Title: Folksonomies: Users Roll Their Own
1Folksonomies Users Roll Their Own
- by
- Tom Zillner
- WiLSWorld
- July 19, 2005
2The slides will be available on the WiLSWorld
2005 websitesoon.
3Folksonomy
4Folksonomy
- A critical mass of people at a single site
- Sharing their tagging
5Tagging
- Attach one-word descriptors to web sites or other
items?tags - Locate these sites or items through a search or
browse of the tags
6Demodel.icio.us
7Del.icio.us
- At heart, Del.icio.us is a social bookmark
manager - But it can be much more if you choose to make it
so - Take note that there is no privacy to speak of
- Thousands of people have decided thats okay
8Del.icio.us
- People leverage others recommendations
- Rich feature set makes it easy to do so
- Subscription
- RSS
9Personal/Social Bookmark Managers
- Use a bookmarklet to capture pages or items
- Description
- Tags
- Often public/private designation
10Tags
- Note that tags can be very bad descriptors
- Single word may not be very revelatory
- Combining multiple words
- eiffeltower
- eiffel_tower
- eiffeltower
- What are the chances when you get a compound
beyond two words that anyone will ever match a
tag?
11Tagging Nothing New Per Se
- Online citation databases
- Keywords accompanying academic papers and their
abstracts - KWIC information retrieval
- Approximated by OPAC keyword searches
12Only librarians like to search, everyone else
likes to find
Roy Tennant
13Only catalogers like to catalog, everyone else
likes to find.
X
14Classification Schemes
- Aristotle-animals
- Library of Alexandria-kinds of writers, literary
forms and other topics - Dewey-created categories of knowledge
- Library of Congress-arose from the workings of
the library
15Taxonomy
Taxonomy (from Greek ta????µ?a (taxinomia) from
the words taxis order and nomos law) may
refer to either a hierarchical classification of
things, or the principles underlying the
classification.
hierarchical
-Wikipedia
16Ontology is Overrated Categories, Links and
Tags
(Shirky Paper)
17Ontology
- The study of entities and their relations
- The question ontology asks is What kinds of
things exist or can exist in the world, and what
manner of relations can those things have with
each other? - Ontological classification organizing a set of
entities (e.g., books) into groups based on their
essences and possible relations.
18Shirkys claims
- Library Its all hierarchical, with an
overarching ontology - But its not really about the concepts, its
about minimizing seek time to the shelvesdont
confuse the container with the thing contained - Books can be about multiple things, but they can
only be in one place - So, for this reason, the scheme is inherently
flawed for web use
19Shirkys claims (cont.)
- In the digital world there is no shelf, so we
dont have to put up with hierarchical
classification schemes - But still, we try
201998
212002
22(No Transcript)
23Hierarchy
24(No Transcript)
25Hierarchy Plus Links
26(No Transcript)
27Hierarchy Plus Lots of Links
28Just the Links
29Yahoo vs. Google
- People chose search over categorization, and
Yahoo sort of threw in the towel
30(No Transcript)
31(No Transcript)
32When Ontological Classification Works Well
(Shirky)
- Domain
- Small corpus
- Formal categories
- Stable entities
- Restricted entities
- Clear edges
33When Ontological Classification Works Well
(Shirky)
- Participants
- Expert Catalogers
- Authoritative source of judgment
- Coordinated users ???
- Expert users
- e.g., DSM-IV (Diagnostic and Statistical Manual
of Mental Disorders)
34When Ontological Classification Doesnt Work Well
(Shirky)
- Domain
- Large corpus
- No formal categories
- Unstable entries
- Unrestricted entries
- No clear edges
35When Ontological Classification Doesnt Work Well
(Shirky)
- Participants
- Uncoordinated users
- Amateur users
- Naïve catalogers
- No authority
- e.g., the Web
36end of Shirky
37But what about tags andfolksonomies?
38Its the third stage of cataloging the online
world
- Experts?Professionals
- Authors (or web-site owners)
- Users?Amateurs (from the Latin to love)
39Experts/Professionals
- Know the most
- Catalog only a small subset of the information
- Arguably catalog what is most important
- Arguably miss what is important in some
fields?some of whats online - Clearly have no time, money, or mission to
catalog lots of useful or cool online stuff
40Authors/Web Site Owners
- Have a vested interest in getting viewers
- META keywords tag in HTML
- Dublin Core
- May take a great deal of care in getting the most
accurate subject terms to attract viewers who
will be interested in their work - On the other hand, may attempt to deceive in
order to attract viewers
41Users
- Are most interested in bookmarking things they
believe are personally important or interesting - Are often willing to share their knowledge
- Are usually naïve in their tagging vocabulary
42The Long Tail and Folksonomies
- In 1906, Italian economist Vilfredo Pareto
created a mathematical formula to describe the
unequal distribution of wealth in his
country?twenty percent of the people owned eighty
percent of the wealth - One of the 80-20 rules 20 percent of the
collection is used 80 percent of the time - There are a whole slew of 80-20 rules in every
field
43The Long Tail (cont.)
Number of Uses of Tag
Tags
Thomas Vander Wal
44The Long Tail (cont.)
- The power of the long tail as applied to tagging
is that - it captures the mainstream as a significant trend
- but also captures unusual and offbeat
nomenclature that a few might find valuable,
i.e., might be able to find when searching
45Amateur Cataloging
46Its bad cataloging, but if users dont do it,
nobody will.
Only catalogers like to catalog, everyone else
likes to find.
47Cataloging
- Theres a lot of stuff online that gets tagged
that has little value beyond the personal. - Theres a little bit of stuff online that has
lots of value often its already widely known to
people it gets tagged, too
48Cataloging (cont.)
- Theres also quite a bit of stuff that has lots
of value to a limited segment of people, and no
value to most people - Amateurs add value by pointing to stuff in their
areas of interest
49Separating the wheat from the chaff
- Evaluating tagged stuff
- The reputation of the taggers
- Gleaned from their selections and the tags they
use - May be well-known within their community of
interest - The number of tags for a particular site
50Hierarchy Plus Lots of Links
51Hierarchy Plus Lots of Links and Tags
- Theres no reason why OPACs cant represent a
hybrid - They already do with direct links to online
databases - SFX and MetaLib, etc.
- But lets envision something more
52Tagging the OPAC The Most Radical Approach
- Provide a Del.icio.us for your catalog that is
linked to it - Allow people to do bookmarking and tagging of any
item or site - Allow them to publish their bookmarks and tags or
keep them private
53Are catalogers an endangered species?
- Theres no substitute for care and quality
- Attention to matching what a thing is about
within an established hierarchy and providing
crosslinks - Plenty of non-electronic resources
- Plenty of electronic resources that require
quality cataloging
54Quality
- Authority control
- Controlled vocabulary
- Overall Quality Assurance and Control
55Invite Users to Join the Effort
- Offer suggested tags for items or sites based on
- artificial intelligence
- tags user just entered
- already-existing tags
- These suggested tags would be from thesauri
constructed by catalogers or other librarian
experts
56What about the erosion of quality with inclusion
of users tags?
- Users like to find
- Can allow patrons to search only the
professionally-created catalog and exclude the
amateur links/tags if they wish - Can clearly label user efforts as
suchdifferentiate the product - Make it clear that the catalog offers more
vetted resources
57An Aside on Privacy
- Permission-based
- If someone freely chooses to give up their
privacy it is not the librarys/librarians job
to act in loco parentis - As an alternative to non-private sharing of
bookmarks and tags, could do so without divulging
identity - Similarly, Most people who checked out this item
and were willing to share their information also
checked out
58Back to the Present
59Other Social Bookmark Sites
60CiteULike
- Targeted at academics to collect citations to
papers and books (can also store full text) - Includes bookmarklet that can automatically fill
in citations from some online databases - Usual tagging options
- Can export citations to BibTeX and EndNote
61(No Transcript)
62furl
- Similar to del.icio.us
- Added attraction saves the web page you
bookmark, so dont have to worry about link rot - Multiple levels of privacy, so you can have
personal pages, shared within a group, or public
63(No Transcript)
64BlinkList
- Yet another bookmark manager
- Claims better features for organizing than
bundles or groups - Allows public and private bookmarks
- Offers autocompletion and suggestions when
attaching tags
65(No Transcript)
66MyWeb
- New (in Beta) from Yahoo!
- Quite similar to del.icio.us
- But, with more privacy
67(No Transcript)
68Technorati
- Not really a bookmark manager
- Search 13.5 million blogs
- Search 1.6 million tags (including del.icio.us)
- Can maintain watch lists
69(No Transcript)
70and many, many more
71and then theres flickr
72flickr
- The best way to store, sort, search and share
your photos online - There are differences between their free and
Pro accounts (24.95/year)?size of storage,
features of photos - Can choose private/public photo options
- You can globally prevent others from downloading
photos
73flickr
- People can comment on your photos
- Organizr can be used to organize your photos
- Browse
- Search
- Create a photoset
- Share photos with your groups
74flickr
- Slideshows
- Creative Commonssix options
- Tagging (of course)
- Easy capture of photos to blogs
- Favorites other peoples photos you bookmark
- Contacts people you want to share photos with
75flickr
- Groups people with like interests
- Can
- Chat
- Instant Message
- Share photos
76flickr
- One of the most popular tags is me (Roy
Tennant) - What is the compelling social interest? (This is
not a rhetorical question!)
77(No Transcript)
78I asked the thinkers on the LITA Top Tech Trends
panel about Folksonomies
- Clifford Lynchreally nothing new
- Karen Schneiderendearing that the civilian
community has discovered metadata. Im waiting
for next year when they discover authority
control. - Roy TennantFolksonomies suck. (Dont hold
back, Roy. Say what you really think.)
79With all due respect...
- ...I think they missed the point
- Folksonomies are a social construct. Its not
just the tags, its the sharing of information
about resources through the tags. Its a
grassroots effort that could only happen through
the Web, and its only happening because of all
that the Web is aboutsize and connectedness and
openness.
(cont.)
80A critical mass has been reached on some services
that brings the Power Curve/Long Tail into play
81Another pundit
- Joan Frye Williams earlier in the session
mentioned a mass collaboration phenomenon blogs,
wikis, the Dean campaign, e-bay - A few basic ground rules
- Large groups of people
- Little intermediation
- They really like it
82Joan Frye Williams (cont.)
- The environment itself is becoming the engine of
productivity - The network is becoming the new locus of trust,
believing this network of strangers rather than
big organizations, e.g., blogs breaking big new
stories - People arent trusting institutions, e.g.,
libraries
X
83Joan Frye Williams(cont.)
- Librarians job includes creating mass
environments carrying whole soup of patron
information?comments, opinions of peers - Libraries are perceived as not providing
information of personal value - BUT, can become even more trusted and central, by
supporting personal information use and
manipulation
84Folksonomies are a part of that
although the panelists didnt seem to think so.
85Scenario 1Shadow CatalogTagging the
Collectionand the Web
86Shadow Catalog
- Already talked about the creation side of this
- Heres a bit about the user side
87(No Transcript)
88(No Transcript)
89(No Transcript)
90- Tagging can go along with other pieces of the
social soup - Reader reviews
- Most people who read this also read
91Scenario 2Library Supports Folksonomies
92Library Supports Folksonomies
- Workshops on Web services, e.g., Del.icio.us
- Workshops on tagging techniques and vocabulary
- Construction of online thesauri or (better)
pointers to thesauri crafted by other librarians
and interested users
93Scenario 3Toolmaking and Tool Distribution
94Toolmaking
- Thesauri construction
- APIs (Application Programming Interfaces) exist
for most services, including Del.icio.us, Flickr,
MyWeb - An example of a possible tool is an aggregator
which would bring together tags from the
disparate services in a single user
interface?aggregating personal information spaces
95Tool Distribution
- Libraries usually do not have the resources
(people, time, mission) to produce tools - But libraries can point to useful tools produced
by others selection, a task that librarians do
very well - There are lots of tools being produced that deal
with tags and folksonomies
96Scenario 4Services Enable Tagging of OPAC
Records
97Services Tagging the OPAC
- Individual users would find an OPAC record for an
item of interest - They would click on a del.icio.us or other
services bookmarklet and create tags - This would probably require some bookmarklet
modification?requires buy in by proprietors of
the service
98The bottom line
- Folksonomies and tagging are here to stay
- Things that would never have been cataloged will
now be cataloged - Folksonomies represent an opportunity, not a
threat - As we embrace our customers, we must also address
their total personal information needs
99- Tom Zillner
- WiLS
- tzillner_at_wils.wisc.edu