►
Description
Sept. 10th, 2021
Moderator: Niall Gaffney
A
A
Over
the
coming
year,
we
had
a
nice
little
experiment
last
spring,
where
we
started
to
have
more
sort
of
topical
panels
coming
and
and
start
to
talk
with
folks
to
look
not
just
at
sort
of
presenting.
You
know
current
existing
infrastructures,
but
to
talk
about
the
needs
and
and
the
utility
of
different
things
put
together
in
in
ways
to
try
and
sort
of
build
out
some
community
around
these
first
one
that
we
did
involved
jay
and
a
few
other
folks
around
sort
of
the
cyber
security
topic.
A
We
had
sort
of
two
meetings
about
that
and
I
think
it
was.
It
was
quite
successful
in
that
there
are
conversations
and
things
happening
on
all
of
that
now,
and
I
think
you
know
that's
sort
of
what
we're
looking
for,
but
before
we
start
getting
into
all
of
that,
I
think,
as
usual,
I
look
around.
A
I
think
I
know
most
everybody,
but
it's
probably
a
good
idea
for
us
to
just
do
sort
of
a
round
of
introductions
for
those
of
you
who
don't
know
me,
I'm
niall
gaffney,
I'm
the
director
for
data
intensive
computing.
My
background
is
astronomy
and
astrophysics,
but
largely
around
archival
science
and
and
data
workflows
and
data
analysis.
A
So
I
previously,
prior
to
coming
to
tac
eight
years
ago,
I
was
working
for
the
folks
who
run
the
operations
for
the
hubble
space
telescope
and
a
few
other
facilities
that
are
orbiting
the
earth
about
300
miles
up
and
have
a
lot
of
background
in
just
you
know
how
to
make
data
workable
and
so
that's
sort
of
where
I
come
from,
but
you
know
enjoying
working
in
a
lot
of
other
fields.
Now
taking
the
experiences
I
have
and
learning
new
things.
B
Yeah,
why
not
here
on
loan
yep,
I'm
here
on
loan,
I'm
filling
in
for
christopher
daly,
I'm
a
project
manager
at
renzi.
I
work
with
the
big
data
hubs,
a
good
bits,
so
I
see
some
familiar
faces.
A
Thank
you,
john.
B
Sure
I'm
john
goody,
I'm
the
executive
director
of
the
mgh
pcc,
I'm
the
steering
committee
for
this
merry
band
now
I'll,
just
a
little
connection
niall
my
father
when
he
worked
for
the
navy
was
responsible
for
moving
the
hubble
space
telescope
from
california
to
to
florida
before
it
was
launched
really
a
little
bit
of
history.
Yeah.
A
Well,
they're
doing
that
with
the
web
telescope
by
by
boat
as
well,
going
through
panama,
so
yeah,
and
there
there
may
or
may
not
be
some
extra
security
on
that
one
right
people
hear
the
words
gold-plated
mirror.
Anyway,
that's
cool
all
right
and
of
course,
everything
shifted
here,
but
I'll
go
with
the
next
person
anton
yeah.
B
Hello,
everyone,
my
name
is
anthony
and
I'm
a
senior
hpc
engineer
at
the
innovation
university,
supporting
the
research
computing
infrastructure,
large
cluster,
so
my
background
is
mostly
the
large
data
and
data
transfers,
plus
I'm
leading
the
cyber
security
efforts
to
improve
around
our
instrument.
Thank
you.
A
Cool
christine.
C
Hi
christine
kirkpatrick,
I'm
at
the
san
diego
super
computer
center,
I
mean
usually
I'm
at
my
house.
Obviously
my
spare
bedroom,
I'm
an
infrastructure
and
a
data
person
and
I'm
a
pi
at
the
the
west
hub,
among
other
things,
my
personal
research.
These
days
is
in
how
to
make
data
ai
ready
in
an
efficient
way,
maybe
leave
it.
There.
D
Thank
you
now
hi,
I'm
jay,
I'm
professor
yen
in
computer
engineering
at
rit,
rochester
institute
technology,
I'm
also
the
director
of
global
outreach
for
global
sciences
institute
at
rit.
So
thank
you
now
for
talking
about
the
earlier
workshop
panel
thing
that
we
did
back
in
spring.
So
since
then
there
were
a
few
different
things
going
on
around
the
world.
They
include
a
workshop
that
sagar
now
we
organized
in
kde,
and
there
was
another
one
that
was
organized
by
university
of
virginia
indian
university,
including
vaughan
and
george
washington
university.
D
That's
talking
about
data
curation
for
cyber
security,
research
and
education.
There
was
a
few
others,
ai
ml
for
cyber
security
workshops
happening
over
the
summer
time,
and
I
think
that
we'll
continue
to
see
more
and
more
of
that
happening
and
then
going
on.
So
thank
you
for
inviting
me
back
to
this
working
group
and
look
forward
to
continue
contributing
to
to
this.
The
mission.
A
Sounds
great
and
yes,
no
invitation
needed,
everybody
can
just
come
yeah,
so
lauren
you
want
to
introduce
yourself.
E
Thank
you,
I'm
lauren
close.
I
actually
work
with
the
northeast
big
data.
Innovation
hub
I've
only
been
in
this
position
for
about
four
or
five
months,
so
I'm
just
kind
of
introducing
myself
to
different
groups,
but
I
work
mainly
in
operations
and
communications
for
the
northeast
hub.
F
Hi,
jim
william
bush,
I
direct
research
computing
at
the
university
of
minnesota,
and
that
includes
the
minnesota
super
computing
institute,
as
well
as
youth
spatial
and
our
informatics
institute.
My
research
is
in
phylogenetics
and
I
still
do
do
that
periodically
in
the
evenings
and
on
weekends.
E
I
am
working
with
some
folks
here
in
to
bridge
the
gap
between
academia
and
the
the
real
world
or
the
industrial
world.
E
I
am
a
cyber
security
data
science
and
data
expert,
so
looking
forward
to
working
with
everyone.
G
Hi
everybody
I'm
melissa,
craigan.
I
work
at
sdsc
with
christine,
formerly
head
of
the
midwest
big
data
hub,
I'm
part
of
a
few
other
projects,
I'm
also
working
on
ai
readiness,
but
from
an
organizational
perspective,
and
and
also
looking
at
federal
data
decision
making
in
ai
readiness
as
well.
I
think
that's
all
I'll
say
for
the
moment
I
I
will
later
on.
A
Excellent
looking
forward
to
it
all
right
well
welcome
everybody.
I
know
you
know
it's
a
little
earlier
on
a
friday,
but
it's
still
friday
after
a
week,
a
compact
week
of
zoom
meetings
for
all.
So
thanks
for
joining
today,
just
a
couple
of
things
before
we
start
talking
about
the
the
new
direction
on
all
of
this.
So
we
we
do
want
to
get
as
much
input
from
folks
about
topics
and
and
other
pieces
that
that
would
be
useful
to
for
us
to
understand.
A
You
know
what
what
potential
topics,
what
potential
things
we
could
do,
there's
a
google
form
that
has
been
circulating
for
a
little
while
folks
haven't
seen
it.
Maybe
we
can
send
out
another
reminder:
email
with
that
to
the
list,
but
jeremy
just
went
ahead
and
posted
the
posted.
The
shortened
url
there
in
the
chat.
A
So
if
you
want
to
right
now,
go
click
and
do
it
rather
than
listen
to
me
drawn
on
about
it,
we'd
like
to
hear
from
folks
on
that
and
also
just
circulate
it
around
to
other
people
who
would
be
interested,
who
you
know,
either
as
somebody
who's
interested
in
either
presenting
or
hearing
about
specific
topics
and
how
they
can
get
to
be
involved
in
these.
A
I
really
do
want
to
try
and
make
this
not
just
about
the
providers,
but
also
bring
in
some
of
the
consumers
of
all
of
this
stuff,
because
that's
where
the
that's,
where
the
real
magic
and
the
in
the
combination
of
conversations
usually
take
place.
A
A
The
very
end
of
zoom
fatigue
on
fridays,
which
I
know
is
real
trust
me,
but
we
want
to
you,
know,
make
it
more
convenient
for
people
and,
as
I
mentioned
to
christine,
make
it
so
that
I'm
the
one
who's
missing
lunch
when
I
come
as
opposed
to
others.
So
you
can't
pick
a
perfect
time
so,
but
the
topic
really
for
today
is
that
we
want
to
hear
from
folks
about
suggestions
or
recommendations
or
just
talk
about
what
sort
of
topics
we
want
to
cover
in
the
coming
year.
A
So
over
the
summer
there
was
sort
of
some
back
and
forth
in
emails,
and
so
we
wanted
to
have.
You
know
something
at
least
set
up
for
october
on
this,
and
I
know
there's
a
lot
of
good
stuff
going
on
around
data
archive,
and
this
has
sort
of
been
the
true
cold
storage
and
real
archiving,
not
not
just
sitting
on
disks
somewhere
but
a
real
preservation.
A
I
believe,
and
so
we've
got
three
speakers
who
are
going
to
come
in
and
talk
about
the
systems
and
the
work
that
they're
doing
around
these
things
in
in
different
areas,
from
high
energy
physics
to
large-scale
tape,
archives,
and
I
think.
A
Talk
about
how
how
what's
working,
what's
not
working
and
and
what
we
need
to
do
to
be
able
to
do
a
better
job
of
of
something
that
I
think
people
don't
appreciate
how
complex
it
is
outside
of
the
people
who
actually
wind
up
doing
this
archiving.
It's
not
just
sticking
something
in
a
closet,
so
be
nice
to
hear
from
folks
we're
going
to
do
sort
of
a
follow
up
sort
of
cyber
security
panel.
A
Florence
is
going
to
help
talk
with
with
vaughn
and
some
other
folks
to
sort
of
get
that
going
and
maybe
hear
about
the
pieces
that
jay
was
talking
about
as
well,
so
we
can
touch
base
back
in
that
in
november
and
then
looking
at
december,
they've
got
right
now
sort
of
a
microscopy
panel.
I
know
we've
got
some
interesting
things
going
on
in
many
different
areas
around
around
different
forms
of
microscopy
and
imaging,
which
would
be
a
nice
panel
to
have.
But
I
don't
know
we
don't
haven't.
A
We
we
haven't
got
a
panel
yet
on
that,
and
so
I
think
we're
sort
of
open
to
discussions
on
that
and
what
we
really
would
like
to
do
with
these
is
to
see
if
we
can
do
the
same
sort
of
thing
that
we
did
with
with
jay's
group,
which
is
to
have
a
discussion
as
the
panel
and
then
maybe
a
follow-up
afterwards
to
start
to
let
people
you
know
once
once
the
talks
have
sunk
in
give
people
a
chance
to
you
know,
talk
to
one
another
afterwards
and
and
start
those
collaborations
or
the
beginnings
of
those
moving
forward
sort
of
not
necessarily
taking
place.
A
You
know
in
the
this
this
time
scale,
but
but
you
know
moving
out
into
other
sort
of
breakout
areas
and
then
hopefully
reporting
back
to
us
about
what's
going
on
and
keeping
us
informed
on
things.
So
so
with
that,
I
will
stop
droning
on
about
all
this
and
open
up
the
floor
and
since
melissa
already
said
she
had
some
pitches.
Why
don't
I
just
go
ahead
and
toss
it
over
to
you,
melissa.
G
Great
thanks
niall
so
I
have
an
eager
award
from
nsf
to
look
at
the
to
look
at
the
public
access
repository
with
respect
to
well
to
to
think
about
community
readiness
for
depositing
metadata
records
on
data
and
and
other
related
products
into
the
nsf
public
access
repository.
G
That
policy
is
anticipated
shortly
to
to
open
up
a
first
voluntary
but
then
later
required
deposition
of
metadata
records
into
the
par
on
these
other
kinds
of
products,
and
so
I've
been
I've
been
developing
just
a
short
presentation
on
the
public
access
repository
and
then
doing
some
discussion
with
community
groups
so
that
that
can
inform
sort
of
what
kinds
of
materials
would
be
fruitful
for
both
community
engagement
and
and
awareness
around
these
things.
G
So
you
know
moving
toward
first
order
products
beyond
publication,
so
anyway,
I'd
love
to
do
that,
maybe
maybe
december
or
or
february.
You
know
I'm
open,
but
but
a
bit
later
in
the
year
would
be
fine.
The
other
is
that
I
I
had
a
call
recently
with
lynne.
G
Prichard
at
brookhaven
she's
been
working
on
computational
reproducibility
for
a
couple
of
years
now,
and
I
know
that
we've
got
some
other
folks
in
in
our
sort
of
colleague,
colleague
group
here
who
are
working
on
on
reproducibility
as
well,
and
I
was
thinking
a
a
nice
webinar
with
a
couple
of
people.
Doing
computational,
particularly
for
machine
learning.
Ai
would
be
would
be
great
because
it
directly
affects
you,
know:
management
of
software
and
workflows
and
and
even
sometimes
hardware
setup.
A
I
think
both
very
good
topics-
reproducibility,
is
just
such
a
huge
topic,
it's
hard
to
cover
in
an
hour.
It's
just
you
know
just
just
saying
software
reproducibility
is
hard
enough,
but
I
think
that
that
would
be
a
fantastic
one
coming
up
and-
and
I
think
yeah
it
would
be
good
for
us
to
hear
more
about
the
public
access,
repo
and
and
and
how
how
we
can
make
this
most
useful
for
folks.
So
I
think
those
both
be
very
good
topics.
B
B
To
bring
up
in
a
different
domain
I'll
end
up
putting
jim
on
the
spot
jim
in
week,
had
an
email
exchange
about
the
work
you're
doing
with
cask
on
research
priorities,
and
I
think
the
two
of
us
we're
thinking
that
that
might
be
the
risk
for
for
a
panel
session
at
some
point
in
the
first
half
of
next
year.
Is
that
still
striking
was
a
good
idea,
and
if
so,
we
can
elaborate
on
it.
F
Yeah
it
does,
I
was.
I
was
looking
back
at
our
sort
of
email,
email
thread,
thinking
that
we
should
raise
that
as
well
and
and
and
think
about.
F
You
know
how
how
to
do
that
in
a
way
that
perhaps
does
a
good
job,
at
least
at
capturing
the
diversity
of
entities
out
there
who
are
working
on
some
of
these
top
priorities
might
might
be
approaching,
and
this
is
where
it
would
be
great
to
get
the
sort
of
thoughts
from
this
group
as
to
you
know
how
to
how
to
bring
focus
to
something
and,
and
so
just
to
put
a
little
more
toward
this.
So
you
could
maybe
comment
effectively
on
it
is,
is
what
cask
is
doing?
F
Is
we
pulled
our
our
90,
some
odd
members
and
had
an
outstanding
response
from
them
to
better
understand
what
are
sort
of
the
top
things
that
keep
them
up
at
night,
and
from
that
we
developed
a
surprisingly
tight
list
of
four
things
that
that
keep
people
up,
and
so
there
was
just
a
tremendous
amount
of
congruence
with
respect
to
our
members.
F
Answers
to
this
poll,
and
and
that's
that's
useful
for
us
as
an
organization
representing
these
groups,
because
we're
going
to
put
in
our
fall
meeting
a
little
bit
more
meet
on
these
four
things
by
having
some
working
groups
within
that
meeting,
to
help
us
make
sure
that
we're
we're
we're
all
on
the
same
page
about
what
it
means.
What
these
top
priorities
mean
to
our
member
organizations.
F
But
you
know
what
could
we
do
with
that
information
in
terms
of
then,
you
know
getting
getting
other
groups
to
together
to
to
comment
on
this
and
put
more
focus
and
attention
on
these
areas.
B
F
F
Yeah
it
does,
I
can.
I
can
give
you
the
the
at
least
running
titles,
of
all,
four
compliance
and
cyber
security.
No
surprise
there
tops
the
list,
support
for
team
science,
slash,
end-to-end
workflows,
so
that's
one
one
topic,
and
so
the
challenges
associated
with
supporting
team
science
or
end-to-end
workflows
is
number
two
keeping
staff
skill
sets
sharp
addressing
diversity
needs
among
our
staff.
Retaining
staff
recruiting
staff
is
number
three
number
four
is
trying
to
improve.
F
Alignment
of
federal
funding
with
the
research
community
needs,
in
other
words,
with
needs
on
our
campus,
and
sometimes
you
know
at
a
at
a
higher
level,
there
seems
to
be
challenges
sometimes
to
find
the
right
funding
for
the
things
that
we
think
are
priorities
within
our
respective
institutions.
F
So
those
are
the
top
five
priorities
again,
like
I
said
those
really
stood
out.
There
are
there's
a
bucket
that
we
have
full
of
other
things
that
are
really
important
priorities
as
well,
but
there
just
wasn't
as
much
agreement
around
those
and
and
we're
going
to
be
treating
those
in
terms
of
our
position.
Paper
is
more
of
these
are
emerging
needs.
These
are
also
important
things,
but
you
know
there
wasn't
as
much
consensus
as
far
as
our
membership
was
concerned
as
there
as
there
was
in
the
the
top
four
that
I
just
mentioned.
A
I
don't
know
I
mean,
but
I
can
clearly
see
doing
you
know
sort
of
the
first
two
where
the
two
that
sort
of
seem
to
align
more
with
you
know
the
data
sharing
with
you
know
clearly
around
cyber
security
and
and
compliance,
but
but
I
think
the
the
team,
science
and
workflows
might
be
interesting,
but
then
maybe
even
aligning
them
or
having
them
aligned
into
some
of
the
other
discussions
that
we
have
so
maybe
have
somebody
you
know
talked
to
what
the
cask
you
know
what
what's
coming
out
of
cask
and
and
those
sort
of
things
is
part
of
the
cyber
security
panel.
A
But
then
maybe
the
microscopy
panel,
we
can
talk,
you
know
I'm
just
thinking.
Maybe
we
could
do
then
have
somebody
or
some
focus
on
you
know.
If
there
are
particular
topics
I
don't
know
if
microscopy
is
probably
not
right,
but
there
are
others
where
team
science
would
come
in
to
play
and
and
talk
about.
You
know
the
shortcomings
and
and
and
what
we
may
want
to
start
to
look
for
proposals
around.
F
Well,
microscopy
is
a
really
good
example
of
the
of
the
second
sort
of
big
bucket.
F
That's
a
that's
a
that's
an
end-to-end
workflow,
where
we
often
find
that
you
know
our
respective
centers
are
asked
to
do
this
piece
right
and
sometimes
the
only
information
that
we
get.
F
Is
this
piece,
but
then,
as
we
sort
of
peel
the
onion
as
we
as
we
like
to
say,
we
discover
that
it's
really
connected
to
something
much
larger
and
if
we
knew
from
the
onset
what
that
larger
thing
was,
we
could
actually
be
much
more
effective
in
terms
of
developing
the
end-to-end
workflows
and,
and
that's
why
we
we
led
with
the
idea
of
team
science,
because
what
we've
discovered,
at
least
amongst
the
you
know,
the
the
working
group
on
this
on
this
topic,
is
that
it
really
is
sort
of
that
one
of
the
challenges
is
sort
of
bringing
together.
F
B
But
that
does
yeah
jim,
I
think,
take
us
full
circle
to
your
original
suggestion,
which
was
to
have
someone
representing
the
perspective
of
the
large
cast
community
in
certain
areas
where
the
first
one
might
have
been.
The
long-term
storage,
but
by
I
agree
microscopy
is
one
that
fits
squarely
in
the
team
science
domain.
A
Well
and
hearing
you
sort
of
frame
it
that
way
as
well.
I
mean
the
I
you
know.
We
just
finished
up
here
this
four-year
darpa
project
around
synthetic
biology,
which
essentially
you
know
in
the
end.
What
we
wound
up
having
to
do
was
to
make
sure
that
we
captured
the
entire.
You
know
you
called
it
the
onion.
You
know
I
I
think
about
it
as
as
the
circle
of
experimental
science.
A
You
know
there's
intent
and
then
there's
what
happens
and
then
there's
aligning
all
of
that
information
because
nobody
can
ever
agree
on
on.
You
know
how
to
describe
even
something
as
simple
as
water
in
chemistry
experiment,
but
you
know
data
normalization
and
then,
and
then
you
know
actually
then
having
common
products
that
are
useful,
especially
in
an
ai
world,
which
was
what
you
know.
That
was
a
lot
of
regression
going
on.
In
the
end,
those
sort
of
things
would
be
useful
to
sort
of
expose.
A
I
think,
and-
and
you
know
also
to
me-
it'd-
be
really
interesting
to
to
make
sure
that
everybody
understands
that.
That's
that's
what's
needed,
it's
not
just
oh
okay!
Then
we
go
use.
You
know
we
go
use
tac
for
this
step
here,
because
it
involves
computers
right.
B
They
didn't
they
ended
up
founding
a
company,
but
but
that
that
notion
and
that
notion
of
this
branch
of
biology
moving
from
research
into
engineering
where,
instead
of
discovery,
you're
actually
trying
to
manipulate
the
world
as
a
both
a
cool
theme
and
and
one
that
fits
totally
in
our
wheelhouse.
I
think.
A
Yeah,
I'd
have
to
make
sure
that
I
mean
it's
darpa
and
it's
not
that
anything
secret,
but
you
know
I.
I
would
have
to
make
sure
that
that
we
can.
You
know
we
can
talk
about
the
details,
but
the
actual
flow,
and
how
we
did
things
are,
is
definitely
something
I
can
talk
to.
You
don't
want
an
astronomer
talking
about
the
biology
trust
me,
but
I
wasn't
suggesting
that
we
do
that,
but.
A
Well
and
then
I
think
that
also
dovetails
quite
well
into
melissa's
reproducibility
for
mlnai,
and
so
I'm
I'm
sensing
a
spring
here
of
some
level.
A
D
A
D
Interrupt
because
the
reproducibilities
of
the
concern,
not
all
the
concerns
on
the
workflow,
I
think
this
is
one
thing
that
I
haven't
heard
a
lot
of
conversation
in
this
group.
So
if
this
is
not
of
interest,
then
don't
worry
about
it.
It's
more
about
you
know
from
the
cyber
security
or
even
espionage
perspective,
that
they
are
more
intellectual,
property,
thefts
or
agents
or
spies.
Really
that
doing
a
lot
of
things
through
this
data.
D
So
there
has
been
conversation
in
open
source
community
and
cyber
security
open
source
community,
particularly
thinking
about
that,
what
cut?
What's
the
trade-off
between
making
everything
open
source
reproducible,
which
is
great,
I'm
academia?
I
like
that.
At
the
same
time,
fbi
agents
is
saying
that
be
careful
and
other
people
from
certain
countries
when
you're
coming
here
to
steal
your
stuff
and
not
necessarily
coming
physically
coming
to
your
domain.
So
so
I
I'm
curious
whether
this
this
group
has
talked
about
whether
the
workflow
right,
the
reproducibility
and
how's.
D
That
trading
off
with
concerns
from
let's
say
fbi
is
that
of
interest
for
this
group
or
not,
and
as
a
researcher
I
I
am
curious
about
it
right.
So
I
won't
put
my
stuff
on
open
source
domain.
I
want
to
put
the
data
there.
I
want
to
put
the
metadata
there
and
I
won't
have
a
workflow
for
my
student.
That
works,
but
am
I
concerned
about
someone's
going
to
knock
on
my
door
saying
that
hey
your
process
is
not
appropriate.
D
E
So,
let's
not
let's
not
confuse
open
source
and
reproducibility
right,
so
reproducibility
is
that
you
can
give
someone
else.
Your
results
or
the
way
that
you
did
it
and
they
can
do
the
same
thing
right,
open
source
means
that
I
let
the
whole
world
see
that.
So
I
think
reproducibility,
even
if
you're
not
going
to
open
source,
is
still
important
because
whoever
you're
collaborating
with
you
know
whether
it's
the
business
or
whether
you're
other
collaborating
with
other
researchers.
E
You
still
need
to
be
able
to
have
that
element
of
reproducibility
and
understanding
of
how
you
got
the
results
that
you
got
so
that
someone
else
could
possibly
do
it.
So
I
I
do.
I
do
see
that
concern
though
jay
throughout
throughout
the
industry.
Like
nobody,
none
of
my
customers
want
me
to
open
source
their
their
ai
that
we're
doing
for
cyber
security.
It's
pretty
hush.
B
Well,
it's
a
and
that
I
agree
with
everything
you
said
lauren
and
then
my
overlay,
that
that
jay
maybe
brought
up
a
interesting
panel
topic,
which
is
you
know,
different
points
of
view
on
open
source,
open
science
versus
hiding
stuff.
A
A
Into
you
know,
yeah
there's,
there's
sort
of
the
security
on
that,
but
then
there's
there's
also
the
portion
around
especially
health
data,
and
that's
the
you
know,
everybody
thinks
of
hipaa
as
being
the
compliance
part,
but
there's
a
lot
of
other
pieces
a
lot
of
other.
A
I
just
call
it
hair
that
hangs
around
that
as
far
as
the
code
or
how
you,
how
you
put
things
together
for
different
systems,
which
yeah
that
have
nothing
to
do
with
reproducibility
until
you
get
it
translated
into
a
layer
where
you
can
talk
about
it
reproducibly,
you
know
it.
B
Might
even
be,
it
probably
is
an
entirely
other
panel,
which
is
how
do
you
navigate
privacy
regulations
successfully.
F
That
tension
is
what
you're
speaking
to
between
you
know
that
that
I
think
is
is
felt
very
much
at
the
universities,
where
everything
is
supposed
to
be
generally
made
available
and
very
few
barriers
to
accessing
data
and
code
and
so
forth,
while
at
the
same
time
there
is
a
very
real
need
to,
in
many
cases,
protect
the
privacy
of
this
data,
and
that
tension
is
very
often
overlooked
in
our
policies
that
we
have
and
data
data
and
openness
policies
that
many
universities
have
that
are
required
by
our
board
of
regents
or,
however,
you're
governed
and
the
reality
that,
when
we
work
with
the
private
sector
not
exclusively
pointing
to
private
sector,
but
very
often
when
we
work
with
private
sector
that
they
do
have
different
expectations
in
terms
of
how
that
how
the
data
will
be
used
and
so
forth.
F
And
it's
it's
not
always
in
a
way
right
that
we
expect.
For
example,
we've
worked
with
very
large
companies
developing
code
that
they
were
explicit
about
making
public
as
rapidly
as
possible,
which
was
interesting
right.
Why?
Why
would
you
do
that?
And-
and
the
answer
is
you
know
getting-
that
into
the
public
domain-
serves
their
purpose
of,
not
being
you
know,
in
a
position
of
paying
royalties
on
something
that
someone
later
actually
captured
and
made
into
a
proprietary
package,
and
so
we've
had.
F
You
know
interesting
engagements
with
private
sector
that
have
both
involved.
You
know
really
keeping
things
secure
to
having
a
very
accelerated
path
to
making
them
public
more
so
than
actually
some
of
our
academic
colleagues,
who
kind
of
want
to
make
sure
that
the
publication
gets
out
there.
First
they're,
often
big
into
the
pre-publication,
to
make
sure
that
they
don't
again
pay
royalties
to
someone
who
makes
it
proprietary
later.
D
Jim,
I
think
that's
a
very
good
point,
so
I
I
thank
you
for
all
of
you
sharing
that
and
jim
sharing
that
the
story,
I
think
a
lot
of
this
is-
will
be
a
good
panel
conversation
in
my
opinion
that
helps
researchers,
practitioners
to
think
about
what's
their
role.
I
think
this
go
back
to
the
workflow
workflow,
not
just
about
data,
but
also
about
source
code.
D
What
what
a
good
practice
is
to
avoid,
in
my
personal
opinion,
not
lawyer
or
fbi
agent
right,
but
in
my
personal
opinion,
that
sharing
it
out
quickly
is
a
good
point,
because
I
think
overprotective
of
a
lot
of
these
things,
I'm
not
talking
about
private
personal
information.
Talk
about
protected
of
of
some
of
these
algorithms
or
innovation.
Advances
is
actually
making
it
harder
for
the
good
guys
to
make
it
easier
for
the
bad
guys
following
the
law
and
not
to
do
anything
about
it.
F
So
yeah
I
I
like
I
like
the
idea
of
thinking
about
this
in
the
context
of
team
science
and
workflows,
it
gets
a.
It
could
be
a
little
bit
difficult
to
manage,
but
maybe
that
could
be
a
way
in
which
we
sort
of
put
context
around
some
of
these
subjects
that
might
be
worthy
of
a
panel
by
themselves.
F
There's
that,
beginning
stage
of
you
know
getting
the
data
from
the
instrument
you
know
to
to
either
researchers
I
mean
to
researchers
in
some
form
that
they
can
begin
to
do.
Analyses
is
a
whole
subject
that
we
spend
a
lot
of
time
on
I'm
sure
everyone
here,
looking
at
various
aspects
of
technology
and
and
compliance
around,
and
so
then
there
are
other
issues
in
terms
of
data
sharing
and
so
forth
that
get
into
some
of
these
compliance
and
security
issues.
I
wonder
what
I'm
getting
at
is.
F
B
A
A
A
You
know,
can
I
can
I
rerun
the
experiment
versus
just
recomputed.
You
know,
and
I
think
all
of
that
that's
why
it
keeps
it
feeds
into
a
lot
of
these
things.
So
probably
it
would
be
good,
I
think,
to
start
off
with
that,
but
then
maybe
go
into.
A
You
know
workflow
and
metadata
normalization,
which
I
think
is
you
know,
that's
that's
a
big
deal
here
for
for
data
sharing
and
you
know,
we've
had
many
people
talk
about
individual
systems
to
do
that,
but
to
have
best
practices,
exposed
or
or
or
ways
to
try,
and
do
you
know
a
better
job
of
not
having
to
reinvent
the
wheel?
Every
time
would
be
good.
A
So
you
know,
I
think
that
might
be.
What
are
the
challenges
to
that?
I
could
see
that
coming
in
yeah
that.
F
That
that's
a
really
good
way
to
look
at
it.
Denial,
I
think,
like
maybe
just
one
thing
would
be
to
try
to
identify
the
challenges
associated
around
around
data
management
with
these
end-to-end
workflows
or
the
team
science
that
we're
addressing,
because
that's
gonna,
that's
gonna
lead
to
things
like
reproducibility
and
maybe,
as
as
a
as
a
topic,
it's
gonna
yeah.
A
No,
no,
no,
I
agree
yeah
and
then
I
I
the
the
other
ones
that
sort
of
resonated
with
sort
of
this.
This
openness
versus
privacy
versus
intellectual
property.
Nobody
said
that,
but
you
know
ip
is
important
and
that's
exactly
what
what
carolyn's
talking
about
with
those
machine
learning
models.
Nobody
wants
to
give
up
the
ip
of
what
they
are,
but
you
know
I
think
those
are
important,
the
the
tug
and
pull
of
that
and
how
you
can
do
that
you
know
have
those
work
together
in
an
environment.
A
I'm
you
know
sure
there's
I
I
agree
completely
that
you
know
openness
can
can
lead
to
potential
security
issues,
but
it's
also
hard
to
hide
from
math.
So
you
know:
does
this
give
us
an
advantage
on
some
of
these
things?
A
I
don't
know
that
that
might
be
an
interesting
one
to
do
on
its
own,
but
then
getting
into
you
know
if
we
go
through
workflows
and
all
these
things
and
then
finally
get
into
what
we're
really
talking
maybe
talking
about,
is
you
know
how
do
we
get
this
whole
system
working
in
a
distributed
fashion?
How
do
you
do
sort
of
this
big
data
sharing
and
have
then
things
like
ai
and
machine
learning
come
into
play?
A
B
B
But
we
have,
I
think,
we've
established
that
we
have
a
target
rich
environment
which
is
great
yeah.
I
think
we've
named
more
than
five
great
panels
and
I
don't
know
if
we
have
time,
but
if
we
were
able
to
do
some
winnowing,
which
I
think
was
where
you
were
going
now,
also
yeah
to
topics
that
we
could
then
go
recruit
speakers
on.
A
So
we've
got
the
notes:
we've
got
other
things,
we're
also
gonna.
You
know
I
don't
have
the
output
of
the
google
forms.
So
I
don't
know
what
florence
is
gonna
have
from
that.
So
what
I
will
take
an
action
to
make
sure
that
I
get
in
touch
with
her
and
get
that
we
can
sit
that
together.
C
B
But-
and
we
can
be
basing
about
it
right,
we
we
have
a
first
cut
at
what
the
landscape
looks
like
we
have
a
private
forum
to
modify
it.
A
If
you
could
just
jeremy,
can
you
link
in
here
in
a
you
know,
just
let's
get
a
google
doc
so
that
we
can
all
throw
in
pieces
of
our
notes
and
maybe
start
to
asynchronously
come
up
with
some
sort
of
I'll
come
up
with
some
sort
of
schedule.
I
will
I
will
go
ahead
and-
and
you
know
put
one
in
as
being
I
guess
I
could
do
it.
A
Yeah,
exactly
and
so
so
I'll
try
and
take
my
bad
notes
and
turn
them
into.
You
know
lists
of
all
of
the
different
things
here
and
then
take
a
shot.
You
know
just
as
as
our
prior
of
what
we
might
think
about
for
the
for
the
spring.
A
I
will
also
say
we
are
still
looking
for
panelists
for
the
november
and
december
meetings.
So
if
people
I
I
will
yeah
we,
I
can
put
those
in
there
as
well.
If
people
have
suggestions
for
other
folks
on
that,
but
we
can
we
can
continue
to
work
on
those
a
little
bit
as
we
go
forward,
but
you
know
go
see
if
we
can
get
a
spring
of
of
meetings
going
through,
and
I
would
propose
that.
A
Actually,
I'm
going
to
say
right
now,
the
first
friday
in
january
is
going
to
be
close
enough
to
new
year's
and
everything
else
that
we
skip
to
the
january
meeting,
and
just
maybe
do
do
you
know
february
march
april
may
and
then
june.
F
A
F
There
is
a
timing
thing
here
where,
for
example,
cast
will
be,
you
know,
cask
is
developing
this
position
paper
or,
and
one
of
the
items
is
specifically
around
what
we're
talking
about.
We
we
could
put
together.
I
could
commit
to
one
of
those
two
meetings,
probably
the
december
meeting,
to
sort
of
bring
the
bring
a
panel
together
to
at
least
frame
the
data
challenges
related
to
end-to-end,
workflows,
team
science
and
then,
and
then
that
way
we
would.
You
know
that
would
be
a
larger
framing
issue.
F
I
mean
we
could
do
it
that
way
as
well.
I'm
just
I'm
worried
that
if
we
don't
have
november
and
december,
we
might
be
looking
for
topics
but
yeah.
Oh
we've
got.
A
Yeah-
and
so
you
know
yeah,
I
think
von's
already
on
the
hook
for
the
november
cyber
security,
I'm
sure
we'll
hear
more
on
things
from
that
and
then
the
december
one
I
mean
come
on,
that's
a
semester
away
in
no
time,
but
you
know
I
I
I
have
at
least
a
few
folks
that
would
have
interest
in
presenting
to
that
and
I'm
sure
there
are
others
that
we
all
you
can
we.
A
Yeah,
no,
so
what
I
was
going
to
do
is
I
was
going
to
go
ahead.
Let's
just
make
the
dock
be
the
schedule-
and
here
you
know,
we've
got
november
december
october
november
december,
already
sort
of
burned
out
and
I'll
take
a
crack
at
flushing
out
the
spring,
and
then
you
all
can
throw
darts
at
it
and
tell
me
where
I'm
wrong
and
start
suggesting
folks.
That
would
be
good
to
talk
for
talk
to
these
panels
and
go
from
there.
A
Sure,
no,
I
was
going
to
do
that.
The
top
of
it
will
be
sort
of
the
bullpen.
You
know
we'll
do
the
scrum
style,
perfect,
wonderful,
so
yeah
we'll
have
we'll
have
the
bolt
band
ready
to
go
and
then
we
can.
You
know
if
other
people
think
that
I've
gone
wrong
or
that
there
are
other
topics
to
add.
We
can
just
do
it
that
way:
yeah
and
yeah.
In
fact.
Maybe
this
is
something
that
we
should
have
as
just.
A
C
I
would
just
have
a
meta
recommendation
that
we
make
sure
that
our
different
ideas
are
communicated
in
a
way
that
can
bring
in
people
who
aren't
aware
they
might
be
interested
like
the
microscopy
is
a
good
example
that
would
be
really
easy
to
advertise.
So
it's
just
the
same
old
people
who
already
know
about
it
versus
saying
you
know
this
exposes
issues
of
large
scale
data
and
you
know
high
data
volumes
or
for
whatever
we
want
to
say.
A
No,
I
completely
agree,
and
so
maybe
that's
something
that
we
ought
to
capture
in
here
as
well
is
is
marketing.
You
know,
where
else
should
we?
You
know
here's
a
topic
for
february.
You
know
what
other
places
should
we
just
send
out
feelers
so
that
people
might
be
interested
in
coming
to
the
meeting
and
then.
B
C
The
second
thing,
though,
is,
I
think
we
also
need
to
hammer
home
the
the
times
when
we
represent
something
that's
not
being
discussed
and
widely
in
other
venues
right.
The
classic
example
is
cyber
security
as
a
big
data
challenge
right
and
so
maybe
also
draw
that
out
as
well
to
help
people
look
more
towards
this
series.
A
C
C
Although
having
the
support
of
of
jeremy
and
ren.