►
From YouTube: SimPEG meeting May 28th
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
C
There's
three
things:
there's
first
of
all,
I'm
discussing
with
duck
for
his
view
of
five.
What
blogging
is
that?
And
so
it's
willing
to
invest
a
little
bit
like
in
producing
one
or
a
few
pending
on
the
priced
videos
that
are
more
professional,
so
that
we
would
have
something
to
find
real
funding
afterwards,
though,
you
both
getting
some
kind
of
a
prototype
that
is
up
to
speed
and
then
doing
further
witness.
So
that's
nice,
so
we're
talking
with
the
studios
now,
because
they're
they're
making
estimations
because
which
are
quite
high.
C
So
if
this
stays
like
that,
it
will
only
add
money
for
one
video
or
delay.
They
are
ready
to
make
some
economy
of
scale
to
develop
several
that
are
in
the
same
line.
Then
there
might
be
a
few
in
it.
So
that's
what
I'm
working
on
now
so
Lindsay
there's
new
stuff
and
you
find
them
to
take
a
look.
You
can
have
a
look,
but
don't
worry
if
you
can't
I
mean
yeah,
it's
just
always
good
to
have
different
views,
but
yeah.
It's
not
a
problem.
If
you
are
not
available,
I'll.
C
So
there's
that
then
there's
the
pitch
so
I
started
finally
treating
some
of
the
simple
work.
I
did
here
in
a
paper
form
that
was
actually
the
work
of
a
master
student
of
mine.
That
I
have
just
read
Amelie
in
simple
for
me
to
train
myself
and
also
to
be
able
to
get
the
work
out
so
I
wanted
to
ask
actually
does
it
have
helped
me
here?
Doesn't
it
I
looked
at
me
an
answer
today,
but
who
wants
to
be
co-author
and
yeah?
C
At
least
we
view
the
paper
and,
if
comments
so
I
guess,
there's
here
my
Tivo
dong
and
directed
I
wanted
to
ask
and
you
can
see
whatever
you
wanna
contribute
or
not.
You
decide
that's
a
thing,
so
I'll,
probably
post
one
once
I
have
the
structure
I'll
post
a
draft
somewhere
where
you
can
have
a
look
and
then
decide.
Okay,.
C
C
Of
the
questions
I
had
actually
I
would
like
to
have
at
least
one
member
of
the
community
to
help
me
structure
which
how
you
typically
describe
like
in
a
material
and
methods
the
algorithms,
but
also
like
what
kind
of
parameters
do
you
guys
put
in
the
text?
And
what
do
you
run
a
put
in
appendix
this
kind
of
stuff?
I
think
it
would
be
good
to
do
it
in
the
same
way
as
you
are
used
to
do
it
so
that
yeah,
it's
confirmed
the
other
stuff.
C
B
C
Expect
big
geophysical
advancement,
so
that's
really
enough
application
of
like
simple
for
agriculture
and
just
seeing
what
are
the
the
possible
problems
with
micro,
typography
uncertainty
and
things
like
that.
So
it's
vacation
than
the
development
of
geophysics
but
okay
and
then
the
third
thing.
What
did
I
write
to
underneath.
A
C
Then
there's
I
asked
evil.
Today,
I
will
be
here
until
the
end
of
July
I.
Think
before
I
leave,
I
will
probably
I
was
invested
too
much
time
with
this
video,
so
I
will
have
me
everything
I
wanted
to
do
in
terms
of
synthetic,
so
I
I
will
have
to
continue
at
home,
of
course,
but
I
think
a
good
thing
would
be
that
I
leave
with
knowing
how
to
contribute
so
simply
get
hoop.
C
Also
the
standards
that
you
yeah,
that
you
have
in
mind
these
things
so
that
at
least
I
know
how
to
do
that
at
home.
Once
I
get
in
a
stage
that
I'm
ready
to
do
that,
I
can
I,
don't
have
to
bother.
You
is
that
from
a
distance,
did
you
see
the
URL?
Thank
you,
sir
yeah
I've
seen
that,
and
so
maybe
that's
a
good
start,
but
I
run
aside
like
how
we
did
with
Doug
the
to
earlier
tutorial
that
we
just
do
some
stuff.
C
C
B
E
A
D
D
Now
you
have
mesh
generation
all
kinds,
and
then
we
have
differential
operators,
and
then
we
have
the
inner
products,
all
tutorials
of
that
are
done,
and
then
the
a
couple
examples
of
just
really
putting
it
all
together
and
solving
a
couple
PDE
using
finite
volume
and
I'm,
not
I,
think
I
sort
of
understand
how
you
put
the
package
together
a
bit
more
now
where,
if
you
had
some
kind
of
term
in
a
PDE,
you
need
to
know
how
to
approximate
the
integral
with
a
function.
And
if
you
can.
D
D
So
that's
sort
of
how
it's
how
the
tutorials
are
set
up
and
yeah.
You
get
time
to
get
a
review
on
it
and
then
I
really
got
spent
a
lot
of
time
on
if
tools
and
3:53,
though
that
might
be
the
last
thing,
I
do
for
a
little
while.
B
Sounds
good
I
can
try
and
take
a
look
this
week,
it's
a
crazy
week
for
me,
but
keep
poking,
because
if
it
doesn't
happen
this
week
next
week,
even
crazier
so
keep
looking
the
thing
that
I
was
thinking
when
I
just
had
a
defer
to
plan.
B
So
if
it
would
be
helpful
actually
go
through
because
this
to
try
new
documents
and
things
like
that,
wasn't
to
documentation
so
that
in
this
product
we
also
get
a
bit
of
a
style
guide
out,
and
so
we
can
put
that
we
actually
have
started
a
community
repo
on
the
suit,
a
gorg
and
so
I.
Think
sort
of
for
general
documents
like
that,
like
the
contributors
guide
that
that
saara,
the
cerumen
and
the
government,
something
like
that
who
lives
there
and
then
like
style,
would
be
a
good
as
well
as
well
as
a
maintainer
guys.
B
B
D
B
D
That
was
that
was
the
only
issue,
though,
because
when
I,
like
I'm
on
my
branch,
I
think
I
made
a
kind
of
it
I
sort
of
made
like
a
examples
like
holding
place
or
like
a
proposed
example.
So
somebody
brought
something
in
that
they
wanted
to
add,
there's
a
place
for
it
to
live
so
that
the
website
will
build,
and
then
we
can
try
and
add
it
and
when
I
put
it
into
that,
it
was
the
only
thing
that
was
causing
the
bill
to
fail.
D
B
So
so
don't
worry
about
that
because
we're
not
deploying
the
docks
anymore
through
read
the
docs
are
being
deployed
to
github
pages.
So
like
it's,
it's
not
worth
it's
not
worth
investing
effort
into
fixing
read
the
docs
because
we're
no
longer
we're
not
so
yeah
cool
thanks,
Devon
I'm
excited
about
excited
about
all
the
work
you've
done,
except.
B
I'd,
do
you
have
some
things
you
want
to
jump
in
on.
F
Not
much
mal,
no
I
would
start,
but
you
know
John
Anna,
everybody
there's
a
there's
more
people
than
you
know
that
excited
about
my
grade
geometry.
My
short
trip
to
South
Africa
I
had
like
a
bunch
of
people
that
want
to
work.
I
want
to
work
on
it,
so
nice
I
think
Jude's
gonna,
be
it's
gonna,
be
a
big,
a
big
kick-off
for
it.
Yeah.
G
G
G
G
So
far,
well,
it's
been
matching
what
I
guess
some
other
surveys
have
been
flowing
around
the
area,
so
yeah
I
haven't
them
still
yet
to
include
the
TMI,
but
the
one
guy,
the
one
guy
I'm
working
with
said
that
it
was
just
a
flux,
gate
and
he's
not
very
confident
in
the
data.
So
he's
not
sure
if
he
would
like
to
use
it
or
not.
Okay,.
G
I
haven't
got
to
that
yet,
because
I've
been
tied
down
with
trying
to
get
the
octree
meshes
to
look
a
lot
like
archon
to
resemble
a
lot
of
our
cancer
mesh
results.
Okay,
our
tensor
mesh
results
are
kind
of
are
actually
a
bit
better
and
until
I
can
start
doing
that,
no
one
really
wants
to
push
forward
on
the
octree,
yet
so
I'm
trying
to
find
out
why
it's
not
what's
going
on.
G
B
A
tiling
front,
Donna
I,
think
there's
an
app
Laura
quest.
That's
been
open
in
simpang
that
I
want
to
ask
you,
but
still
relevant
work.
It's
been
surpassed,
I'm
just
pulling
it
up
here.
Oh,
though
the
local
problem
one
is
that
still
is
that
still
one
we
need
to
clean
up
and
get
in
or
I
will
send
you
the
link
yeah.
F
B
B
F
C
G
No
I've
just
been
working
on
the
Grady
Amma
tree
stuff
and
yeah,
trying
to
get
that
octree
to
match
up
with
a
tensor
stuff
or
with
the
results
so
mother,
dad
I'm
kind
of
at
a
standstill
until
I
started
getting
out
a
little
bit
pushed
forward.
A
bit.
I
did
make
some
surface
way.
Tools
for
octree
as
well.
I
can
add
that
in
once,
I've
got
some
satisfactory
results.
Excellent.
H
Yeah
I
guess
I
don't
have
that
much
to
share
right
now.
I've
just
been
looking
more
into
the
documentation,
but
some
pegging,
seeing
what's
really
necessary
to
include
in
the
training
that
proclaimed
in
my
amore
and
I.
Don't
know
if
Ashley
said,
but
we
to
this
group
before,
but
we
did
actually
get
funding
for
the
project,
so
it's
gonna
be
happening
and
thanks
to
people
like
Lindsay
and
soggy
and
any
others.
H
B
Let's,
let's
find
a
time
even
just
for
a
quick
chat
and
for
sure
we
can.
We
can
prioritize
a
few
things.
I
think
one
thing,
though
it
would
be
good
to
chat
about
actually
like
how
to
distribute
simpang
when
internet
connections
are
spotty
and
because
I
know
domme
this
ya.
B
B
Because
it
would
be
nice
if
we
can
sort
of
try,
try
out
an
installation
on
a
comp,
welcome
pewter
or
get
somebody
to
test
out
to
test
it
out
there,
but
one
thing
that
so
what
I
was
thinking
and
I've
seen
this
done
before
and
we
would
just
need
to
try
it
out.
But
basically
the
thing
to
do
would
be
to
download
anaconda
multiple
versions
of
anaconda,
for
Mac
Linux
and
for
Windows
you
and
USB
stick
could
anaconda
is
the
heaviest?
B
Is
the
heaviest
piece
right
like
that's
the
most
expensive
installation,
yeah
right,
you
can
download
all
versions
of
that,
and
then
we
can
all
have
unloaded
source
code
for
for
simpe
and
all
dependencies.
Basically,
so
basically
they
pick
install
everything
onto
a
USB
stick
and
then
you
pass
the
USB
stick
around.
So
we'll
need
to
try
this
out.
I
haven't
done
that
for
Python
code
before
but
in
theory,
it
should
work.
Yeah.
F
B
I
know
so,
like
the
pip
install,
wouldn't
work
in
that
case,
but
the
if
you
actually
install
from
source
we're
fine.
It
should
not
error
on
you.
It
should
just
say
the
tree
mesh
is
not
available
and,
like
that's
fine
yeah.
So
what
we're
gonna
have
to
test
this
out
and
so
it,
but
it's
just
something
to
sort
of
put
on
the
radar
and
and
we
can
chat
further
yeah
yeah.
C
H
I
H
F
B
C
B
B
So
if
you
haven't
seen
the
loop
project,
they're,
basically
like
working
on
trying
to
bring
together
Etra
physical
geologic
and
geophysical
data
in
the
interpretation
process,
so
all
the
things
that,
like
we've,
been
thinking
about
and
talking
about,
they
want
to
make
it
open
source
and
so
they're
talking
about
developing
open
source
software
already
in
touch
with
Miguel,
who
does
gem
pie,
which
is
like
the
3d
structural
geology,
modeling
in
Python
and
I
saw
that
they
were
like
looking
to
start
doing
geophysical
stuff
and
probably
planning
to
write
their
own
codes.
B
So
just
wanted
to
reach
out,
have
a
conversation
with
them
and
just
say
like
hey,
he
exists,
don't
write.
Your
own
inversion
framework
like
come
help
improve
this
one.
So
got
connected
with
two
marks
who
were
at
UWA?
Not
the
mark
that
we
know
from
he
was
a
trio,
but
now
he's
at
a
start-up
and
Laurent
who
is
at
Monash,
University
and
so
yeah
so
connected
with
all
of
them.
B
Had
a
great
conversation,
they're
interested
in
using
synthetic
they've
passed
along
a
couple
papers
of
things
that
they're
looking
into
and
Tebow
I,
think
you've
already
seen.
You've
probably
already
seen
those
papers
they're
relevant
to
your
work
yeah,
and
so
they
are
still
a
few
months
out
from
basically
picking
anything
up
there
right
now
in
the
hiring
stages,
so
they're
working
on
hiring
three
developers
and
three
like
scientific
postdocs
and
so
yeah.
This
was
perfect
timing.
B
To
have
this
conversation
is
because,
hopefully
we
can
actually
I
mean
it
would
be
amazing
if
one
of
their
developers
is
like
part-time
working
on
something
that
would
be
super
cool,
so
they're,
probably
gonna,
be
in
touch
I.
Think
mark
already
joined
the
slack
group.
One
of
the
Marcus
did
yeah.
So
just
like
have
a
look
at
the
the
loop
stuff
and
we'll
see
where
all
that
goes,
but
it'd
be
nice
to
get
like
a
contingent
in
Australia
using
synthetic.
That's.
B
F
B
I
they're
pushing
on
it
I,
don't
know
we
didn't
try
too
much
in
detail
about
where
it
is,
but
it
is
like
a
listed
partner
of
the
project.
So
if
you
see
on
the
loop
3d
sites,
it
is
let's
see
in
Aachen
in
Germany,
the
are
wth.
Aachen
is
that's
where
Miguel
is
based,
and
so
that's
that
connection
there.
F
B
B
With
that
one
of
the
things
and
this
one
is
brought
up
with
the
community
stuff
I
think
we
should
also
give
some
thought
to
like
just
trying
to
document
a
bit
more
sort
of
decision-making
processes
and
all
of
that
sort
of
stuff,
as
we
have
people
who,
like
are
just
not
able
to
join
the
meetings
and
like
aren't
in
daily
contact
with
us,
it'll,
be
important
to
like
communicate.
What
we're
thinking,
how
decisions
are
made
and
and
all
of
that
outwardly.
B
B
Let
me
see
if
I
can
find
this
real,
quick,
but
I
don't
know
if
you
guys
all
remember
that,
like
right
now,
when
we're
in
the
data
class,
you
can
basically
index
into
the
data
class
using
a
source
and
receiver,
and
then
you
get
back
a
datum
which
is
kind
of
cool,
but
at
the
same
time,
it's
now
a
bit
challenging
if
we
want
to
host
and
like
hold
both
the
data
and
the
uncertainties
on
the
same
object.
B
So
I
put
together,
I
had
there's
a
conversation
with
Rowan
and
Franklin
on
slack
a
little
while
back
and
from
that
I
put
together
just
like
a
few
notes
of
what
I
was
thinking
just
as
to
how
to
simplify
that
I,
just
posted
it
in
the
Google
hangout
and
I'll
post
it
in
the
notes
here,
but
the
idea,
basically
being
instead
of
indexing,
the
data
class
to
get
a
single
datum
back.
We
would
just
have
a
method
on
that
class.
B
So
there's
like
a
bit
of
a
sketch
there's
a
bit
of
a
sketch
in
the
the
link
that
I
just
posted.
If
you
have
questions
or
thoughts
right
now,
we
can
certainly
chat
about
it.
But
if
not
just
like,
take
a
look
and
and
add
thoughts
along
the
way
and
we
can
get
that
included
with
like
the
simulation
upgrades.
B
No
like
transparent
way
to
do
it,
it
ends
up
like
I,
was
trying
to
have
that
work
at
first,
but
it's
you
end
up
having
to
like
nest
too
many
layers.
It's
like
custom,
Python
objects
that
are
a
little
bit
magical
and
so,
like
I,
think
the
maintenance
load
and
just
like
the
transparency
of
what's
going
on,
gets
too
complicated
to
you
quickly.
F
B
We're
not
really
attaching
uncertainties
to
it,
and
so,
if
you
want
to
then
attach
uncertainties
to
it,
each
of
those
is
a
vector,
it's
sort
of
like
fundamentally
how
we
want
to
work
with
it
within
simpe,
but
then
to
populate
that
you
want
to
be
able
to
index
it
by
source
and
receiver,
and
so
then
you
have
to
start
thinking
about
okay.
So
is
this
like
a
fancy
numpy
array
that
we're
trying
to
attach
to
it?
And
so
that's
what
I
was
looking
at
trying
to
do?
B
But
by
building
like
a
fancy,
numpy
array,
then
you
sacrifice
basically
the
ease
of
just
treating
it
like
a
vector
because
it's
not
really
a
vector,
and
so
there's
this
ambiguous
question
that
happens
of
like
okay.
Is
this
thing
a
fancy
object,
or
is
it
just
a
vector?
And
so
like
you
sort
of
either
like
we
have
to
explicitly
choose
one
or
the
other,
or
you
have
to
expect
that
the
user
comes
with
disentangle
disentangle.
Am
I
using
this?
F
I
guess
I'll
have
to
see
it
in
action.
I,
don't
understand
why?
Because
it
like
data,
the
actual
measurement
and
the
uncertainties
cannot
just
be
two
properties
of
the
object
and
then
they
all
queried
to
the
same
location.
To
the
same,
like
you
know,
transmitter-receiver
pair,
that's
not
how
you
so.
B
B
B
F
B
F
B
F
B
I've
started
sort
of
the
base
classes
and
then
this
thing
sort
of
like
said,
be
done,
but
now
the
ejector
either
getting
it
plugged
into
him.
She
can
be
reasonably
straightforward.
B
Yes,
because
this
is
because
is
like
a
fairly
continued
fix,
simulate
you
class.
Unfortunately,
it's
gonna
end
up
being
a
big
PR,
because
there's
not
like
an
easy
way
to
make
them
smaller,
but
we
can
perhaps
bring
that
in,
like
on
a
dev
branch
and
like
have
a
little
bit
of
a
sandbox
ground
for
us
to
play
with
it
at
first
and
then
merge
it
into
master.
Once
we've
worked
out,
the
kinks.
F
B
B
B
D
Maybe
I
go
one
one
thing
about
the
discretize
package,
so
I
think
I
brought
it
up
and
made
a
couple
of
issues
on
on
github
about
having
the
full,
comprehensive
differential
operators
and
boundary
conditions.
So
we
we've
defined
our
divergence
is
being
faced
to
sell
centers
we've
got
our
curl
operator
going
edges
to
faces
and
we've
got.
We
have
the
ones
that
are
set
up
to
do
our
a.m.
D
problems,
but
we
don't
have
like
a
general
general
post
of
the
differential
operators
and
we're
not
able
to
impose
Dersch,
lay
annoy
minh
on
on
all
of
them
and
I
think
the
ability
to
put
in
boundary
conditions
for
the
tree
meshes.
We
don't
have
that
yet.
So
we
like
to
do
some
tricks
like
use
the
transpose
of
our
curl
operator,
to
go
from
faces
back
to
edges,
and
I
just
yeah.
B
Thanks
for
reading
that,
Devin
no
I
think
that
that's
that's
a
helpful
thing
to
have
outlined
and,
if
you're
willing
to
like
even
flesh
that
out
a
bit
Morris
like
what
you
see
is
missing.
I
know
you
mentioned
the
boundary
conditions
and
like
yeah.
If
you
want
to
add
some
pointers,
that's
great
from-
and
this
is
this
is
my
perspective
and
I
think
like
how
we've
done
a
lot
of
development.
B
So
far,
a
lot
of
it
is
on
and
as
need
a
sis
if
there's,
if
there's
a
problem
or
an
application
that
drives
the
need
of
that,
then
like
absolutely,
let's
get
cracking
on
it.
If
not,
then
it's
still
it's
extremely
useful
to
have
it
documented,
so
that
when
somebody
comes
along
and
does
actually
need
this,
we
can
point
them
to
the
issue
and
say
you
know:
we've
thought
about
this:
a
bit
there
hasn't
been
a
need
to
implement
it.
B
Yet,
but
here's
enough
to
get
you
going
and
like
we're
happy
to
help
you
implement
this
and
make
it
happen.
It
is
a
great
way
to
onboard
people,
so
that
would
be
sort
of
like
perhaps
the
way
to
think
about
like
really
filling
out
the
functionality
of
discretize
yeah
is
like
as
needed,
but
for
sure
keeping
track
of
like
what
are
the
things
that
we
don't
have
and
can
they
can
they,
like
onward
people,
to
bring
those
things
in
if
they
need
them?
Yeah.
D
And
these
pretends
are
mesh,
it
seemed
like
it's
pretty
low-hanging
fruit,
I,
don't
know
how
much
more
involved
it
is
for
the
tree
mesh
or
sill
mesh,
but
yeah.
It
seemed
like
for
for
tensor
mesh.
I
was
kind
of
surprised.
I
couldn't
find
some
of
those
things,
but
yeah
I
can
do
a
bit
more
better
documentation
and
then
yeah.
Whenever
it
becomes
important,
then
we
can
just
kind
of
knock
those
things
off
the
list
that.
B
Cool
okay:
are
there
any
other
items
for
people
to
bring
up?
B
No
so
next
week,
I'm
gonna
be
a
way,
so
I
won't
be
able
to
join
but
feel
free
to
to
still
have
a
meeting.
If
that's,
if
that's
useful,
if
not,
we
can
catch
up
the
following
week,
yeah
and
Kevin.
Actually,
if
you
have
time,
we
could
chat
for
another
twenty
minutes
right
after
this.