of 'what terms would i type in the search engine if i want to find this book again'.

    That may help in the quality of social tags that are contributed.

    The 'contribution' would also mean that there must also be social 'cleanup' (for lack of better term) which requires readers to trim irrelevant tags or spam tags.

    Without building a pool of dedicated people to contribute and cleanup the social tags, it would be difficult to maintain a well-tagged catalogue.

    As for the use of, our small library used to have help in creating an e-catalogue for our collection. However, it was not maintained because it is double workload to update the catalogue, and when the catalogue is not updated, it became ineffective and eventually a white elephant. What a waste.
    So unless we have a way to facilitate the maintenance of the electronic catalogue at, we will still be closing our library, limiting it to staff only. Which, really, nobody has the time to go in there physically to be overwhelmed by the books without a good way to search the collection.

    We need help, RFID help, scanners, electronic systems. Those books are screaming to be read!!!

    If we cannot even get these books read the first time, social tagging would be a non-existent issue.

    If we don't even have an electronic catalogue, tagging systems whether LCSH or social tagging would be a non-existent issue.

  5. Well, yes you can see a "tag" like LCSH. Although a "tag" is seen more as "uncontrolled vocabulary". So if you use LCSH, you need to adhere to the established conventions. Or else it wouldn't be "LCSH".

    About your cataloguing woes, I'm reminded of what I like to tell friends who have small libraries as you've described -- that a simple spreadsheet program could work just as well as a library circulation system.

  6. It could as that is what I would do, but it wouldn't be online. Physically having to find a book at the end of the room and to come back to the computer to type the data in does not make sense. (We cannot afford a laptop) So people are just more comfortable with pen and paper, tick tick tick is easier.

    Which will then mean whether or not transferring the data to the spreadsheet or online system is duplicate work.

    However, if we have a scanner that is linked, whoa~!!! All done in just beep, beep, beep! as easy as tick, tick, tick!

  7. Heh, you want online then you must be prepared to spend extra time and effort. If it's a small library, I would think it's OK for users to take that extra steps. Using a spreadsheet programme is far from perfect, but I'd think it would be a reasonable compromise, given all the constraints. Also, you can always duplicate the records and/ or make them read-only. So that you never have to expose your working records to users.


Join the conversation. Leave a comment :)