►
From YouTube: SimPEG meeting January 15, 2019
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
Cool
well
I
thought
we
might
try
meeting
notes.
I,
don't
know
if
that
is
a
useful
thing,
but
I
figured
it
might
end
up
being
pretty
useful
when
things
like
this
happen
and
like
saga,
you
can't
come
just
have
like
a
written
record
that
he
could
skim
through
in
five
minutes
and
see
you
know
what
we
talked
about.
So
if
someone
is
willing
to
take
notes,
that'd
be
awesome
because
I
find
it
hard
to
like
speak
and
keep
track
of
the
cameras
and
take
notes.
B
A
Yeah
so
aren't
there
I've
got
three
items
at
this
point,
so
talking
just
a
little
bit
about
the
plan
for
the
discs
or
ties
Docs
and
I'm,
hoping
to
get
some
input
from
Dom
on
the
upgrades
for
the
potential
fields
and
then
maybe
just
go
over
like
if
there's
pull
requests
that
need
motion
on
them.
I
there
anything
else
that
we
should
add
to
this.
A
A
Well,
let's
then
start
with
the
docs.
If
that
works
for
everyone,
yeah
sure,
okay,
I
am
just
going
to
share
my
screen
and
what
I'm
actually
gonna
do
is
go
to
the
Verde
Docs,
because
this
is
what
I've
been
following
in,
like
in
contact
for
our
Docs
I'll,
just
post
link
in
the
notes,
Oh
excellent
I
already
got
it
so
first
thing
to
notice.
I
think
is
like
the
the
subheadings
are
on
the
the
outline
here
so
getting
started
the
user
guide
and
then
the
reference
documentation.
A
So
what
we
have
currently
is
we
sort
of
have
like
the
user
guide
and
reference
documentation
all
muddled
together.
It's
like
the
API
is
sort
of
documented
in
some
of
the
usage
which
I
think
is
confusing
and
hard
for
people.
So
I
think
that's
one
of
the
first
things
to
note.
There's
some
really
nice
I
really
like
how
he's
done
the
user
guide,
and
so,
if
you
look
at
these,
these
are
actually
basically
generated
by
like
the
Sphinx
gallery,
and
so
it
is.
A
So
if
you
go
to
bairdai
datasets
checker
board,
we
can
then
see
like
what
what
the
docs
for
that
are
he's
doing
some
stuff.
That
I
think
would
be
nice
to
get
in
there,
but
were
perhaps
a
little
ways
out
and
that's
okay,
I,
really
like,
like
the
the
way
that
he
also
includes
which
examples
use
that
function
so
like
that
kind
of
thing
would
be
nice
to
get.
But
I
think
is
a
first
pass.
A
B
A
So
we
might
like
force
impact
will
probably
need
to
break
this
down
into
like
by
method,
so
inverting
mag
data
or
Meg,
probably
forward
and
inverse
for
each
of
the
different,
like
styles
of
inversion,
same
thing
for
for
a.m.
and
all
of
that
yeah
yeah.
So
we
might
have
like
sort
of
two
levels
here
so.
A
I
think
for
the
user
guide
we
need
to
like
try
and
get
in
the
head
of
somebody.
Who's
like
I
want
to
invert
time
domain
iam
data.
How
do
I
do
that
they
don't
actually
care
what
the
classes
are
right
and
then
here
in
the
reference
documentation,
that's
then,
where
we
actually
document
the
classes,
does
that
make
sense,
yeah.
C
A
So
what
I've
been
playing
around
with
for
the
discretized
docs
is
something
similar
so
having
a
getting
started.
What
did
this
content
was
stuff?
That's
that
was
already
there
doing
a
user
guide,
so
I
think
like
having
same
style,
where
we
just
show
like
here's,
a
tensor
mesh.
Here's
how
to
build
it.
Here's
what
the
origin
is
those
kinds
of
things
having
same
thing
for
a
cylindrical
like
having
one
for
each
mesh
types
so
having
curvilinear
mesh
octree.
All
of
that
having
a
bit
of
an
overview
of
finite
volume.
A
A
A
If
you
do
the
question
mark
in
the
Jupiter
notebook,
which
is
like
so
much
friendlier
than
looking
at
the
the
:
like
the
double
:
parameter,
which
is
in
the
the
current
standard,
sphinx-like
documentation
style,
so
I'll
try
and
get
some
of
the
documentation
ported
over
to
this,
but
I
will
need
help
from
other
people
to
get
all
of
it
and
we'll
just
sort
of
try
with
all
future
documentation
to
go
with
this
style.
Does
that
work
with
everybody?
Do
you
think
that's
makes
sense?
Is
that
something
you'd
support.
A
Cool
yeah,
so
I'll
get
like
maybe
one
one
module,
maybe
like
the
tensor
mesh
or
something
like
that
totally
cleaned
up,
and
then
we
can.
We
can
go
from
there
yeah
and
then
once
this
is
done,
then
doing
this
impact
Docs
I
think
will
be
timely.
Cuz.
This
impact
Doc's
are
just
a
bit
bigger
and
like
it's
easier
to
flush
stuff
out
in
dis,
critize
first
I
think.
A
A
A
C
A
It
doesn't
need
to
be
each
line,
but
like
what
is
hard
right
now
is
I
can't
tell
what's
an
important
change
versus
what
is
not
and
so
like,
because
the
optimization
is
such
a
core
part
of
the
package.
Just
trying
to
understand
like
where
are
the
important
changes
like
what
should
I
think
about
I
guess
I
was
having
trouble
just
comparing
these
two
things.
C
A
I
guess
like
so
it's
either.
You
know
like
half
hour
on
your
end,
or
it's
also
I
mean
because
this
is
a
core
part
of
the
code.
I
like
I
want
to
get
my
head
wrapped
around.
What's
changed
so
I
can
spend
some
more
time
and
like
just
try
and
send
you
specific
questions,
I
realize
it
is
just
CG,
but
it.
C
C
A
Fair
enough,
because
that
solve.
C
B
C
C
A
A
A
Yeah,
so
if
somebody's
willing
to
take
a
look
at
that,
that
would
be
that'd
be
helpful.
It
implements
getting
electric
fields
from
and
currents
from,
DC
fields
objects
which
we
did
in
current
or
we
didn't
have
before
or
data.
Sorry,
we
had
the
ability
to
get
it
in
the
fields,
but
we
couldn't
get
electric
field
data.
So
now
we
can.
A
C
A
A
I
had
a
really
good
chat
with
Matt
Rockland,
who
develops
tasks
and
just
getting
an
overview
of
like
desk
delayed
and
how
that
works.
When
we
have
like
multiple
things
that
need
to
be
delayed
and
you
can
actually
pass,
you
can
safely
pass
gas
delayed
like
functions
or
computations
as
arguments
to
other
desk
delayed
things.
A
C
A
C
A
C
A
Yeah
cuz
I'd
be
super
helpful
to
have
and
the
other
thing
I
was
chatting
with
soggy
over
the
weekend,
just
a
bit
about
the
simulation
class
and
try
and
provide
that
and
get
that
going
again
pretty
quickly,
because
I
think
that
that
will
actually
simplify
a
lot
of
the
parallelization
and
also
I
mean
because
we
need
all
the
classes
to
serialize
for
the
more
complicated
codes.
In
order
to
be
able
to
do
that
so
yeah,
oh.
A
A
A
C
A
C
A
Yeah,
okay,
so
that
will
take
there's
a
few
things
that
need
to
get
done.
We
need
to.
We
need
to
take
another
look
at
the
directives
and
be
careful
about,
because
one
of
the
things
that
I've
noticed
in
some
of
them
is
in
a
lot
of
cases,
we're
assigning
we're
like
randomly
attaching
properties
to
the
directive,
and
we
need
to
be
careful
about
that,
because
that
can
cause
confusion
and
like
in
some
ways.
That's
what
happened
with
the
confusion.
A
C
A
C
C
A
But
I
mean
in
order
to
do
top
level,
we
have
to
be
able
to
serialize
all
of
the
individual
objects
and
so,
like
I,
think
the
directives
are
reasonably
clean
place
to
start
it's
cuz
it's.
It
should
be
pretty
clear
like
what
the
exact
parameters
that
the
directive
needs
to
know
in
order
to
recreate
itself,
then
same
thing
like
we'll
repeat
the
pattern
for
optimization
all
of
that
sort
of
stuff,
yeah.