►
From YouTube: SimPEG Meeting April 21st
Description
Weekly SimPEG meeting from April 21st, 2021
A
So
again,
thank
you
guys
for
all
for
showing
up
today,
a
few
other
things
going
on,
so
I'm
not
quite
sure
who
else
is
going
to
be
stopping
by,
but
might
as
well
just
get
rolling
with
this
again,
please
add
yourself,
as
I
say,
as
always,
to
the
quick
reports
if
you've
got
something
to
chat
about
as
well
as
just
the
attending
list
and
things
like
that.
A
As
far
as
agenda
items,
there's
still
some
good
talks
happening
with
the
transform.
So,
if
you'd
like
to
get
involved
like
the
theater
was
telling
us
before,
they've
had
a
lot
of
views
so
some
good
stuff
there.
It's
really
good.
C
I
I
don't
know
honestly,
because
I
think
the
only
way
is
just
watching
the
youtube
views.
I
guess
it
was
about
100
100,
something
190.
I
don't
know
exactly
yeah
in
the
slack
room
there
was
about
40-ish
people,
I
guess
yeah
theater
like
do
you
get
any
analytics
about
that?
How
many
people
were
there
and
like?
Is
that
just
watching
number
of
views
in
the
youtube
is
that
sort
of
the
way.
D
C
A
A
Okay
with
that,
I
can
kick
us
off
with
some
quick
reports
items
here,
so
I
thought
I
made
a
lot
of
progress
on
the
boundary
conditions.
I've
as
far
as
writing.
I've
written
tests
that
tests
that
have
you
know
pde
solutions
actually
converging
properly,
with
with
boundary
conditions
and
applied
as
well.
A
So
just
for
all
the
truth,
the
different
types
of
boundary
conditions
we've
been
dealing
with
for
pers
the
simple
poisson
problem
right
now:
I'm
also
have
I'm
finishing
up
tests
on
like
em
light
problems,
so
curl
curl
equations
things
like
that
in
2d
in
3d,
so
this
finishing
up
and
that
should
be
shortly
it'll
I've.
I've
got
poisson
with
boundary
conditions
for
the
cell
centered
and
nodal
problems
yeah.
It
works
pretty.
Well,
it's
nice!
So
as
soon.
A
Yep
so
it's
tested
on
it's
tested
on
tensor,
archery
and
the
curvilinear
meshes,
that's
awesome
and
it
was.
I
don't
know
what
was
going
on.
I
was
getting
these
really
interesting.
Like
numerical
error
patterns.
A
A
And
it
seemed
like
very
interesting.
I
had
no
idea
why
it
was
doing
it
the
other
thing,
so
that
should
be
in
short
and
we'll
be
able
to
get
as
soon
as
that's
in
we'll
be
able
to
move
a
lot
of
the
2d
and
1vmt
work
that
lindsey
and
john
and
I
have
been
working
on
into
simpeg
and
the
ones
that
rely
on
the
boundary
conditions
to
drive
the
systems
so
it'll
be
nice
formulations
for
those.
A
The
other
thing
is
again
is
my
reminder.
Every
week,
please
test
out
the
d
properties
branch
of
discretize,
peter
I'm
assuming
you've
been
okay,
like
that
was
the
only
after
you've
had
that
one
issue
that
was
all.
A
E
F
Yeah
so,
and
if
all
of
the
simpeg
stuff
also
runs,
I
don't
know
if
anybody's
tested
that.
But
if
the
simple
documentation
builds
with
that.
F
A
Days,
okay,
yeah.
My
only
real
concerns
about
it
at
this
point
is
that
I'm
missing
some
edge
cases.
As
far
as
like
the
way
people
have
initialized
things
or
possibly
how
they
are,
people
are
bringing
like,
deserialization,
serialization
and
deserialization.
I'm
probably
missed
a
few
edge
cases
with
that.
Those
are
the
only
things
that
I
think
might
have.
Issues
with
everything
else
should
be
is
fine.
Everything
else
should
be
just
fine,
but
those
are
the
only
things
I'm
really
concerned
about.
D
C
Yeah
yeah,
I
think
I,
like
your
perfection
kind
of
but
the
as
long
as
the
like
a
test
is
passing.
Then
I
think
you've
done
your
job.
I
guess
so
because
that's
their
problem
like
if
they
really
wanted
to
keep
kind
of
be
updated.
They
should
have
put
the
test
about
that
in
this
impact
and
make
sure,
because
that
I
think
as
like,
as
the
fast
like
test
is
passing,
I
think.
That's
not.
I
guess
your
problem
too.
Maybe
there
are
problems.
A
Yeah,
let's
just
make
sure
that
all
the
the
tests
pass,
at
least
before
we
merge
in
the
discretized
stuff.
F
I
caught
well,
I
copied
in
a
couple
emails,
but
joe
did
you
want
to
go
over
the
poster.
A
Next
week,
at
on
a
session,
it's
kind
of
like
open
source,
open
source
building,
open
source
communication,
open
source
science
conversion
stuff,
like
that,
so
it's
we're
putting
together
a
poster
and
geophysical
inversions
with
something
like
kind
of
really
broad,
just
introducing
sent
back
to
this
group
of
people.
As
far
as
I
understand
it,
we're
wanting
to
have
examples
like
such
like
the
myanmar
example.
A
We
thought
would
be
a
good
one
to
put
in
there
mining
examples,
probably
nice,
because
one
of
the
things
that
they're
trying
to
do
is
also
help
connect
like
eu
and
russia.
Scientists,
that's
one
of
the
other
topics
of
this
session
like
promote
open
source
communication
or
promote
communication
between
those
two
entities.
So
that's
one
of
the
reasons
they
brought
a
mining
example
as
far
as
the
russian
contingency
and
then.
A
C
F
Well,
so
the
myanmar
work
includes
a
nice
example
of
humanitarian
geophysics
shows
the
dc
resistivity.
We
can
show
the
octree
code
a
bit,
and
then
I
I
mean
I'm.
I
think
we're
pretty
open
on
sort
of
ideas
for
a
second
example,
but
what
it
would
be
nice
to
show
is
something
that's
relevant
to
this
audience,
something
that
shows
off
a
bit
more
complex
physics.
F
So
if
we
show
the
dc
resistivity
problem,
it
would
be
nice
to
show
either
something
em
or
to
show
something
like
the
mvi
work.
That's
a
bit
more
of
an
involved
inverse
problem,
something
that,
like
he
said,
this
is
also
you
know,
a
compelling
research
tool.
F
Yeah,
so
don't
have
like
strong
opinions
on
what
that
second
example
would
be,
but
I
think,
having
two
that
kind
of
have
different
applications
and
also
use
different
physics,
would
be
nice
to
nice
to
show.
So
I
don't
know
if
others
have
thoughts
on
like
what
you
see
as
sort
of
a
compelling
use
of
of
simpeg.
That
you'd
want
to
show
off
to
this
community.
E
Well,
I
mean
right
now
just
with
getting
the
the
static
utils
stuff,
almost
ready,
that's
it's
kind
of
at
a
point
where
I
think
it's
more
more
versatile
to
use.
So
I'd
I'd
like
to
promote
that,
because
it's
it's
cleaned
up
and
you
can
do
a
lot
of
good
stuff
with
it.
The
3d
pseudo-section
plot
converting
3d
surveys
to
2d
slices.
E
D
C
F
Yeah,
so
it's
sarah
gary
who
invited
us-
and
so
I
put
in
the
notes
here-
the
session
description-
is
in
the
I've
linked
it.
Let
me
just
copy
that
for
you,
but
the
so
the
session
is
called
breaking
the
barriers
and
it's
kind
of
about
open
source
projects,
open
data
and,
like
the
the
role
of
these
and
sort
of
promoting
collaborations,
and
this
event
specifically
is
sort
of
targeted
at
helping
facilitate
collaborations
between
researchers,
both
in
the
eu
and
in
and
in
russia.
F
That's
sort
of
the
target
audiences,
but
I
mean
the
we're.
Also
a
you
know,
a
broad,
diverse
project
and
not
necessarily
everyone's
in
the
eu
or
russia,
and
that's
fine.
But
I
think
we've
got
some
perspective
to
offer
on
you
know
collaborative
projects,
so
I
think
really
the
role
of
the
poster.
F
We
want
to
show
off
some
of
the
capabilities,
but
also
like
really
it's
it's
something
to
to
talk
about
and
sort
of
prompt
discussion
with
folks
who
are
doing
sort
of
environmental,
near-surface,
geophysics
and
so
devon.
To
to
your
point,
what
I
would
potentially
see
is
like,
I
think
the
myanmar
work
is
a
great
sort
of
jumping-off
point
for
discussions
about
the
dc
capabilities,
and
things
like
that.
So
we
might
not
necessarily
go
into
like
a
you
know,
a
big
list
of
what
is
currently
possible.
F
But
if
somebody
starts
asking
about
that,
it's
great
that
we
can.
You
know,
use
that
as
a
bit
of
a
jumping
off
point
for
for
conversation
and
actually
anyone's
welcome
to
join.
This
is
a
free
event.
So,
if
anyone
wants
to
come,
you're
totally
welcome
to
be
great
to
have
great
to
have
a
number
of
people
there.
It's
on
tuesday
morning.
E
F
Yeah,
so
also
I
mean
soggy.
If
you
want
to
show
off
some
of
your
airborne
em
work,
that's
also
totally.
We
could
use
the
second
example
for
that.
I
think
it's
it's
pretty
flexible.
So
it's!
Basically,
you
know
if
there's
anyone
here
who
wants
to
highlight
some
recent
work,
we
should
definitely
use
this
as
a
chance
to
do
that.
C
Lindsay
are
you
going
to
do
a
demo?
Is
that,
like
a
or
just
a.
F
F
We
can
point
to
things
that
people
could
go
and
follow
up
on,
so
we
could
definitely
point-
and
it's
probably
a
good
thing
to
point
to
both
the
transform
talks
and
the
m
rs,
maybe
as
well
but
yeah.
It's
just
a
static
poster.
D
Maybe
to
your
question,
so
I
never
heard
of
that
event,
which
is
interesting
because
I'm
always
getting
news
from
the
local
chapter
eig
and
I've
not
seen
anything
so
far,
which
is
a
bit
weird,
but
so
it's
a
russian
european
workshop.
I
guess
that's
where
these
things
are
coming
from
the
conveners
search
for
topics
that
they
can
fit
into
this.
C
B
C
Yeah
yeah,
that's
that's
what
I
recognize.
B
B
F
F
Cool
okay,
a
couple
other
things,
so
I've
started
making
a
bit
of
progress
on
the
governance
document.
It's
still
a
work
in
progress.
I
would
say
it's
not
worth
spending
your
time
on
it
right
now,
but
just
I
put
it
there.
F
So
if
you
would
like
to
see
the
state
of
it,
you
certainly
can
I'll
ping
when
it's
actually
worth
taking
a
read
through
when
we've
got
sort
of
a
full
draft
and
we'll
do
that
as
a
pull
request
and
we
can
walk
through
if
folks
are
unfamiliar
with
how
to
leave
comments
and
things
like
that
in
a
pull
request.
If
you've
got
suggested,
changes
for
the
language
or
questions
or
anything
like
that,
so
that'll
be,
I
think,
the
way
we
can
do
a
bit
of
a
review
process.
F
There's
one
issue.
I
would
appreciate
input
on
it's
very
simple,
but
for
this
community
repository
we've
got
a
blend
of
markdown
documents
and
rst
documents,
and
I
think
we
should
pick
one
and
so
feel
free
to
add
your
thoughts
there.
I
would
vote
we
go
with
markdown,
but
take
a
look
and
if
you
have
other
thoughts,
that's
that's
fine,
but
I
think
for
our
to
keep
things
simple.
It
would
be
nice
to
just
just
work
with
one
or
the
other.
F
F
It
sounds
like
they're
going
to
do
another
one
of
these
mt
challenges
with
ford
modeling
and
inversion
to
be
held
with
an
event
on
november
8th
to
12th,
and
so
I
think,
it'd
be
awesome
if
we
got
in
on
that
and
sort
of
show
some
simpeg
examples-
and
I
think
you
know
the
fact
that
we've
pushed
so
hard
on
the
mt
code
recently,
it
would
be
really
nice
to
be
able
to
to
come
to
this
workshop
and
and
show
something
reasonably
compelling
and
kind
of
yeah
have
a
seat
at
the
table
there.
D
D
B
I
think
this
would
be
really
valuable
to
to
shoot
for,
and
the
mt
community
right
from
the
get-go
has
had
a
lot
of
these
challenges
to
mostly
driven
by
alan
jones.
Just
to
try
to
you
know
see
what
codes
are
working
and
having
test
examples.
B
I
I've
been
to
a
couple
of
these
are
really
illuminating,
because
people
are
kind
of
sitting
back
and
they're.
Saying
oh,
didn't
realize.
I
didn't
realize
that
or
you
know
how
do
we
actually
work
with
this?
So
it's
a
very,
very
good
learning
experience
as
well
as
the
fact
that
you
pick
up
a
lot
of
tips
from
what
other
people
have
done
and
if
you've
got
something
that's
being
developed,
you
know
in
the
in
the
open
source
that
is
going
to
be
extremely
extremely
appreciated.
B
A
A
F
Excellent
well
I'll
make
sure
to
keep
these
emails
on
the
radar,
and
I
don't
know
if
the
models
have
been
distributed
yet,
but
when
stuff
comes
out
we
can
we
can
get
working
on
that
and
maybe
even
start
a
dedicated
channel
and
slack
for
anyone
who
wants
to
participate
and
because
it
would
be
great
if
multiple
people
run
simulations
and
inversions
in
a
sense
kind
of
deter
like
what
we
did
with
the
csm
case.
F
Even
if
we're
using
the
same
code,
you
know
people
will
choose
different
mesh
designs
and
stuff
like
this,
so
I
think
it
would
be
a
fun
exercise
and.
D
It's
actually
because
you
mentioned
a
comparison
we
should,
I
will
write
it
afterwards
in
the
channel,
because
raphael
rocklets
also
added
empty
capabilities
to
custom,
so
he
might
also
be
interested
in
in
participating.
F
Absolutely
okay,
then,
the
last
large
email
that
I
have
copied
a
chunk
in
is
there's.
This
cooperate.
Em
group
and
I
link
the
title
to
the
the
website
and
there
might
be
more
info
on
there.
I
didn't
actually
look
yet
they've
got
a
slack
channel
and
right
now,
it's
kind
of
an
email
list
of
just
folks
who
do
electromagnetics
they've
been
running
seminars.
F
Basically,
gonna
do
a
bit
of
a
roundtable
of
folks
who
are
doing
stuff
with
marine
em,
who
have
either
funded
projects
or
projects
in
mind
that
want
to
do
mourinho
and
figure
out
kind
of
how
to
work
together
on
things
like
that.
So
I
don't
necessarily
have
you
know
a
specific
project
in
mind.
Others
here
might,
but
even
if
not,
I
think
that
it
would
be
worthwhile
if
a
few
of
us
could
could
try
and
be
there.
F
D
A
C
C
My
background
is
pretty
noisy,
I
guess
somebody's
cleaning,
I
kind
of
well
not
really
a
hackathon,
but
that
I
sort
of
created
that
a
small
hackathon
with
theater
so
we're
trying
to
bringing
in
the
emg
3d
into
simpac.
It's
the
task
to
buy
this
weekend
was
developing
that
j
vector
product
in
emg
3d.
C
So
so
far
I
am
at
the
stage
I
tested
a
theater's
jt
back,
and
that
is
passing
our
second
order.
So
that's
good
and
we're
I'm
just
gonna
modify
his
jbac
function
into
a
g
like
a
j
t
back
function
into
a
j
back
and
yeah.
Hopefully
we
can
bring
that
in
like
once
that
passes.
It's
just
mechanistical
things
to
just
bringing
into
syntax,
so
yeah
it'll
be
kind
of
fun
to
actually
run
the
mourinho
inversion
in
in
syntax,
so
yeah.
D
C
That's
where
I
am,
and
what's
actually
interesting,
that
I
have
found
so
what
we're
doing
in
terms
of
jvac
and
jt
back.
We
are
first
discretizing
then
take
the
derivative,
but
the
sort
of
what
theater
is
doing
is
more
traditional
approach.
You
start
from
continuous
derivation
of
the
agent
state
and
then
discretize,
so
there
is
actually
an
advantage
just
a
bit
of
detail,
so
anybody
want
to
talk.