Slides from the talk I gave at the TCDL 2009 conference on May 27, 2009. Full writeup at http://onebiglibrary.net/story/tcdl-2009-talk-better-living-through-linking.

1. Better Living
through
Linking
Daniel Chudnov - 2009-05-27 - dchud at umich edu
Texas Conference on Digital Libraries 2009

2. MY OPINIONS
DO NOT REFLECT
THOSE OF:
MY EMPLOYER
THE US GOV’T
PRES. BARACK OBAMA
THE DETROIT RED WINGS

3. • doing “web”
better
• linking data
• steps you can
take today

4. 1.
doing “web” better

5. we’ve been building
a web
for a long time

16. Photo cc-sa by osunick on flickr, http://tinyurl.com/oy8qw942

20. which brings us
to about, say,
1996.

21. xml
1997

22. db-backed sites
1998

23. xml-backed sites
1999

24. blog tools
2001

25. better search
2002

26. rss
openurl
oai-pmh
srw/u
opensearch
microformats

27. redesign
redesign
redesign
redesign
redesign
redesign

28. scripts
includes
templates
frameworks
cms
seo

29. User. Generated. Content.
OMG ROFLMAO LOLZ

30. how we use the web
evolves

31. how we build the web
evolves

32. it’s not done yet,
but

33. we are still
building a web

34. here’s what’s next

35. 2.
linking data

36. links matter

37. just ask Google

38. but we knew that

39. for example

48. which, basically,
was this

50. the web
is made of links

51. but wait
there’s more

59. which, basically,
was this

61. but
nobody knows that

62. ...unless they
know to look...

63. ...which is the
problem.

64. more to the point

65. Google won’t find this.

66. http://authorities.loc.gov/cgi-bin/Pwebrecon.cgi?
AuthRecID=2852966&v1=1&HC=3&SEQ=2009052703254
5&PID=XGImKyGgg1Cofw8uO3IFxAvS

67. what’s wrong
with this picture?

68. system info
http://authorities.loc.gov/cgi-bin/Pwebrecon.cgi?
AuthRecID=2852966&v1=1&HC=3&SEQ=2009052703254
5&PID=XGImKyGgg1Cofw8uO3IFxAvS
unknowable
session info info

69. unbookmarkable

70. screenscraping required

71. unscriptable

72. item record authority record

73. not linked

74. a holdings record
doesn’t help
Google

75. but wait, there’s more

85. not connected

87. not findable

88. not relevant

89. ouch.

90. despite our efforts

91. the expense of
digitizing

92. care and feeding
of
cranky coders

93. the overhead
of
authority control

94. we count on
Google to guess
at relevance

95. inside our systems
we tweak
and improve

96. across our systems
we can
only hope

97. time to
do
better

98. enter
Linked Data

99. “Linked Data”
is
data in the web

100. “Linked Data”
is
a movement

101. “Linked Data”
is
dogma

102. “Linked Data”
is
worth a close look

103. so what is it?

104. a way to
link
better

105. a way to
make
concepts explicit

106. a way to
connect
our stuff

107. a way to
connect
these together
item record authority record

108. draw our stuff
deeper
into the web

109. not just
files to download
but part of the web

110. how?

111. use URIs as names for things
use HTTP URIs
provide useful information
include links to other URIs
http://www.w3.org/DesignIssues/LinkedData.html

112. break down concepts
make them “pages”

113. the new
the old

114. bad URL
screenscraping
required
few links out
no links in
the old

115. the new
good URL
alternate
formats
some links out
more links in

116. LCCN Permalink
alternate
formats
more links out
clean URI
http://lccn.loc.gov/84759993

117. discoverable, parseable
alternate formats

118. apply these rules
everywhere

120. rel=alternate
resource URI
rel=contents
(already has)
resource URI
resource URI
resource URI
rel=alternate

121. arrows
become
cross-references

122. why?

123. it becomes
crawl-able
mine-able

124. integrated access
through
smart crawl / index

125. precision improves
with
rich metadata
explicitly linked

126. recall improves
with
explicit links
through authority recs

127. bring
precision / recall
of clunky OPACs
to open web

128. this is
doing web stuff
better

129. just by doing
HTML / HTTP

130. use URIs as names for things
use HTTP URIs
provide useful information
include links to other URIs
http://www.w3.org/DesignIssues/LinkedData.html

131. doesn’t fit the web,
it is the web

132. many examples already

135. LIBRIS
id.loc.gov
chronicling america

136. if i squint just right...

137. the “LOD cloud”
a
“ready reference”
collection

138. make
this
useful
turn this
into
linked data

139. seems obvious,
right?

140. 3.
steps you can take today

141. how to
link your data

142. (or, About All That
Other Stuff You Have)

143. use URIs as names for things
use HTTP URIs
provide useful information
include links to other URIs
http://www.w3.org/DesignIssues/LinkedData.html

144. first:
clean, cool URIs
for data concepts

145. instead of this:
/search?subject=Strayhorn,+Billy

146. do this:
/subject/lccn/n81072976

147. easy to implement
today

148. easy to redirect
tomorrow

149. next:
offer data views

150. use
<a rel=___ href=___>
<link rel=___>

151. offer up
MARC,
or JSON,
or RDF,
or ...

152. next:
relate your data
to established
authorities

153. your search results
backed by shared
authority records

154. put it all together

155. visit
/subject/lccn/n81072976
• get your content results
• get alternate data views
(concept record)
• get links to other useful data
• get links to authorized form

156. all that’s easy

157. i promise

158. the next part’s
hard

159. how to make it
last

160. if your site breaks
when links break
cache and link yourself

161. instead of just this

162. do this
UChicago
same as
cache

163. just like
we do
in ILS/OPACs

164. but, yknow,
the web way

165. so if a remote
link breaks
your local links
still work

166. also,
proxy concepts

167. your
/subject/lccn/n81072976
==
my
/subject/lccn/n81072976

168. but in your app

169. make every cache
its own
linked data source

170. instead of just this
UChicago
same as
cache

171. enable this
UChicago same as
cache
same as same as
same as
same as

172. if one goes down
the others live on

173. if others live on
it’s all still linked

174. if it’s all still linked
we can keep adding

175. linked data
that lasts

176. i think

177. summary
• making the web
better
• follow Linked Data
practices
• easy changes to
make now

178. summary
• get more value
from our data
• get more users
to our data
• make relevance
explicit

179. thanks!
dchud at umich edu
onebiglibrary.net
twitter.com/dchud