►
From YouTube: 2020-03-24 Apache Cassandra Contributor Meeting
Description
Agenda for this week:
4.0 Status
Drivers: Addressing support for 4.0
Thrift issues in 3.x
A
Now
we
were
recording,
alright
I
was
good.
I
I
was
gonna,
say
I,
really
appreciate
the
weekly
update
emails.
Those
look
like
they
take
a
long
time
and
I
love
that
rotation.
That's
going
on.
That's
really
awesome,
but
I've
heard
a
few
people
out
in
the
wild
say
wow.
These
are
great.
They
really
give
me
a
great
snapshot
of
what's
happening
in
the
project.
Right
now
and
oh
is
a
weekly
reminder.
Oh
there's
stuff
that
I
could
maybe
contribute
to
so
just
a
huge
kudos
to
the
I.
A
C
Mean
it's
also
like
so
when
I
did
the
status
update.
Last
week
we
had
a
pretty
big
milestone.
I
might
be
buried
in
there
a
little
bit
that
we
had
all
of
the
Alpha
tickets
assigned
for
the
first
time
ever
in
the
project.
So
that's
really
rad
and
then
also
great
to
see
the
folks
who
have
worked.
The
testing
fixes.
C
But
the
larger
testing
plan,
I
kind
of
started,
encouraging
people
that
were
either
assigned
or
somehow
indicated,
involvement
and
the
tickets
to
start
thinking
about,
like
a
prioritized
list
of
what
they'd
want
to
test
in
those
areas,
saw
some
movement
on
that
in
the
I
think
like
packaging
area.
So
thanks
for
the
folks
who
picked
that
up
and
are
working
on
that,
you.
D
Might
as
well
just
chime
in
here
with
a
couple
things
so
I'm,
not
sure
how
many
of
you
were
aware,
or
if
this
was
already
discussed
at
a
previous
meeting,
but
the
docs
have
reached
a
pretty
damn
good
state.
We
no
longer
have
a
single
to-do
item
in
there
like
just
missing
Doc's,
like
they
used
to
just
be
data
modeling
with
it
to
do
so,
all
those
all
those
years
were
merged
in
them.
D
So
when
we
release
it,
we
can
say
things
like
you
know:
compaction
is
25%
faster
or
whatever
some
additional
numbers
around
around
screaming
be
able
to
provide
additional
guidance
to
folks,
because
there's
still
a
lot
of
nonsense
that
people
talk
about
like
I,
hear
people
say
like
oh
no,
more
than
1
terabyte
per
node
I'm
like
it's,
not
2014
like
we
can
do
better
than
that.
So
we're
really
trying
to
help
people
see
the
advantage
of
moving
up
as
opposed
to
sticking
with
old
versions.
C
That's
just
recipes
it'd
be
cool
to
to
get
some
of
the
work
that
Netflix
presented
at
the
Bay
Area,
like
Cassandra
user
designer,
like
Cassandra
runners
meeting
thing.
That
became
this
thing,
but
they
had
presented
some
work
on
testing
the
new
messaging
service
and
performance
there.
That
could
be
interesting
to
share
as
well.
I,
unfortunately,
don't
see
any
of
them
on
it.
C
D
Would
love
to
see
the
blog
actually
be
active
if
we
can
get
more
contributors
on
it?
I
think
I'd
be
great
if
we
were
rolling
out
a
couple
posts
a
month,
because
right
now
it
to
the
outside
world,
it
looks
like
the
project's
is
basically
dead.
So
I
need
to
I
want
to
make
sure
that
we
kind
of
revive
it
and
convince
the
world
that
cassandra
is
not
a
dead
project,
and
maybe
you
know
we're
actually
doing
things.
A
A
What
what
I
think
has
been
helpful?
This
is
kind
of
my
world.
A
bit
is
just
like
a
common
spreadsheet,
something
like
that.
Where
we,
you
know,
you
can
light
out
some
subtopics
and
let
people
pick
them,
take
them
and
run
with
them,
and
then
we
can.
It
just
gives
us.
You
know
it
gives
everyone
an
idea.
Oh
here's.
Here's
a
great
bunch
of
stuff!
A
Well
I,
can
add
my
two
cents
to
this
one
or
something
like
that
so
John,
maybe
if
you
and
I,
can
spend
a
minute,
we
could
probably
come
up
with
at
least
a
few
months
where
it's
a
blog
post.
That
way.
A
D
A
F
A
A
D
G
Most
recently,
the
the
drivers
have
received
support
for
transient
replication
metadata
and
it's
the
multiple
posts
or
ports
per
host.
The
thing
we're
still
tracking
is
the
revised
protocol
level
coming.
I
think
Sam
is
working
on
that,
but
we
haven't
seen
a
prototype
yet
so
no
further
development
on
that
front
for
the
drivers
in.
D
The
dimension
version
here
so
on
the
Java
driver
is
the
plan
to
only
include
the
4.0
support
in
the
4.0
version
of
the
Java
driver,
because
that
was
such
a
big
change.
That
I
don't
expect
people
to
be
able
to
actually
make
application
level
changes
at
the
level
that
they're
required,
especially
if
they're
pulling
configuration
for
anything
other
than
the
new
configuration
file
thing.
That's
in
there.
D
G
Our
general
stance
has
been
or
we're
doing,
forward,
development
on
the
driver
for
o-line
and
we
would
generally
like
to
avoid
backporting
things,
but
there's
they're
still
open
source
Apache
project.
So
if
people
have
the
desire
and
the
bandwidth
to
back
port,
we
we
welcome
polar
requests
and
we
can
also
talk
about
doing
that
work.
But
we
haven't
started
it
yet.
D
Okay,
that's
it's
a
little
okay,
yeah
I
I
think
that's
gonna
present
a
a
major
major
problem
for
the
community.
D
If
we
roll
out
for
oh
and
like
you
have
to
upgrade
your
driver
in
the
do
a
driver
migration
to
for
oh
I,
didn't
see
a
single
person
use
the
four
overs
and
the
driver
at
LP
and
I've
actually
tried
to
port
applications
to
it,
and
I
was
I,
hit
significant
hurdles
to
the
point
where
it
was
a
complete
non-starter
for
me,
like
I
I,
just
like
I
did
a
bunch
of
client
work
and
like
I,
had
to
get
cut
off
because
I
was
taking
so
much
time
to
do
it.
It's
like.
G
D
B
Kind
of
hope
the
bear
open
Pandora's
box
like
who
should
do
that
implementation
in
the
driver,
because
their
open
source
project
and
it's
apache
license
so
I
mean
if
data.
D
Foundation
if
data
sex
wants
to
own
the
drivers,
then
that's
sort
of
it.
If,
if,
if
data
sex
is
gonna,
say
wheel
in
the
drivers,
then
you're
basically
claiming
responsibility
for
it.
B
D
If
the
folks
at
data
sacks
want
to
contribute
the
drivers
to
the
project,
I
would
be
a
firm
plus
one
on
that.
I
think
that
having
it
not
as
part
of
the
project
is
a
giant
mistake.
I
also
think
that
dataset
kind
of
holds
the
project
over
a
barrel,
because
they
got
you
guys
get
to
just
decide
what
to
do
with
drivers
and
whatnot
like
you,
could
theoretically
not
do
any
work
and
never
support
something
like
transient,
replication
or.
B
D
In
in
my
ideal
world,
we
would,
and-
and
you
know
in
my
perfect
world-
I'd
love
to
see
all
the
dry-
it
like
a
new
Cassander
drivers,
repo
all
of
them
being
list
all
of
them
just
being
in
there
and
everybody.
That's
currently
commits
to
those
things
we
just
fast
track
everybody
and
as
a
committer,
with
the
understanding
that
you're
already
committing
to
the
drivers.
We
trust
you
to
keep
doing
it.
D
D
Let's
just
assume
all
that
stuff
is
done
correctly,
I'd
love
to
see
that
in
there
and
then
at
least
puts
us
in
a
position
where
you
know
that
we
we
have
there's
some
incentive
for
people
to
work
on
the
drivers,
because
right
now,
like
let's
take
transient
replication,
you
know
area
like
there
was
a
bunch
of
work
done
on
it
and
then
it's
sort
of
like
okay.
Now
we
hope
that.
G
D
F
D
Had
the
most
friendly
community
standpoint,
so
let's
just
take
it
back
a
year
if
Dana
Stax
wanted
to
say
not
be
involved
at
all,
you
guys
could
just
say
it
we're
just
doing.
Dsc
drivers
I,
don't
I'm,
not
saying
that
you
are
doing
that.
I
want
to
be
really
really
clear,
I'm
saying
the
potential
for
a
problem
is
there
and
by
moving
things
to
the
neutral,
ASF
ground,
it
makes
it
so
that
we're
we're.
Not
it's
not
a
risk
anymore
right.
D
B
D
D
Gonna
sue
you,
so
we
we
actually
can't
yes
you're
right,
we
they
could
be
for
it,
but
then
it
has
to
be
again
a
benevolent
third
party
willing
to
maintain
it.
We
can't
pull
it
into
it
about
you,
so
I
would
love
to
just
say:
hey,
let's
fork
it
right
now
who
bring
it
in.
It
just
doesn't
work
that
way.
How.
B
There's
a
risk
of
fragmentation
like
I
know
back
when
I
was
working
on
features
in
the
code
base
like
haka
would
barely
do
with
ghetto
implementation
in
Python,
so
I
could
actually
test
it
inside
the
codebase,
but
it
was
not
pretty
and
it
required
somebody
else
to
clean
it
up,
because
I'm,
not
a
Python
driver
dev,
certainly
not
go
sequel,
c-sharp,
c+
whatnot,
you
know
I'm,
not
gonna,
go
in
there
and
do
a
good
job
of
all
that.
So
how
do
we
as
a
project?
If
we've
got
five
or
six
different
drivers
inside
a
foundation?
D
I
think
that
I
think
that,
folks,
that
work
on
drivers
are
gonna
keep
working
on
drivers,
I,
don't
think
you're
gonna
see
like
you
know,
Alexi
Benedict
Blake
decide
that
they're
gonna
spend
most
of
their
time
on
the
driver
side
of
things
and
just
improve
one
driver
and
not
touch
the
others.
It
seems
very
like
counter
intuitive
it's
one
of
the
reasons
why
my
in
my
head,
it
makes
sense
to
just
say:
if
you're
already
committing
to
the
drivers,
then
you
should
continue
to
be
a
committer
like
it's
sort
of
to
me.
D
It's
kind
of
like
a
package
deal
I,
don't
know
if
it
is
like
that's
how
it
really
works,
but
again
in
my
fantasy
it's
just.
It
should
not
change
how
the
drivers
are
developed
right
now.
It
really
just
changes
the
like
the
entity
that
owns
it
in
a
way
that
prevents
you
know
hit
by
a
bus
problems
from
happening.
B
A
G
C
From
the
whole,
like
entry
out
of
tree
conversation,
not
that
we
can't
keep
talking
about
that,
but
I
think
the
the
other
conversation
that
we
almost
started
to
have
before
that.
That
was
kind
of
interesting
is
like
setting
some
sort
of
standard
like
we
did
for
alpha
beta
and
release
candidate
about
what
is
expected
from
the
drivers,
because
it
sounds
like
there's
pretty
big
disconnect
on
like.
C
If
we
want
the
current
major
version
to
support
everything
or
the
new
major
version,
or
do
people
need
to
change
their
driver
to
upgrade
to
Flora
I
mean
those
seem
like
important
things
to
figure
out
in
the
whole
like
Cassandra,
4'o
is
ready
when
fluoro
is
released
thing,
because
you
need
a
client
to
like
use
the
database
and
for
the
record
I'm,
actually
not
in
the
camp,
that
the
driver
needs
to
support
all
the
features.
But
it
does
seem
like
an
important
conversation
to
have.
C
D
Think
if
we
don't
have
a
fully
I,
don't
think
if
we
don't
have
a
fully
functional
Java
driver
when
it's
released,
then
there's
no
point
even
doing
it.
It's
just
like
hey
we're,
really
a
database,
and
maybe
one
of
the
drivers
would
work.
I'd
love
to
see
the
Python
one
as
well.
I
mean
that
to
me
you
definitely.
D
D
The
the
issue
is
is
that
the-
and
some
of
this
is
speculation
on
my
part,
but
the
I
noticed
that
the
changes
in
the
4.0
version
of
the
driver
basically
involve
a
instead
of
configuring,
the
driver
through
code,
it's
now
through
a
configuration
file,
and
so,
in
addition
to
a
lot
of
things
moving
around
within
the
driver,
most
of
which
make
perfect
sense.
It
changes
the
way
that
you
can
that
you
have
to
configure
the
driver.
So
it's
now
closer.
D
If
you
want
to
do
it
programmatically
it's
closer
to
like
a
Python
multi-dimensional
map
than
it
is
a
nice
fluent
API
and
I
noticed
that
when
there
was
the
last
data
stacks
conference,
there
was
the
you
download
this
file
and
everything
just
magically
works.
Unfortunately,
the
effort
to
port
from
an
existing
application
to
the
new
one.
D
C
D
Of
the
client
right
now
is
a
lot
of
work,
and
it's
not
something
that
can
be
done
incrementally.
It
has
to
be
all
done
all
at
once,
so
any
code
that
any
code
base
he's
going
to
have
a
commit
that
touches
every
file
that
uses
Cassandra
in
some
way,
and
my
experience
like
I
said
with
it
was
it
was
such
a
disruptive
change
that
it
was
like,
don't
want
to
do
it,
so
we
essentially
have
the
driver
now
sandbagging
Cassandra,
because
it
it
just
so
if
you
can't
use
the
new
features,
then
from.
B
B
Is
like
the
age-old
question
of
like
when
do
you
break
api's
or
not
like?
When
do
you
remain
tied
to
the
pound
versus
moving
to
the
future,
and
it
probably
helps
the
Heike
do
a
project
to
have
an
answer
for
that
before
changes
go
off
and
be
made
like
it's
kind
of
gray
space
on
when
we
make
changes
that
could
be
API
breaking
with
the
drivers
or
not
I,
think
I,
agree
to
say
John.
C
As
you
know,
say
putting
that
part
aside
for
a
sec,
not
that
I,
don't
think
it's
an
interesting
discussion,
but
just
like
on
the
technical
side
for
one
more
moment
like
I,
guess
that
you
know
you
had
said
that
the
most
embarrassing
thing
is
that
someone
goes
and
turns
on
transient
replication
and
the
client
breaks
I.
Think
to
me
the
more
base
and
there's
something
as
someone
upgrades
to
forró
and
their
existing
client
breaks.
C
Well,
yes,
I
agree
with
you
completely
that
I-I've
watched
having
to
upgrade
lots
and
lots
and
lots
and
lots
of
clusters
to
three
oh
and
I've
watched
how
hard
and
how
long
it
was
to
get
people
to
change
clients,
but
that
needs
to
definitely
be
like
a
minimal
friction,
yeah
and
then
now
minimal.
Friction
client
needs
to
be
able
to
talk
to
floor
Owen,
so
at
least
I
mean,
like
my
opinion
aside,
it
seems
I
think
Josh
start
to
get
this
like.
C
It
seems
useful
for
us
to
spell
out,
like
we
did,
for
alpha
beta
and
RC
what
is
expected
of
whatever
it
is:
the
3.0
and
4.4
not
x-series
clients
moving
forward
and
maybe
to
help
solve
some
of
what
you
spoke
about
like
we
can
discuss
that
as
a
community
caveat.
Obviously,
like
data
snacks
is
providing
the
majority
of
the
resources
for
the
work
and
is
going
to
have
a
lot
of
say
because
of
that
mm-hmm.
F
So
can
we
just
summarize
what
John
saying
is
that
we
don't
want
that
you
to
see
these
of
the
tribal
to
like
explore
when
people
obtain
to
design
and
ground
for
that?
Oh
so
that
so
that
we
can
just
summarize
as
a
let's
just
have
a
graceful
failure
mode
when
we
update
to
for
auto
on
the
under
silver
solid,
so
but
more
or
Jon.
F
You're
trying
to
say
is
that
we
also
need
to
support,
or
at
least
have
support
for
new
features
in
the
three
ceilings
of
the
driver,
which
is
which
is
a
bigger
discussion
exceedance.
So
if
people
upgrade
to
the
four
dot
or
driver
leads
to
a
lot
of
application,
refactoring
and
work,
but
but
the
baseline
that
you
you
have
is
everybody
needs
to
you
know
you
don't
want
people
to
update
to
for
auto
in
order
to
get
in
the
Ford,
auto
capabilities
on
the
server
side,
because
it
will
lead
to
too
much
work
on
the
client-side.
F
But
as
far
as
the
button
by
waistline
is
concerned,
is
that
if
people
are
played
on
the
other
side
too,
for
not
oh
they
a
I
mean
they
will
be
in
much
more
at
for
a
while.
While
we
are
paving
the
nodes
individually
and
at
that
point
they
need
to
use
the
three
Donald
Driver
they
can
potentially
preemptively
update
the
for
auto
driver,
but
I.
Don't
think
that's
a
that's
a
bigger
heavier
discussion.
So
so
we
just
don't
want
the
driver
to
fail
in
case
of
people
upgrading
the
Ford
auto
on
the
server
side.
D
Yeah,
that's
that's
what
I'm!
That's
what
I'm
getting
I
I
just
want
to
reduce
friction
to
four.
Oh,
because
there's
been
like
I
think
we've
done
a
in
the
history
of
the
project,
there's
a
bit
of
bunch
of
things
that
have
been
friction,
one
of
which
is
you
know,
data
loss
and
horrific
bugs
and
new
features
being
broken.
I
think
we've
we've
addressed
a
lot
of
this
stuff
on
the
server
side.
I
think
this
is.
D
This
is
going
to
be
the
best
release
that
we've
ever
had
as
a
project
like
there's
so
much
good
work
going
into
this.
It's
amazing
and
the
thing
that
I
don't
want
is
for
folks
to
go.
I
want
to
upgrade
to
four.
Oh,
but
there's
this
uncertainty
around
the
driver
like
there's
this
mystery
do
I
have
to
change
my
whole
codebase
like
yeah,
I
I,
just
I
want
to
get
people
to
this
version.
I
think.
A
This
is
you're
driving
the
right
point,
which
is
anything
that
gets
in
the
way
of
people.
Adopting
the
right
thing
is
the
wrong
thing,
and
we
can
probably
address
this
many
ways
with
code
education
like
that
blog
post,
like
here's
the
word,
this
is
how
you
do
it.
This
is
what
needs
to
be
done.
This
is
what
you
can
rely
on,
and
but
it
shouldn't
be,
and
you
jump
through
fiery
hoops
and
good
luck,
bro
yeah,
so
I'm
with
you
on
that
yeah.
D
And
I'm
honestly
I'm,
not
against
the
idea
of,
especially
since
I
now
see
that
there's
per
there's
a
programmatic,
API
I'm,
not
against
evolution
in
the
driver
like
I,
want
to
be
really
clear.
Like
I'm,
not
like
hey,
we
have
to
stick
with
the
same
API
I
made
the
point
earlier.
There
were
a
lot
of
really
good
changes
that
simplify
things.
D
I
believe
that
going
away
from
like
lavas
futures
to
built
in
Java
wands
to
fix,
like
the
the
chaining
of
API
calls,
was
a
nice
improvement.
The
organization
of
the
library
made
a
lot
more
sense.
There
were
so
many
really
good
improvements.
I.
Just
think
that
what
I'd
like
to
see
is
something
a
little
bit
more
gradual,
so
that
we
can
get
people
over
that
initial
hump.
And
then
we
say:
hey
we're
gonna
have
another
release
in
a
year:
okay
cool!
Then
you
can
move
on
to
the
next
version.
B
C
D
D
C
I
think
the
standard
that,
like
the
minimal
amount
of
client,
upgrade
to
get
to
four,
oh
and
I,
don't
know
what
that
means.
Does
honestly
I
should
be
a
little
bit
more
familiar
with
the
clients,
but
if
that
means
like
putting
a
few
things
in
the
300
series
of
the
client
and
then
before
datos
series,
having
like
all
the
features
or
something
but
just
like
again,
watching
the
pain
of
getting
a
bunch
of
people
to
upgrade
the
client
before
upgrading
their
database,
the
easier
we
can
make
that
the
better
I
think
and.
F
One
of
the
things
that
we
don't
very
sorry,
though
the
point
I
was
trying
to
make
earlier
and
even
I
mean
one
of
one
of
the
options
that
might
be
is
basically
keep
the
same.
Public
API,
but
internally
use
the
for
our
co-drivers
internals
right
like
people
care
about
the
public
API,
they
usually
don't
care
about
the
internals.
The
only
other
thing
that
they
care
about
is
that
the
new
version
of
the
driver
doesn't
pull
in
dependencies
that
break
their
existing
applications
either.
Two
main
things
that
I've
at
least
experienced
in
the
past.
F
As
long
as
you
keep
the
same
public
API
and
you
don't
pull
any
dependencies,
that's
gonna
cause
their
applications
to
bloat
or
you
know,
explored
in
production.
That's
as
long
as
you
keep
those
two
things.
You
know
people
don't
care
what
you
do
internally
right
so
again,
is
there
a
ready
town
where
we
can?
We
can
potentially
add
a
shim
I.
C
G
D
B
No
idea
what
that
would
look
like,
but
so
Jordan
you've
mentioned
this
whole,
like
we
probably
should
have
a
position
as
a
project
on
what
it
means
to
be
ready
for
a
release
with
drivers
able
to
connect
to
said
release.
So
that's
something
I,
definitely
think
we
don't
want
to
drop
on
the
ground
because,
like
you
mentioned
John
traffic
replication
a
couple
times,
it's
still
March
as
experimental.
So
it's
like
do
we
want
experimental
feature,
support
to
be
part
of
a
blocker
for
a
release
or
not
like.
B
C
100%
I
don't
know
if
that
means
a
separate
deck
or
just
taking
the
doc
zoom
on
started,
and
we
all
collaborated
on
and
extending
it
either
way.
I
mean
it
might
be.
You
know
whatever
it
is
I
think
we
should
try
to
state
our
values
there
as
a
project
and
I.
Think
Josh
brings
up
good
points
around
the
nuances
and
those
things
I
think
that's
a
definite
action
item
from
this
meeting
is
is
moving
that
discussion
to
the
mailing
list.
E
D
C
More,
we
made
some
statements
about
that
actually
in
the
original
doc.
I
think
we
just
need
to
refresh
our
memory,
but
I.
Remember
us
having
a
debate
I
think
in
like
it
can
change
through
alpha
and
can't
change
in
beta,
but
it's
in
the
doc
that
we
let
some
I'll
start
it.
We
agreed
on
Netflix,
okay,
agreed
on
that
Netflix
agreed
on
online,
with
discussion
in
Vegas
was
where
I
was
going
so
multi
works.
That's
like
sorry,
yeah.
A
A
H
A
A
I
C
A
A
And
again,
I'm
trying
to
keep
track
of
the
notes
here,
but
in
this
recording
will
be
available
for
it.
If
we
want
to
get
others
involved
in
this
conversation
they
can
see
where
everyone
was
talking
so
I'll
have
that
up
today,
probably
and
then
the
notes
as
well.
So
if
you
want
to
invite
your
friends
and
family
in
to
join
in
the
conversation
about
this
I'm
sure
all
the
times,
I
will
have
a
compliment.
I'm
sure
of
it
as
well-
and
you
know
everyone's
sitting
at
home
now
is
probably
you
know.
A
E
Yeah,
go
ahead,
yeah
so
being
recently
working
on
two
one,
two,
three
up
three.
So
there
are
a
lot
of
clusters
and
we're
trying
to
upgrade
them
to
three.
Oh
and
a
couple
of
things
we
have
noticed
so
far.
One
is
so
there
is
no
in.
There
is
no
cross-training,
obviously
between
two
one
and
three
yo,
so
I'm
trying
to
take
a
look
at
if
that's
a
solvable
problem
to
have
at
least
three
or
stream
from
two
one
and
second,
is
we
did
up?
Do
an
assistive
will
upgrade
on
one
of
the
two
one-fifth
clusters?
E
A
B
A
One
thing
I
was
gonna
I,
just
want
to
add,
as
kind
of
a
PSA
for
a
community
which
is
I,
am
currently
working
with
a
lot
of
groups
right
now,
like
community
organization
groups
around,
what's
happening
right
now,
with
our
pandemic
that
we're
all
enjoying,
but
there's
just
a
lot
of
activity
going
around
like
hey,
we
want.
We
have
a
lot
of
nerds.
A
lot
of
people
need
help.
A
How
do
we
connect
those
I
just
wanted
to
bring
this
up
as
a
something
that
potentially
you
know,
I'm
gonna
try
to
get
some
organization
around
it,
but
I
think
there's
a
lot
of
people
in
the
Cassander
community,
especially
in
the
deaf
community
that
have
some
amazing
skills
that
could
be
helpful
for
some
organizations.
So
just
a
heads
up,
I'm
hoping
to
get
all
this
kind
of
narrowed
down.