TheDynamicWeb Zoetrope: WWW Interac-ng with the Ephemeral Web - PDF document

6/8/09
 The
Dynamic
Web
 Zoetrope: WWW
 Interac-ng with the Ephemeral Web Eytan
Adar,
Mira
Dontcheva
 James
Fogarty,
Dan
Weld
 University
of
Washington
&
Adobe
Systems
 The
Now
Web
 Is
this
really
a
good
price?
 See
the
web
through
a
 Access
services
through
 browser
of
“now”
 an
index
of
“now”
 Temporally
 insensi-ve 
 (ignore
temporal
context
and
historical
values)
 1


6/8/09
 Is
this
really
a
good
price?
 What
is
the
relaRonship
between
 weather
and
traffic
in
SeaTle?
 When
is
really
the
right
Rme
to
buy?
 What
is
the
relaRonship
between

 traffic
and
temperature
in
SeaTle?
 traffic
and
temperature
in
SeaTle?
 an
alternaRve…
 Manually
 Temp t collect,
page
 Travel -me t by
page
 Now
what?
 2


6/8/09
 Related
Work
 Related
Work
 • EvoluRon
of
the
Web

 • EvoluRon
of
the
Web

 [FeTerly
et
al.,
WWW
‘03]
 • Monitoring
 [Ntoulas
et
al.,

WWW
‘04]


 [Sugiura
&
Koseki,
UIST
’98]
 Internet
Archives
 [Greenberg
&
Boyle,
GI
‘06]
 Photo
 Photo
 Related
Work
 Our
Approach
 • EvoluRon
of
the
Web

 • Monitoring
 • Clipping
&
AggregaRng
 Photo
 [schraefel
et
al.,
WWW
‘02]
 [Dontcheva
et
al.,
UIST
‘07]
 3


6/8/09
 Zoetrope
Architecture
 Content
Streams
&
Operators
 Zoetrope
Front
End
 R 1
 R 2
 R 3
 {…
<T i
 ,
C i >
,
<T i+1
 ,
C i+1 >
,
…}
 Zoetrope
Engine
 …
 …
 O 1
 O 2
 O 3 
 Data
Store
 XML/Images/Text/…
 Time
 Content
(typed,
but
can
be
 stamp
 anything:
HTML,
images)
 Web
 






Crawler
 Content
Streams
&
Operators
 Construct
everything
from
operators
 Pick
image
 {…
<T i
 ,
C i >
…}
 Generate
 Crop
to
 Input:
Content
Stream 
 based
on
 Images
 Rectangle
 slider
 {…
<T i
 ,
C i >
…}
 Transform
 Filter
 Render
 Render
into
 Operator
 Operator
 Operator
 lens
 Lens,
locked
 Output:
“Visual”
Output 
 Output:
Subset
of
Content
Stream 
 Output:
Content
stream
with
 to
coordinate
 modified
data
payloads
(C i ’) 
 4


6/8/09
 Construct
everything
from
operators
 44.10
 <…>
$44.10
<…>
 <…>
$50.20
<…>
 49.10
 43.10
 <…>
$4310
<…>
 <…>
$49.10
<…>
 50.20

 Extract
 {…
<T i
 ,
C i >
…}
 Select
DOM
 Render
into
 numerical
 forest
 Rme
series
 value
 LENSES
 Visual

(same
place
on
screen)
 Structural
(same
place
in
HTML)
 DOM
Structure
 5


6/8/09
 Content
(same
text)
 Filters
 RENDERERS
 6


6/8/09
 Binding
Lenses
 Demo
 7


6/8/09
 ContribuRons
 Future
Work
 • Visual programming toolkit 
for
temporal

 • EvaluaRon
 Web
content
 – User
Study
 • Seman-cs for
temporal
data
streams
 – Performance
 • Operators for
fast
processing
of
temporal
 data
streams
 • Indexing structures for
fast
processing
and
 interacRon
with
Web
content
over
Rme
 Future
Work
 InformaRon
Arbitrage
 • EvaluaRon
 • Combine
observaRons
to
improve
granularity
 – User
Study
 – Performance
 • Data
gathering
 cnn.com
 Rme
 Weather.com
 8


6/8/09
 Future
Work
 Possible
Distributed
Architecture
 • EvaluaRon
 • Server
collects
snapshots
 – User
Study
 • Client
collects
finer
resoluRon
 – Performance
 – Gaps?
 • Data
gathering
 • DHT
for
collected
data
(good
for
intervals)
 • Scaling
up
 • Query:
operator
chain
+
data
 – Release
on
similarity
threshold
 – Repeat
 Thanks!
 Jaime
Teevan,
Susan
Dumais,
Rob
Miller
,
GRAIL,
DUB
 NSF
Graduate
Fellowship,
ARCS,
WRF/TJ
Cable
Professorship
 ?
 9