►
Description
Cindy Bruyere: Cindy Bruyere is a scientist at NCAR (National Center for Atmospheric Research). She is the director for NCAR’s Capacity Canter for Climate and Weather Extremes. A group that conducts actionable science to advance our understanding of weather and climate extremes on time scales from sub-seasonal to multi-decadal. Her background is in meteorology and environmental management.
Mike Daniels: Mike Daniels’ career has been dedicated to the leadership and development of complex computing, software engineering, data management and cyberinfrastructure systems for the geosciences. His recent projects range from executive leadership and governance, real-time data acquisition and access, development of science workflows, data management, and linking data artifacts to research by way of semantic architectures.
Date: 11/06/20
Presenter: Cindy Bruyere & Mike Daniels
Institution: UCAR
West Big Data Hub
A
B
B
So
so
john
and
I
are
going
to
take
the
leadership
across
the
four
hubs
we
spoke
to
the
other
eds
about
the
whole
cyber
security
space,
and
so
that
includes
the
trustworthy
data
working
group
stuff.
We're
also
going
to
be
looking
at
the
cyber
security
initiatives
across
the
hubs
and
try
to
figure
out
what
we're
all
doing
and
how
we
can
help
each
other
there's
a
social
cyber
security
step
in
the
south,
and
we
have
a
new
connected
healthcare,
cyber
security
workshop
working
on
the
northeast
and
a
bunch
of
stuff.
B
A
Sure,
let's
put
a
pin
in
that
for
a
little
discussion,
and
maybe
so
we
do
have
a
just
I'll
sort
of
let
everybody
know.
We
do
have
a
little
bit
of
other
discussion
that
I
wanted
to
get
started
on
after
this,
which
is
sort
of
around
the
notion
of
a
theme
for
the
spring,
so
that
we
can
start
looking
for
people
to
recruit
for
talks
and
and
and
go
into
having.
A
B
Could
be
cyber
security
and
the
other
thing
I
just
want
to
mention
last
time
I
I
we
were
talking
about
the
covet
information,
commons
and
we've.
We've
we've
delivered
our
our
third
search
mechanism,
which
has
pi
data
in
it.
So
now
we're
looking
at
the
kick
follow-on
and
we
spoke
to
the
hpcc
crowd
this
morning.
You
know
and
john
towns
and
the
crowd
and
so
we're
looking
at
how
we
can
work
together
on
this
kick
follow-on,
and
I
think
I
mentioned
it.
The
last
time
you
got
all
excited
niall.
C
A
C
Thanks
I'll
just
put
a
couple
of
links
in
the
chat
here
for
folks
for
a
couple
of
upcoming
activities,
so
the
the
hub
is
working
with
some
partners
in
our
region
and
nationally
on
what
we're
calling
the
water
data
forum,
which
is
a
series
of
events
throughout
2021,
focused
on
bringing
together
different
communities
that
are
that
are
doing
work
in
the
water
space.
You
know
it
spans
everything
from
academic
researchers
to
industry,
to
government
and
quasi
government
agencies.
C
C
D
Yeah
it's
along
the
lines
of
cyber
security,
so
purdue
host
is
hosting,
I
should
say,
a
series
on
cyber
security
and
higher
ed,
a
cui
workshop,
and-
and
so
we
had
our
kickoff
just
yesterday
the
day
before
yesterday
and
had
about
33
other
institutions
participating
in
that
event,
a
lot
of
great
knowledge
sharing
and
given
that
this
was
the
kickoff.
I
think
we'll
hear
more
about
that
and
I'll
certainly
report
out,
as
as
we
progress.
E
E
Those
there
was
a
whole
sort
of
contest
with
that
and
prizes
were
awarded
a
couple
weeks
ago,
I'll
see
if
I
can
find
the
link
to
that
that
that
web
page
and
and
drop
that
in
the
chat
in
a
few
minutes,
but
that
seemed
to
be
successful.
It
was
really
interesting
for
me
to
sit
in
on
part
of
that
because
of
the
breadth
of
people
that
participated
in
it.
E
We
had
people
from
high
school
and
local
community
college,
some
transfer
students
and
faculty
and
researchers
from
a
range
of
organizations,
so
that
was
great
and
those
data
sets
are
available
still
and
we're
also
working
to
get
them
cataloged
and
available
in
a
couple
of
different
ways
as
well.
So
that's
all
I
have
to
report
today,
there's
probably
other
things
going
on,
but
I'm
afraid
I
I
didn't
get
a
briefing.
I
had
a
call
today.
A
A
You
know
how
to
do
the
the
all
hands
meeting
and
things
like
that,
but
those
will
be
topics
probably
more
for
the
spring,
for
people
to
report
out
on.
So
hopefully
we
can
get
a
little
info
in
in
next
month,
either
I'll
ping,
folks
for
it
or
or
make
sure
that
we
get
people
here
not
to
pick
on
you
twice
melissa
but
I'll
pick
on
you
twice
anything
new
and
interesting,
coming
out
of
the
open
search,
osen.
E
Yeah,
yes,
yes,
we
are
running
a
webinar
series
right
now,
the
aim
of
which
will
be
a
series
of
concept
papers.
That'll,
be
shared
immunity
quite
broadly,
and
also
help
us
move
toward
the
proposal
for
the
next
phase
of
the
osn
I'll
also
drop
the
link
in
for
those
you
do
have
to
register
for
those
webinars
I'll
grab
those
and
drop
it
in
the
chat
as
well,
and
so
those
are
happening.
E
A
E
E
And
lastly,
we
are
working
with
a
couple
of
organizations
to
see
if
we
can
develop
a
third
party
catalog
separately,
so
that
there
would
be
that
there
would
actually
be
landing
pages
for
publicly
available
data
sets
through
other
catalogs.
So
we're
working
on
that.
But
we
we're
we're
nowhere,
we're
nowhere
near
done
in
development
with
that
and
so.
A
E
A
I
know
florence
had
the
suggestion
that
we
talk
a
little
bit
about
cyber
cyber
security
and
and
sort
of
coordinating
efforts,
and
maybe
that
should
be
topic
one.
So
why
don't
we
see
if
that?
If
that
fits,
but
let
me
introduce
the
sort
of
spring
topical,
deep
dive.
I
I,
when
we
were
talking
about
this
last
time
I
scribbled
down
some
notes
that
I
think
I've
translated
into
english
or
at
least
into
words
that
I
can
convey.
A
I
know
we
there's
a
lot
of
work
that
goes
around
that
a
lot
that's
played
to
it,
I
think,
even
just
especially
getting
people
to
understand
the
different
forms
of
reproducibility
is
sort
of
important
because
it
isn't
just
about
making
sure
you
get
the
same
number
every
time
there
are.
There
are
other
forms
of
it
improving
collaboration
on
workflows.
A
I
know
we've
got
some
things
from
some
of
the
darpa
projects
that
we
have
that
we
could
talk
to
and
I'm
sure
there
are
other
ones
around
that
to
talk
about
those
things
and
how
cyber
infrastructure,
around
things
like
containers
and
and
more
flexible
schema
databases
and
other
things,
can
all
help
in
that
from
the
ci
point
of
view,
but
then
also
how
users
who
aren't
gurus
can
participate
in
that
I
put
down
sort
of
pathways
towards
fair
data
for
more
users.
I
know
everybody
likes
to
say
things
about
fair
data.
A
If
we're
going
into
you
know,
hazards
or
or
climate,
or
you
know
any
of
the
other
ones
that
we
could
dive
into.
So
I
think
that's
sort
of
an
intro
on
that.
Why
don't
I
throw
it
open
to
everybody
to
sort
of
give
opinions
on
these
and
see
if
there's
anything
that
really
stands
out
as
something
we
should
cover
in
the
spring?
A
A
B
Absolutely
on
that
on
mute,
so
there
are
two
things
one
is
you
know:
jim
has
been
working
with
us
on
the
trustworthy
data
working
group,
jim
basney
and
maybe
jane
wilkin
bush
too,
and
you
know
john
mcmullen
and
I
have
been
participating
quite
a
bit
from
the
hubs,
and
so
there
are
a
couple
of
things.
We're
thinking.
One
is:
how
do
we
leverage
the
trustworthy
data
working
group
work
to
benefit
our
constituency?
You
know
the
scientists
and
researchers,
so
we
have
a
couple
thoughts
and
then
how
do
we?
B
How
do
we
increase
awareness
that
people
have
to
be
thinking
about
this
stuff
and
how
we
help
them?
Actually
do
it?
I,
I
think
I
think
people
need
to
be
thinking
about
it
more
and
making
sure
that
they're
doing
what
they
should
do
for
good
hygiene
and
all
the
new
things
that
keep
coming
about
for
cyber
security
and
data.
So
a
couple
things
we
were
thinking
about
was
to
work
with
the
trustworthy
data
working
group
to
identify
some
of
the
best
practices
and
tools
people
are
already
using.
B
You
know
that
we
could
communicate
in
some
way.
Then
we
could
think
about
either
doing
a
webinar
to
communicate
the
information
and
or
having
a
collaborative
workshop
to
discuss.
You
know
the
increasing
challenges
in
cyber
security
for
our
scientific
and
research.
You
know,
partners
and
and
how
we
can
work
together.
So
those
are
just
some
ideas
that
we
had
on
working
with
tdwg
the
toastery
data
working
group
data
and
then
separately.
B
C
That
was
great
thanks
florence.
I
guess
one
of
the
discussions
that
we
have
one
of
the
open
questions
that
we
had
was:
who
is
the
audience
for
some
of
these
activities?
And
so
you
know,
would
it
be
like
a
cso
level
discussion
around
tools
and
and
issues?
Would
it
be
more
of
a
operations,
hands-on
workshop
kind
of
thing,
so
I
think
we're
still
trying
to
decide
what
makes
the
most
sense
for
the
the
people
that
we
have
engaged
in
this.
F
D
Yeah,
those
are
really
good
questions.
I
think
we're
going
through
cmmc
right
now.
You
know
with
specific
interest
in
in
cui
and
defense
related
information
and
and
what
what
I
found
is
that
the
all
the
groups
that
you
just
mentioned
florence,
like
think
of
this
in
very
different
ways,
not
surprisingly
right
and
and
what
I
guess
where
I've
benefited
from
the
discussion
that
kicked
off
on
wednesday
in
purdue
was,
you
know,
really
hearing
more
from
people
outside
of
research
computing.
D
D
You
know
kind
of
explicitly
across
the
areas
that
you
mentioned.
You
know
I
I
think
one
way
to
maybe
do.
That
would
be,
if
we
invite
speakers
to
say
this
is
going
to
be
sort
of
a
technical
examination
of
some
of
the
some
of
the
requirements,
but
then
really
try
to
make
sure
that
all
audiences
are
are
welcome
might
be
an
approach,
because
those
communities
can
really
learn
a
lot
from
even
the
deeper
dives
and
then
also
contribute
to
the
discussion.
D
But
it's
a
it's
it's
a
great
topic.
I
would
totally
vote
for
that.
If
we're
voting,
I
think
that
the
timeliness
of
something
like
what
you
proposed
is
is
great
and
I
think
we'll
there's
a
lot
of
different
venues
that
we
can
push
that
information
out
to
and
potentially
get
a
broad
audience
to
contribute
to
the
discussion.
A
So
I'm
even
thinking
of
of
you
know
I
mean
I'll
just
put
this
out.
We
can
change
the
entire
format
of
of
this.
We,
you
know,
we
we've
been
doing
sort
of
the
we
have
a
meeting.
We
have
lightning
talks
for
two
years.
Do
we
want
to
turn
it
into
even
in
the
fall?
Having
some
meetings
just
be?
You
know
five
minute
talks
of
a
panel
and
then
open
discussion,
kind
of
things
for
the
rest
of
the
hour,
and
you
know
I
mean
along
these
lines.
A
You
know,
I
think
you
know
just
cover
the
policies
that
we're
all
talking
about
in
concrete
ways.
You
know
what
is
800
171
versus
kua
vs
hipaa
in
real
ways,
not
in
the
fear
ways
that
everybody
talks
about
them
and
keeps
anything
from
happening,
but
in
realistic
implementation
and
protection
ways
or
I
wrote
down
horror
stories
and
how
how
to
not
be
that
project,
but.
B
B
B
A
That
would
be.
That
could
be
another
one.
You
know
why
this
matters
to
the
researcher,
why
it
matters
to
the
cso.
Why
it
matters
to
the
developer.
You
know
put
it
in
terms
for
the
developer.
Maybe
that's
what
we
do
is
those
are,
let's
pick
four
cyber
security,
I'm
just
I'm
brainstorming
like
I
do
you
all
know
me.
D
In
terms
of
format,
I
I
like
the
idea
of
of
a
panel
in
the
sense
that
we
could
have
you
know
a
series,
but
then
there's
there's
so
many
individuals
that
and
institutions
that
would
have
something
to
share
that
that
one
that
one
subject
could
be
from
the
vantage
of
of
four
different
participants
and
exactly
and
and
sort
of
just
sweet.
You
know
and
relatively
short,
but
enough
that
the
the
panelists
could
say
and
if
you
want
to
learn
more
about
this,
go
to
this
url.
A
Yeah,
I
think
we,
you
know
we
ended
with
hey.
Can
we
you
know
for
those
people
that
are
interested
in
doing
a
deeper
dive
on
even
these
topics?
You
know,
here's
a
here's,
a
you
know
schedule
that
we
might
be
able
to
set
up.
You
know
we
can
have
these
meetings
anytime
these
days,
it's
not
like.
We
have
to
have
them
on
the
first
friday
and
we
can
have
breakout
sessions
during
the
month
and
maybe
somebody
can
come
and
report
as
to
what
came
of
that
at
the
next
meeting.
A
B
Let's
talk
about
that
having
enough
people,
so
these
calls
are
tend
to
be
rather
small.
This
is
like
the
people.
You
know
who
sit
around
our
thanksgiving
dining
room
table
right.
These
are
these
tend
to
be
small.
This
is
like
the
core
family.
We
love
it
when
more
people
come,
but
it's
not
it's
not
broad
right
at
this
point.
So
the
question
is:
if
we
really
want
to
do,
have
broader
impact.
B
We
could
actually,
you
know,
put
the
word
out
ask
for
people
who
would
like
to
be
on
a
panel
and
share
the
good,
the
bad
and
the
ugly
yeah
right
either
they
have
a
great
answer,
they've
lived
through
it
and
they
want
to
warn
other
people
or
something
else,
and
then
we
attract
people
by
the
different.
You
know
cohorts
as
we're
saying
they'd
all
be
represented
on
the
panel
like
three
or
four
constituencies,
and
then
we
have
breakout
sessions
for
each
of
those
constituencies.
B
So
they
can
talk
about
you
know,
what's
your
reality
gosh,
what
do
we
really
need
and
maybe
part
of
what
we
present
are
some
of
the
best
practices?
You
know
that
come
out
of
the
the
follow
on
to
the
tdwg
and
jim
bassey
I'd
love
to
hear
what
you
have
to
think
about
how
we
would
get
those
answers
from
the
crowd.
So
maybe
we
we
try
to
do
something
where
we,
you
know
we
go
out
to
our
northeast
hub
distribution
lists
with
this
idea.
A
A
A
B
D
D
A
Know
I
agree
so
I
I
just
looked
at
the
clock,
so
I'm
gonna,
I'm
gonna,
I'm
gonna,
take
you
know
the
role
of
of
keeping
us
on
time,
or
at
least
trying
to
we
can
go
a
little
longer
on
the
talk.
That's
not
a
big
deal,
but
I
wanted
to
say
it
sounds
to
me
at
least
like
by
the
discussion
that
this
one
has
has
a
a
good
start
right
now
to
get
us
going
in
the
spring.
A
I
would
rather
we
pitch
it
and
actually
make
it
be.
Let's
have
a
panel
and
then
if
there
are
people
who
want
to
go
and
do
more,
that's
sort
of
the
matchmaking
aspect
of
what
we
try
and
do
and
those
groups
are
free
to
use
our
resources.
You
know
zoom
meetings
are
free
now,
so
anybody
can
give
them
one,
but
you
know
we
could
do
that
in
some
way
and
then
maybe
just
report
on
that
back.
But
it's
not
like
it's
not
an
official
thing.
A
A
Well,
can
we
shift
gears
to
talking
a
little
bit
about
workflows
which
are
less
less
stressful
than
cyber
security
and
just
as
important,
I
think
so,
but
you
have
to
have
secure
workflows
to
make
them
work.
So
let
me
flip
some
tabs
here,
and
so
I
just
want
to
welcome
cindy
briere,
I
believe
is
that
how
you
pronounce
last
name.
Thank
you.
A
A
Oh
okay,
I
saw
the
csu
post
on
there
too
so
but
they're
coming
to
us
from
boulder
and
they're
here
to
talk
about
accelerating
scientific
workflows
and
work.
Well,
city's
been
working
on
capacity
center
for
climate
weather
extremes,
which
is
you
know,
taking
all
of
the
weather
and
weather
data
and
climate
models
and
many
things
over
the
multiple
decades
and
trying
to
make
it
usable,
which
is
a
huge
task.
A
Having
you
know,
it's
even
more
complicated
in
atmospheric
science,
that
is
an
astronomy,
and
I
tried
to
do
an
astronomy
for
many
many
years
and
mike
is
working
with
this
group
as
well.
Around
complex
computing
software
engineering,
data
management
and
cyber
infrastructure
systems
for
geosciences
and
he's
he's
been
working
on
governance,
real-time
data
acquisition
and
development
of
science,
workflows
and
they're
here
to
tell
us
about
some
of
those
science
workflows
so
rather
than
me,
taking
up
all
the
time
introducing
them
cindy.
Are
you
starting.
F
So
I've
been
really
interested
in
this
from
that
perspective,
so
we
have
a
pilot
project
funded
through
nsf
earthquake
to
that
we
call
asset
that
looks
at
accelerated
workflows
and
what
we
were
really
interested
in.
Seeing
that
there's
so
many
tools
that
has
already
been
developed
and
is
part
of
the
earthquake
repository.
F
We
wanted
to
start
using
those
tools
and
making
it
easier
for
people
to
use
the
tools,
but
we
also
landed
up,
and
what
I
want
to
talk
about
today
is
the
whole
part
of
why
people
are
not
doing
this.
So
there
are
a
lot
of
data
available,
but
people
are
not
really
picking
up
somebody
else's
workflow
or
somebody
else's
software
and
doing
something
with
it.
F
So
if
I
can
just
seem
to
move
on
okay,
so
the
team
that
we
have
is
myself
and
mike,
we
also
partnered
with
scott
beckham
from
cu
and
yolanda
gill
from
ucs,
and
then
we
have
collaborators
across
from
ucs
doing
a
lot
of
the
workflow
development.
We
have
partners
across
encore
and
we
also
have
a
partner
at
ucr,
and
that
quarter
is
is
very
important
in
mind.
F
Harrison.
I
can
never
pronounce
her
name
correctly.
That
partnership
for
us
is
also
very
important
because
we
look
at
workflows
from
a
scientific
perspective
and
she
will
look
at
workflows
from
a
physical
perspective,
so
collecting
soil
sample
data
and
how
you
can
use
the
same
workflow
systems
with
what
we
use
also
in
a
physical
sense.
F
So
here's
the
problem
that
I
think
we've
all
seen
and
all
face-
and
you
guys
have
just
talked
about
this.
We
often
talk
about
the
80
20
problem.
In
other
words,
we
spend
80
of
our
time
looking
for
data,
finding
data,
manipulating
data
and
only
20
of
our
time
doing
science,
and
if
we
can
flip
this
around
and
spend
less
time
on
developing
code
and
manipulating
data,
we
should
accelerate
the
time
to
science
significantly.
F
But
there's
there's
in
my
mind,
there's
two
piece:
two
pieces
to
it:
you
sort
of
broke
them
down
into
big
components
and
there's
a
lot
of
nuances
in
between
those
and
the
one
is
really
the
data
and
the
manipulation
of
the
data.
The
other
part
of
it
is
with
reproducibility
and
that
one
is
really
complicated.
There's
very
there's
levels
of
reproducibility
and
here's
a
classic
example
that
I
want
to
show
you
guys,
and
that
is
we
asked.
F
One
was
published
in
2008
and
one
was
published
in
2015
and
both
of
these
people
happen
to
have
used
the
exact
same
data
source
and
came
up
with
two
very
distinctly
different
answers.
One
came
up
with
zero
trend
losses
due
to
hurricanes,
and
one
came
up
with
136
million
a
year
increase
because
of
hurricane
losses.
F
So
you
know
that
that
asks
a
question
of
what
is
the
problem?
Was
it
that
they
actually
use
slightly
different
data?
Did
they
manipulate
the
data
differently,
or
did
they
just
have
different
techniques
in
looking
at
the
data,
and
it
was
more
a
coding
and
a
science
perspective
that
that
is
bringing
out
this
very
big
difference.
F
So
what
we
did
in
this
project
is,
we
did
a
very
small
experiment.
We
actually
did
it
twice
and
I'm
just
going
to
show
you
the
one
experiment.
We
did
an
experiment
saying
well,
if
you
have
two
scientists.
First
of
all,
I
want
to
sort
of
highlight
where
I
think
the
the
problems
are,
and
that
is
you.
You
would
have
two
scientists.
F
Looking
at
the
exact
same
question,
they're
both
doing
data
discovery,
they
both
develop
tools
and
they
both
analyze
the
data
and
typically
they're,
probably
going
to
spend
more
or
less
the
same
time
on
all
three
of
these
topics,
and
they
could
potentially
come
up
with
two
different
answers.
So
a
trend
versus
no
train-
and
we
have
a
number
of
times
said-
if
only
we
can
utilize
what
one
person
has
done-
hand
it
over
the
fence
to
another
person.
That
means
that
we
could
potentially
solve
a
number
of
things.
F
F
F
We
might
ask
identical
questions
for
different
parts
of
the
world
or
even
for
different
time
periods.
So
you
might
ask
a
question
of
what
a
tropical
storms
look
like
in
1960s
versus
what
tropical
storms
is
going
to
look
like
in
2015,
or
you
might
ask
the
question:
how
is
tropical
storms
going
to
look
in
2050
in
australia
versus
america?
F
F
The
other
part
is
because
we
ask
different
questions.
We
do
not
need
all
the
tools
that
has
been
developed
by
a
specific
researcher.
We
need
parts
of
the
tools
that
is
developed
and
we
very
often
as
far
as
tools
are
concerned.
We
either
do
not
share
and
it's
just
not
available.
So
we
do
code
it
up
or
one
of
the
things
that
we
found
with
the
earthquake
tools
is
that
there
there's
a
lot
of
tools,
but
they
package
it
and
nobody
really
wants
the
entire
package.
F
People
want
bits
and
pieces
and
modules
out
of
the
packages
rather
than
the
entire
package,
and
then
the
other
piece
that
is
very
important
is
the
fact
that
we
are
all
different
people.
So
our
skill
levels
and
computing
environments
are
different
and
even
if
something
works,
100
on
one
computer,
it
might
not
work
on
a
different
computer
or
somebody
is
maybe
coding
in
r
and
my
skills
are
in
python.
F
Understand
so
this
is
the
tools
that
was
developed,
so
our
asset
workflow
design
tool.
It's
available
on
this
website
that
I
have
up
top
here.
I
do
want
to
point
out
that
this
is
a
beta
version.
It's
probably
an
alpha
version
still
with
the
code
right
mike,
so
because
it's
a
pilot
project
and
we
do
not
have
funding
to
continue
developing
it.
It
is
going
to
stay
in
an
alpha
version
for
the
foreseeable
future,
but
basically
what
it
is.
F
So
I
just
want
to
give
you.
I've
got
two
two
or
three
more
slides.
I
want
to
give
you
a
quick
overview
of
what
we've
done
to
try
and
tease
out
where
the
really
big
bottlenecks
in
workflows
are.
So
what
we've
done
is
we've
asked
a
different
question.
We
asked
what
the
environment
associated
with
intensifying
hurricanes
are
and
we
wanted
to
look
at
them
in
australia.
F
F
They
were
also
given
the
ability
to
code
up
in
any
language,
so
we
were
really
trying
to
create
a
real
world
simulation,
all
the
documentation,
the
workflow
and
the
notebooks
was
handed
over
to
researcher
too.
What
we
did
here
to
simulate
a
more
real-world
environment
was
to
say
that
this
person
needed
to
look
at
current
and
future
climate
data,
and
they
must
use
data
from
a
climate
model.
So
the
data
is
different.
It
is
the
same
region
and
instead
of
using
surface
pressure,
they
needed
to
use
a
field
called
maximum
potential
intensity.
F
So
this
is
a
breakdown
of
the
time
that
they
spend
on
the
different
components,
and
I
want
to
point
out
quite
a
few
things
on
the
slide.
First
of
all,
during
conversations
and
even
thinking
of
myself,
documentation
always
seems
to
be
a
really
hard
thing
to
do,
because
we
always
do
it
at
the
end
of
the
project.
F
Both
of
the
people
spent
a
very
small
amount
of
time
on
documentation
and
they
did
hand
in
a
lot
of
good
documentation
for
this
project,
and
they
both
also
said
that
they
were
surprised
that
they
did
not
spend
as
much
time
on
documenting
their
workflows
as
they
as
it
felt
like,
while
they
were
doing
it
and
as
they
thought
it
would
be.
So
I
think
this,
oh
it
always
takes
too
much
time
to
do.
It
is,
is
probably
something
that
we
could
could
try
and
get
people
to
understand
that
it
really
doesn't.
F
There
were
times
that
researcher
b
simply
wanted
to
give
up
and
actually
said,
hey.
Can
I
please
just
write
my
own
code
because
it
would
be
easier
when
we
analyzed
that,
in
the
end
of
the
day
researcher,
I
spent
13
of
this
if
we
compared
the
same
amount
of
time
between
the
two
researchers
spending
30
of
that
time,
figuring
out
the
workflow
and
figuring
out
a
new
language
which
was
still
even
adding
that
up
with
co-development
was
significantly
less
than
it
was
writing
the
code
from
scratch.
F
There
was
a
27
time
save
in
for
researcher
b
and
on
the
overall
time
that
this
research
has
been
more
almost
half
of
the
the
time
that
they
used
on
analysis.
But
if
you
take
this
27
percent
out,
they
spend
two
thirds
of
the
time
on
science.
So
we
we
didn't
quite
get
the
80
20
flip,
but
we
really
got
close
in
terms
of
that.
F
The
other
piece
that
we
specifically
asked
them
to
do
is
as
they
document
they
needed
to
make
sure
that
each
one
of
these
pieces
that
they
document
was
a
module
on
their
own,
and
the
report
back
that
we
got
was
that
that
really
made
a
big
difference.
It
wasn't
one
big
module.
It
was
pieces
of
modules
that
they
can
pull
out
what
they
needed.
F
This
is
an
opportunity
to
create
notebook
and
containers
that
would
help
with
the
skill
level
that
is
different
between
users,
as
well
as
just
the
fact
that
we
don't
all
have
the
same
computer
language
behind
us,
and
if
we
can
put
things
in
containers,
we
can
move
containers
across
the
face
which
would
be
much
better
than
just
moving
scripts
across
the
pants,
and
with
that
I
would
leave
it
to
mike.
Do
you
have
anything
extra
to
add.
G
Yeah,
all
I
would
say
is,
as
cindy
initially
mentioned
it.
This
is
a
this
is
a
pilot
project.
That's
funded
through
earth
cube,
and
so
the
problem
that
that
earthquake
has
is,
you
know
we
have
a
very
diverse
science
community
and
many
members
of
the
science
community,
they're
small
research
teams
and
so
we're
trying
to
get
a
handle
on
what
are
the
common
gaps
that
these
research
teams
have,
and
so
we
viewed
a
workflow.
You
know
workflow
sketching
tool,
it
outputs
json,
for
example.
G
At
the
end,
you
know
having
having
the
scientists
to
design
and
develop
their
workflows
with
a
very
simple
tool
can
create
a
library
of
these
that
we
can
then
mine
for
information.
Oh
while
everybody's,
having
or
using
this
data
set
or
using
this
tool
to
kind
of
get
a
an
idea
of
the
landscape
and
so
cindy
as
a
scientist
did
or
did
a
deep
dive
on
on.
You
know
her
particular
area
of
science
using
this
tool
for
this
express
purpose
of
improving.
G
C
G
They
get
graduate
students
coming
in
all
the
time
and
they
have
to
explain
you
know
their
processes,
which
of
course
change,
but
they
have
no
kind
of
basis
in
small
university
departments
to
to
you
know
kind
of
describe
their
their
their
work.
So
at
a
high
level,
that's
kind
of
what
we
looked
at
for
this
project
that
we
just
we
just
got
funding.
You
know
to
do
this
as
a
pilot,
and
so
therefore
we
you
know
cindy
was
able
to
do
a
real,
deep
dive
on
a
particular
science
area,
so
that
helps.
A
He
does
sort
of
neurological
interpretations
of
fmri
data
coming
out
of
his
open
fmri
initiative
and
fed
that
to
several
people
with
existing
tools
and
existing
backgrounds,
and
they
too
came
up
with
very
different
results.
In
fact,
I
mean
it's
a
nature
paper
that
that
you
know
it's
fantastic,
because
what
it
means
is
reproducibility
isn't
happening
and
it's
a
real
challenge
and
a
real
problem.
So
I'm
glad
to
see
people
talk.
A
A
Everybody
else's
software
is
the
right
way
to
go,
even
if
you
have
to
learn
a
different
language
for
it,
and
so
I'm
I'm
glad
to
hear
that
you're,
seeing
what
I
sort
of
glommed
on
to
and
I
think
other
people
who
are
successful
at
it.
It
doesn't
matter
what
language
it's
written
in,
but
the
other
comment
I
was
going
to
make,
though,
is
that
you,
you
pointed
out
that
the
data
discovery
portion,
which
we
focus
a
lot
on
and
is
an
important
thing,
especially
given
how
much
data
is
out
there.
A
You
know
filtering
out.
The
wheat
from
the
chaff
is
important,
but
a
lot
of
that
data
discovery
portion
is
also
a
data
trust
portion
of
what
you
do,
and
I
don't
think
you
can
ever
get
rid
of
that.
You
know.
That's
if
you
hand
me
a
bucket
of
data,
I'm
going
to
spend
a
long
time,
even
if
I
can't
do
anything
else,
you
know
use
any
other
data,
I'm
still.
I
still
have
to
come
up
with
that
and
that
to
me
that
my
gut
says
that's
10
of
your
time.
A
F
A
G
I'm
curious
related
to
data
discovery
and
I
totally
agree
with
what
you
say
not
to
mention
things
like
uncertainty.
What's
the
uncertainty
of
these
data,
you
know
the
the
domain.
Scientists
know
that,
where
the
where
the
people
that
just
picked
the
data
it
may
not,
but
are
the
hubs
using
this
schema.org
mechanism
for
some
I
know
it's
just
a
rudimentary
data
discovery
right
now
in
terms
of
the
types
of
data
stored
in
the
repository,
I'm
just
wondering
if
that's
permeated
through
these
to
these
larger
data
centers
that
you're
representing
here.
E
So
if
it's
okay,
gents
I'll
I'll,
respond
to
that,
so
the
hubs
actually
don't
have
repositories.
The
hubs
were
developed
to
build
public
private
partnerships
and
cross-sector
capacity
and
research
helps
support
research
development
for
great
just
to
to
find
challenges
in
each
of
the
regions.
So
there
is
actually
no
infrastructure,
that's
funded
and
run
out
of
the
hubs.
Okay,.
A
So
I
see
some
places
where
things
like
schema.org
do
come
in
to
play,
but
there's
a
lot
out
there,
where
I
think
you
know
it's
the
same
sort
of
thing:
it's
a
data,
trust
issue
and
there's
always
going
to
be
data,
translation
and
data
pieces
around
that.
I
think
that
falls
into
that
same
10
of
the
data.
A
G
Yeah,
I
would,
I
would
just
say,
in
earth,
cube
they
they've
they've
really
streamlined
the
implementation.
I
think
they've
done
a
good
job.
You
know
the
folks
that
are
doing
that
work,
I'm
not
one
of
them,
but
I'm
just
aware
of
it.
I'm
actually,
the
chair
of
the
leadership
council
for
earth
cube,
so
I'm
more
on
the
government
side,
but
I
think
they've
done
a
good
job
in
making
the
overhead
to
doing
you
know
it's
basically
some
json
ld
code
that
you
add
to
your
website.
G
A
Sorry
I
had
to
do
an
emergency
door,
try
and
keep
the
dog
from
barking
in
here.
Oh
absolutely,
so
well
we're
a
little
bit
over
six
minutes
over,
but
I
just
want
to
make
sure
there's
not
anything
else
that
other
people
want
to
talk
about.
But
if
not
that
long
pause
I
will
go
ahead
and
say
well.
Thank
you,
everybody
for
coming
and
thank
you
for
a
very
nice
talk
and
I
think,
there's
some
interesting
many
interesting
pieces
around
that
I
know
you're
working
with
from
usc
as
well.