The document describes how people communicated and socialized in the past compared to today. In the past, people would write long emails and wait to get photos printed. They would listen to full albums of music. Now, communication happens through brief social media updates, short videos, and one photo at a time. Researchers have found it difficult to directly observe mobile communication practices as they happen. However, the document proposes following people around with video recording, as people now record many aspects of their daily lives, to understand how people use different technologies to stay connected throughout the day. The study involved following four groups of close friends/family for half a day each to see their communication from each person's perspective.
Apidays New York 2024 - Scaling API-first by Ian Reasor and Radu Cotescu, Adobe
Integrating Local and Remote Worlds Through Channel Blending
1. Once
upon
a
*me,
long,
long
ago,
people
communicated
and
socialized
in
curious
ways.
1
2. They
wrote
whole
paragraphs
at
a
*me
–
some*mes
more
-‐-‐
and
sent
them
through
email,
delighted
that
the
recipient
might
read
their
message
that
very
day!
2
3. They
waited
un*l
they
had
taken
24
or
even
36
photos
before
geFng
them
printed
on
paper..
3
4. Then
they’d
get
together
with
people
in
person
and
tell
the
stories
depicted
in
the
photos.
4
5. And
some*mes,
also
when
they
were
in
the
same
room
together,
they
watched
a
single
video
for
two
whole
hours
at
a
stretch.
5
6. People
listened
to
12
songs
by
the
same
ar*st
-‐-‐
all
in
a
row
-‐-‐
stored
on
preKy
silver
disks.
6
7. Now,
of
course,
the
world
has
changed.
-‐
Content
now
comes
in
bite-‐size
pieces:
status
updates
and
text
tweets,
3
minute
videos,
and
people
send
one
photo
at
a
*me.
-‐
People
capture
their
experiences
in
many
more
ways
than
just
photos
and
videos.
Now
they
track
their
loca*ons,
their
exercise
and
food
intake,
their
game
playing,
even
their
sleeping
paKerns.
-‐
They
access
this
content
from
just
about
anywhere
via
mobile
phones,
laptops,
tablets,
music
players,
ereaders,
and
so
on.
-‐
And
people
connect
with
others
from
almost
anywhere
via
tex*ng,
video
chaFng,
Facebook,
TwiKer,
phone
calls,
email,
and
so
on
-‐ We
wanted
to
understand
how
all
these
changes
have
affected
the
way
people
stay
connected
with
friends
and
family
while
they’re
mobile.
Most
ethnographic
studies
of
mobile
communica*on
tend
to
focus
on
the
use
of
a
single
technology,
but
we
wanted
to
see
how
people
use
the
whole
assortment
of
technologies
available
to
maintain
an
ongoing
connec*on
over
the
course
of
the
day.
7
8. We
had
in
mind
that
it
might
look
something
like
this.
Maybe
people
would
have
a
face-‐to-‐face
interac*on
in
the
morning,
then
perhaps
exchange
emails
or
IMs,
then
later
in
the
day
maybe
they
send
a
photo
or
text,
and
then
they
get
back
together
in
person
at
the
end
of
the
day.
The
ques*on
was,
how
could
we
study
this?
The
best
way,
of
course,
is
to
be
there
all
day
so
you
can
watch
it
when
it
happen,
but
this
is
difficult.
8
9. Grinter
&
Eldridge
in
2003
were
interested
in
studying
tex*ng,
said
they’d
ideally
like
to
“directly
observe”
teenagers’
text
messaging
prac*ces,
but
found
“direct
observa*on
highly
imprac*cal,”
so
they
felt
they
had
to
use
indirect
approaches
such
as
interviewing
and
logging.
9
10. Two
years
lager
Ito
&
Okabe
also
noted
that
it
is
“notoriously
difficult
the
flee*ng
par*culari*es
of
mobile
communica*on,”
so
they,
too,
studied
tex*ng
using
“second-‐
hand”
methods.
10
11. But
we
started
to
think
that
things
have
changed
since
those
studies
were
done.
People
are
now
video
recording
everything,
everywhere
–
in
restaurants,
train
sta*ons,
drug
stores,
out
on
the
street
–
so
that
they
can
post
their
videos
on
YouTube.
So
we
though,
maybe
we
can
do
it
too.
Maybe
we
can
follow
people
around
and
video
record
them
and
it
won’t
seem
that
remarkable
to
people.
We
were
crazy
enough
to
try
it.
11
12. We
found
four
groups
of
close
friends
and
family
and
observed
each
of
them
for
half
a
day.
Three
were
in
different
parts
of
the
San
Francisco
Bay
Area,
and
one
was
split
between
California
and
Texas.
To
observe
them,
we
had
a
different
researcher
follow
each
person
in
the
group
as
they
went
about
their
ac*vi*es
over
the
same
*me
period
and
we
watched
how
they
stayed
in
touch,
either
through
technology
or
in
person.
So
we
saw
each
person’s
point
of
view
as
they
connected,
engaged,
and
disconnected.
12
13. These
images
give
you
a
sense
of
how
this
worked.
In
this
case,
we
studied
two
roommates
who
also
worked
together
–whom
we’re
calling
Anita
and
Cathy.
I
was
following
Anita,
and
afer
work
she
first
drove
to
a
shopping
center
where
she
would
meet
up
with
Cathy.
13
14. She
did
a
liKle
shopping,
some*mes
on
her
own,
some*mes
browsing
together
with
her
friend
Cathy.
14
15. Then
they
came
together
to
pay
for
their
purchases.
You
can
see
Yutaka
there
video
recording
Cathy
while
I
recorded
Anita.
15
16. Then
they
went
to
dinner
in
a
restaurant
while
each
of
them
connected
with
other
people
through
their
phones
16
19. Then
returned
to
her
apartment
in
California
and
had
a
video
chat
with
some
friends
in
New
York
19
20. while
Cathy
listened
in
from
her
room
while
also
doing
her
own
thing
–
and
I’ll
be
showing
you
a
video
of
this
shortly
20
21. Then
at
the
end
of
the
night
they
got
together
to
watch
TV
while
surfing
also
the
net
and
chaFng
with
friends
on
their
phones.
21
22. What
did
we
find?
In
a
nutshell,
we
saw
a
collec*on
of
related
behaviors
that
we
are
calling
channel
blending.
First
I’ll
define
channel
blending
and
then
I’ll
show
you
a
few
examples.
Channel
blending
is
the
integra*on
of
interac*ons
and
content,
over
mul*ple
channels
–
and
face-‐to-‐face
counts
as
a
channel
–
into
one
coherent
conversa*on.
The
conversa*ons
usually
involve
both
local
and
remote
par*cipants
–
meaning
there
are
mul*ple
people
in
the
room
communica*ng
with
one
or
more
people
in
a
remote
loca*on.
Some*mes
people
blended
channels
not
all
at
once
but
as
they
carried
on
a
conversa*on
over
*me.
22
23. You
can
think
of
channel
blending
as
both
similar
to
and
the
opposite
of
mul*tasking.
So
on
the
lef,
two
people
are
face-‐to-‐face
watching
a
basketball
game
while
video
chaFng
with
two
others
who
are
also
watching
the
game.
They’re
looking
up
informa*on
on
their
tablet
and
phone
related
to
the
game
and
maybe
chaFng
with
someone
else
about
the
game
on
the
laptop.
So
they’re
using
many
media
but
converging
them
into
one
coherent
conversa*on
about
the
basketball
game.
On
the
right,
you
have
the
same
configura*on
of
media
but
now
there’s
just
one
person
in
each
space
having
mul*ple
interac*ons.
The
local
person
is
watching
the
game
while
also
video
chaFng
with
someone
on
the
laptop,
and
they’re
both
engaged
with
their
phones
for
different
purposes,
and
the
local
person
is
also
text
chaFng
with
someone
about
something
else.
They
key
difference
is
that
on
the
lef,
there’s
one
conversa*on
going
on,
and
on
the
right
there
are
mul*ple
conversa*ons.
The
discussion
might
move
on
from
basketball,
of
course,
but
everyone
would
be
included
as
the
conversa*on
moved
on
to
new
topics.
In
the
channel
blending
case,
there
are
two
people
in
the
local
space,
and
two
in
the
remote.
This
was
very
common
–
to
have
more
than
one
person
together
in
one
place,
possibly
connec*ng
with
mul*ple
remote
others
in
another
space.
23
24. There
were
two
categories
of
channel
blending.
Some*mes
people
channel
blended
interac(ons
–
bringing
together
mul*ple
people
across
mul*ple
channels.
So
here
it’s
the
video
chaFng,
face-‐to-‐face,
and
text
chaFng
all
blended
together.
And
some*mes
they
channel
blended
content
from
mul*ple
sources
into
the
one
conversa*on.
So
here
it’s
the
game
and
all
the
things
they’re
looking
up
on
their
devices.
24
25. And
of
course
you
could
have
blending
of
both
interac*ons
and
content
in
one
conversa*on.
25
26. Now
I’d
like
to
show
you
a
case
of
channel
blending
an
interac(on
over
mul*ple
channels.
It
involved
Anita
and
Cathy
when
they
were
back
at
their
apartment
in
California
and
Anita
was
chaFng
with
two
friends
in
New
York.
In
the
adjacent
room,
Cathy
was
packing
for
a
2-‐month
trip
to
Dublin,
Ireland,
where
she
would
be
running
an
internship
program
for
her
company.
Anita,
too,
would
soon
be
going
to
Los
Angeles
to
run
another
internship
program
there.
Here
they’re
talking
about
the
upcoming
travel.
In
it,
they
talk
about
a
couple
of
movies
involving
danger
in
foreign
ci*es
called
“Taken”
and
“Hos*le.”
You’ll
no*ce
how
Anita
acts
as
what
we
call
a
“pivot
person”
to
blends
the
two
spaces
together,
relaying
comments
back
and
forth
and
keeping
everyone
involved
in
the
one
conversa*on.
[PLAY
CLIP]
They
key
things
to
no*ce
here
are
how
Anita
is
interac*ng
with
people
remotely
and
locally
at
the
same
*me,
and
she
is
integra*ng
the
two
“channels”
into
one
conversa*on.
As
the
pivot
person,
she
some*mes
aKributes
Cathy’s
comments
to
her,
as
when
she
says
“I
love
how
they
ask
as
if
you
have
a
choice”
or
at
the
end,
she
just
integrates
Cathy’s
comments
as
her
own,
saying
“Anita
loves
LA!”
26
27. In
an
earlier
interac*on
at
dinner,
they
both
had
their
phones
out
and
they
both
communicated
with
others
while
s*ll
talking
to
each
other.
Instead
of
carrying
on
separate
conversa*ons
locally
and
remotely,
they
shared
with
each
other
what
they
were
reading
or
typing
to
others,
blending
the
local
and
remote
spaces
into
one
conversa*on.
This
diagram
shows
their
conversa*on
over
the
9
minutes
before
the
waiter
took
their
order.
The
orange
are
topics
they
generated
from
their
current
environment
(things
about
the
restaurant
or
work
that
day),
and
the
blue
are
topics
that
emerged
from
interac*ons
they
were
having
on
their
phones.
You
can
see
how
they
blended
them
together
into
one
interac*on.
They
both
read
aloud
what
they
were
reading,
and
verbalized
what
they
were
typing
as
they
texted
or
responded
to
a
tweet.
So
in
this
case,
they
both
acted
as
pivot
people
for
their
interac*ons
with
different
remote
others.
We
also
thought
it
was
interes*ng,
as
a
side
note,
that
their
conversa*on
–
with
its
quick,
bursty
nature
about
many
different
topics
–
mirrored
exactly
the
type
of
online
interac*ons
they
were
having
over
text
chat,
TwiKer,
Four
Square,
and
Facebook.
27
28. This
next
example
shows
how
people
channel
blended
content
from
mul*ple
sources.
In
this
case
they
were
all
face-‐to-‐face,
and
they
brought
in
content
from
their
phones
and
a
laptop
as
the
conversa*on
unfolded.
This
example
is
interes*ng
because
it
also
illustrates
channel
blending
over
*me.
These
three
people
are
musicians
and
a
few
days
before
this
conversa*on,
the
one
on
the
lef
had
posted
a
link
to
a
recording
of
a
violin
concerto
that
was
played
by
a
guy
who
at
the
*me
was
15
years
old.
It’s
a
very
difficult
piece
and
it’s
amazing
that
someone
so
young
played
it
so
well,
but
there’s
one
sec*on
in
the
piece
where
he
screws
up
and
has
trouble
hiFng
a
high
note.
It’s
a
funny
clip
because
of
the
squeaky
notes.
They
had
all
played
the
clip
on
their
own
and
laughed,
but
now
that
they
got
together,
it
comes
up
in
conversa*on
and
they
want
to
play
it
again.
So
you’ll
see
the
woman
figuring
out
how
to
access
the
clip
on
her
phone,
with
help
from
the
others.
Afer
this,
they
con*nue
to
discuss
musical
pieces
and
each
of
the
men
in
turn
brings
out
their
phone
to
access
a
different
song
and
the
one
on
the
lef
opens
his
laptop
when
he
can’t
find
it
on
his
phone.
Then
he
uses
it
to
find
a
picture
of
the
sheet
music
on
Facebook.
[PLAY
CLIP]
So
each
used
their
own
devices
–
phones
and
laptops
–
to
access
content
based
on
28
29. Those
are
just
a
few
of
the
many
examples
we
saw,
and
more
are
described
in
the
paper.
So
when
we
go
back
to
our
original
concep*on
of
what
we
might
find
when
watching
how
people
communicate
throughout
the
day,
we
discovered
that
instead
of
looking
like
this…
29
30. It
looks
much
more
like
this.
With
mul*ple
people
together
in
a
shared
space,
communica*ng
with
mul*ple
remote
others
using
mul*ple
devices,
and
sharing
different
types
of
content.
We
want
to
be
clear
that
other
have
no*ced
different
aspects
of
this,
like
people
carrying
on
a
conversa*on
with
both
local
and
remote
par*cipants,
or
carrying
on
a
conversa*on
over
*me.
What
we’re
proposing
here
is
that
it’s
useful
to
think
of
these
cases
as
part
of
a
larger
phenomenon
that
involves
blending
different
channels
into
one
conversa*on.
We
think
this
may
happen
rela*vely
ofen,
and
it’s
likely
that
at
least
some
of
the
cases
that
are
reported
as
mul*tasking
may
actually
be
channel
blending.
30
31. Now
let’s
think
about
the
technology
people
are
currently
using.
Most
of
it
is
designed
to
connect
just
two
people,
with
the
assump*on
that
they
are
each
alone
with
a
single
device.
This
is
the
case
with
phone
calls.
31
33. Other
technology
is
designed
for
many-‐to-‐many
interac*ons,
but
even
s*ll,
the
assump*on
is
s*ll
that
each
one
is
alone
in
their
space
with
a
single
device,
not
sharing
with
people
in
their
local
space.
33
34. Yet,
what
was
saw
was
that
communica*on
looked
much
more
like
this.
Where
you
have
mul*ple
people
in
the
same
space,
and
small
groups
of
people
are
trying
to
connect
–
not
two
people,
but
more
like
3
or
4
or
5.
Here
the
two
men
are
on
the
phone,
and
she’s
chiming
in
as
he
relays
comments
back
and
forth.
At
one
point
he
uses
his
computer
to
enter
in
some
informa*on
they
give
him
34
35. Here
she’s
IM-‐ing
with
that
same
person
and
now
she’s
relaying
comments
back
and
forth.
They
also
discuss
a
Words
With
Friends
game
they’ve
been
playing
on
their
phones.
35
36. This
one
is
a
case
of
massive
content
blending
–
they
all
got
together
to
watch
a
sequence
of
videos
projected
from
a
computer
onto
the
TV
while
also
looking
up
related
informa*on
on
their
phones
and
laptops.
At
one
point
they
all
pulled
out
their
phones
to
compare
their
sleep
cycles,
which
they’d
all
recorded
with
an
app.
36
37. In
this
case
this
woman
is
video
chaFng
with
her
husband
while
also
text
chaFng
with
a
woman
in
this
IM
window
–
and
she’s
relaying
comments
back
and
forth
as
they
work
together
to
coordinate
an
upcoming
visit,
which
they
had
all
discussed
in
an
earlier
email.
37
38. This
one
looks
simple
but
is
complicated.
The
daughter
is
talking
with
a
friend
about
some
documents
on
her
screen,
which
he
had
emailed
her
and
is
asking
her
to
print
out
and
bring
to
him
when
they
meet
up
later.
Meanwhile,
the
mother
is
talking
with
another
daughter
–
who
this
daughter
had
just
been
talking
with
before
she
handed
her
mother
the
phone.
38
39. Finally,
these
three
sisters
are
having
a
three-‐way
phone
call
while
two
of
them
also
look
at
and
discuss
a
website
on
their
computers,
but
the
third
one
can’t
share
in
it.
In
fact,
she
then
complains
about
the
problems
with
sharing
in
this
short
clip.
39
41. So
what
does
all
mean
for
technology
design?
If
people
are
doing
a
lot
of
channel
blending,
how
should
we
be
suppor*ng
it?
We
think
technology
needs
to
support
small
groups
of
people
–
not
just
one-‐to-‐one
or
many-‐to-‐many
Some
together,
some
apart
–
not
just
with
each
person
alone
in
separate
loca*ons
With
fluid
levels
of
par*cipa*on
–
like
with
the
Pepper
Spray
example,
people
in
the
same
space
ofen
monitor
conversa*ons
and
engage
to
different
degrees
at
different
*mes
As
they
show
&
tell
&
react
to
the
stories
of
their
lives,
since
that’s
what
people
are
doing
Using
mul*ple
media
and
devices.
People
aren’t
just
using
a
single
device
at
a
*me
or
even
a
single
device
per
person
Over
mul*ple
interac*ons
–
people
not
only
con*nue
their
conversa*ons
across
*me
and
media,
they
also
want
to
re-‐share
the
same
content
when
they
get
together
face
to
face
or
afer
they
separate
and
reconnect
remotely.
41
42. Once
upon
a
*me,
long,
long
ago,
people
communicated
and
socialized
in
curious
ways.
42