►
From YouTube: Create Group Conversation
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
Hello,
everyone-
this
is
James
Ramsey
from
the
creating
we've,
also
I,
think
doubt
was
on
here
as
well.
The
slides
are
linked
in
the
Google
Doc,
which
I'll
quickly
paste
in
the
chat
Channel
and
our
Francis
has
already
put
a
few
questions
in
there,
which
I've
answered
asynchronously.
A
A
The
question
directly
is
take
a
look,
the
slides,
and
that
has
an
excerpt
from
the
categories
page,
and
so,
if
you
have
questions
what's
in
the
create
stage
and
the
DevOps
lifecycle
versus
the
create
group
and
the
slides
will
answer
that
and
also
explain
that
there
will
soon
be
no
longer
a
create
group,
so
the
create
stage
will
be
different.
So
there
are
five
groups
within
the
create
stage
source
code
which
is
going
to
be,
which
is
repositories
and
merge
requests.
There
is
knowledge
which
is
design
management
and
wikis.
A
There
is
editor
which
includes
the
web,
IDE
live
coding,
snippets
and
search,
and
then
there
is
the
giddily
group
and
then
there
is
the
Deuter
group.
So
that's
a
long
preamble
to
answer
the
question
about
the
Ginley
group
and
we
are
working
on
h,
a
we're
currently
working
on
the
beta
of
high-availability,
which
is
working
on
a
naive
implementation
of
eventual
consistency
for
replication.
A
Once
we
have
that
in
place
and
we
can
start
testing
it,
then
we'll
be
able
to
refine
our
estimates
for
when
we
might
have
a
generally
available
version
that
a
customer
could
actually
use
for
high
availability
purposes.
Currently
we're
aiming
to
have
G
a
version
of
AJ
in
some
form
by
the
end
of
the
year,
but
there's
a
high
degree
of
uncertainty
around
that
estimate,
because
it
is
quite
a
complicated
project.
You
can
see
the
different
stages
of
the:
how
we're
playing
on
tackling
that
problem
in
the
epoch
that
I've
linked.
A
I'll
jump
down
to
the
get
a
question
get
also
a
different
group,
but
dinner
is
kind
of
a
separate
product
and
I
happen
to
also
be
the
interim
product
manager
for
Guetta.
It's
not
part
of
the
self-hosted
installation
of
your
lab
because
it
is
kind
of
a
separate
product.
It's
really
a
SAS
product
at
the
moment
where
you
go
to
get
a
dot,
IO
or
get
I
am
don't
I,
can't
remember
the
URL,
it
autocompletes
my
browser
but
yeah.
A
It's
it's
a
community
building
product
and
we're
trying
to
improve
it
and
increase
the
integration,
forget
lab,
and
but
it's
got
kind
of
a
different
direction
at
the
moment
in
terms
of
just
improving
it.
Retaining
the
customers
that
we've
got
and
I
think
ultimately,
in
the
coming
months,
and
probably
next
year
realistically,
will
start
to
really
build
deeper
and
deeper
and
more
interesting
things.
We've
geta,
intergate
lab
and
there's
interesting
opportunities
that
we
can
explore
in
the
future
around
using
it
as
a
way
to
escalate
conversations
from
asynchronous
or
synchronous
men
like
channels.
A
So
so,
if
there's
an
incident
or
an
outage,
we
could
use
that
as
an
interest.
Instant
messaging
solution,
I'm
sort
of
progressively
escalate
into
increasingly
real-time
forms
of
communication
and
then
move
back
down.
So
if
still
I
guess
early
days
for
what
we're
doing
with
that.
But
yeah,
probably
not
part
of
the
pitch
for
how
we
still
get
lab.
A
Control
I
would
love
that
I'd
love
to
be
taking
notes
and
markdown
in
actual
text
files,
rather
than
this
weird
Google,
Docs
format,
I
think
there's
a
lot
of
people
internally,
gitlab
that
probably
feel
the
same
I
think
Kies
the
product
manager
for
the
editor
team,
so
editor
group,
which
will
be
working
on
this
later
in
the
year.
So,
if
you
want
to
leave
some
comments
on
the
epic
I
think
we
plan
to
start
some
discussion.
We
still
haven't
worked
out
how
to
start
tackling
the
problem
yet,
but
hopefully
we'll
get
there
soon.
A
B
So
I
that
I-
that's
this
in
the
channel
late
last
night.
Actually
but
I
was
wondering
I'd
come
across
that
we
show
language
percentages
in
the
repository.
But
I
was
looking
at
the
repository
API
and
that's
not
easily
queer
book
queryable
so
that
an
administrator
could
say
of
all
my
projects.
What's
that
combined
a
percentage
or
how
many
projects
do
I
have
of
various
languages,
and
so
that
way
gave
an
answer,
but
I
just
wanted
to
verbalize
that
that
seems
cool.
C
Absolutely
and
if
there's
things
that
we're
not
currently
tracking
in
the
database
in
an
equitable
way,
let
us
know
what
you'd
like
as
a
tracker.
We
can
probably
make
that
work
right
now,
we're
only
tracking
things
where
we
kind
of
need
to
for
caching
purposes,
for
just
expensiveness
of
calculation,
which
is
why
we're
we
have
to
date
a
programming
language
stuff
in
the
database.
Otherwise
we
generally
refer
to
either
get
us
from
the
repo
directly
or
to
cash
it
in
Redis.
To
save
some
of
these
calls,
but
then
the
purpose
is
necessarily
to
allow
querying.
D
C
Sorry
I
missed
a
question
there.
You're
asked
about
specific
versions
of
languages.
People
are
using
correct,
I
mean
in
some
cases
you
can
get
that
information
from
files
in
the
repository,
like
that's
ruby
version
or
you
know
whatever,
but
we're
not
currently
tracking.
That
know
where
we're
going
off
heuristics
like
the
filename
extension
and
we're
going
off
things
like
batch
files
that
might
have
a
you
know
a
hash
thing
at
the
top,
which
indicates
what
language
to
written.
It's
not
smart
enough,
yet
to
look
at
hints
for
what
version
it
might
be
now.
A
Yes,
I
think
that,
as
you
as
you
point
out,
the
first
step
is
to
Detective.
Nhs
is
being
used
and
I
think
the
plan
for
how
we
implement
that
and
the
detection
method
that
was
proposed.
It
should
be
possible
to
back
port
that
to
previous
versions
in
the
same
way
that
we've
back
ported
the
rugged
patches
and
hopefully
I
think
we
can
make
that
a
non-breaking
change
and
safely
enable
that
the
patch
release
great.
E
A
There
are
a
lot
of
different
features
in
gollum
I.
Don't
know
that
I
think
we
need
to
actually
review
each
feature
before
we
enable
it
because
I
think
there
are
some.
There
is
possible
issues
in
how
they
work
and
I
know
that
it
also
hasn't
enabled
them
all.
So
we
should
probably
do
a
review
and
understand
what
they
are
and
the
way
they
behave.
They
may
introduce
security
issues
and
there
may
be
other
problems.
A
I
think
there
are
some
specific
ones
that
we
should
look
at
implementing,
but
in
terms
of
general
prioritization
I
think
the
wiki
has
been
relatively
low
on
the
priority
list.
We've
been
mostly
fixing
performance
issues
and
nasty
bugs
in
the
wiki
I
think.
Hopefully
we
will
be
able
to
find
more
capacity
in
the
future,
but
I
don't
have
a
firm
timeline
on
that
and
I
think
ky-
picks
up.
A
E
For
sure,
thanks
for
that
also
for
the
next
one
we
have
two
web
IDE
and
I
was
super
wrong
about
this,
because
I
I
told
you
like,
oh
well,
monaco
working
on
iOS,
we'll
just
wait
for
Microsoft
to
fix
that.
But
I
tried
it
it's
a
horrible
experience.
It's
basically
unusable
right
now.
Do
you
think
we?
How
hard
would
it
be
to
fix
some
very
annoying
things
on
iOS?
We're,
like
your
your
cursor
strokes?
Don't
come
true
and
things
like
that.
A
Fixing
the
error
keys
is
almost
impossible
if
you
go
to
any
from
the
keyboard
events,
if
you
go
to
any
other
editor
project
online,
that
has
like
a
web
interface.
There
are
people
complaining
about
this,
and
the
reason
is
that
the
key
press
events
from
the
keyboard
they
get
reported
to
the
operating
system,
don't
get
forwarded,
entry
into
the
browser,
and
so
there's
no
JavaScript
event
to
handle.
A
D
A
So
apparently,
various
code
online
I
had
no
chance
to
check
it
out,
but
I
was
chatting
to
spend
from
Thea
and
the
other
day
and
apparently
they've
done
a
bunch
of
work
to
make
there's
code
on
line
work
on
camels
and
they
market
it
as
such
and
and
it's
also
built
on
top
of
Monaco.
So
it's
likely
that
they
have
fixed
some
of
those
issues
themselves.
A
A
A
We
should
follow
up
on
them
now
that
they've
and
test
their
Scott
online
on
that
I
haven't
compared
it
to
our
idea
and
see
exactly
what
the
differences
are.
I
know
that
the
Theia
project
is
planning
to
do
the
same,
because
they've
got
the
same
frustration
because
they're
also
builds
on
moniker
and
it
just
doesn't
work
on
an
iPad
yeah.
A
So
it's
really
easy
to
enable
and
get
a
connection
and
then
after
threading
we'll
continue
to
work
on
improving
the
native
integration
between
yet
lab
and
get
er
so
that
there's
better
rooms
and
Kishin
linking
kind
of
like
there
is
to
github,
where
you
can
link
to
much
requests
and
issues,
and
you
get
nice
unfilled
links.
Okay,.
E
C
Yeah
I
mean,
of
course,
Jenny
mastic
search
up
and
running
for
every
project.
Think
it'll
come
is
the
goal,
but
there's
one
really
big
hurdle
in
our
way,
which
is
just
the
sheer
index
size
of
Xing.
All
these
repositories
right
now,
if
we
index
all
of
the
git
repositories,
the
total
size
of
the
index
would
be
about
sixty
percent
of
the
total
size
of
the
bear
upholsteries
themselves.
C
So
if
we
were
to
enable
annalistic
search-
and
you
know
create
all
these
indexes,
we
would
certainly
get
a
much
much
higher
data
usage
bill
or
rather
storage
usage
bill,
but
that
is
that
epic
number
three,
which
is
links
from
the
slice
as
well,
where
we
are
looking
into
ways
of
reducing
that
size.
When
we
get
there,
we
will
also
need
to
figure
out.
How
are
we
going
to
shark
this
efficiently
effectively?
C
One
of
the
things
that's
it
proposed
a
little
while
ago
is
to
dizzle
for
group
basis,
which
is
one
of
the
options
being
considered,
but
some
concerns
of
wolves
have
been
raised
there
and
then
there's
another
issue
that
talks
about
the
general
problem
of
English.
Starting
being,
you
know
badly
distributed
highly
sued
in
its
distribution,
so
our
group
is
one
thing
or
considering
but
more
discussion
that
will
take
place.
C
But
our
first
kind,
of
course,
on
the
elasticsearch
topic,
are
getting
this
NPC
up
on
the
get
up
of
comb
with
only
to
get
the
workgroup
and
get
that
comb
group
as
well.
If
we
can
make
that
work
and
then
next
up
improving
the
a
paint
controls,
because
a
lot
of
our
customers
who
are
using
elasticsearch
have
been
have
been
telling
us
that
their
experience
isn't
so
great
because
we
haven't
been
dogfooding
it.
So
once
we
start
dr.
C
putting
it
on
get
the
comb,
we
can
focus
more
on
actually
getting
those
ethnic
controls
into
a
place
where
they
can
be
used
to
manage
an
elasticsearch
installation
and
get
that
integration
of
this
skill.
And
then
the
next
step
is
once
those
are
in
place
and
once
the
customer
experience
is
better,
is
to
kinda
keep
focusing
on
getting
this
rollout
for
all
of
them.
But
there's
still
some
discussions
there
on
that
rods
and
if.
E
They
might
uninformed
opinion
that
you
can
ignore,
but
I've
heard
a
lot
of
other
founders
that
have
massive
massive
problems
when
they
try
to
skill
elasticsearch
to
pick
sizes,
and
they
tend
to
have
like
teams
of
people
to
take
care
of
that.
So
having
kind
of
treating
it
more
like
Italy,
where
you
just
have
multiple
servers
that
have
that
mind,
their
own
business
might
be
a
way
to
avoid
that,
because
elasticsearch
is
great
as
long
as
the
index
size
is
not
overly
broad
and
I.
C
That
is
true,
although
of
course
that
might
change,
and-
and
we
also
have
customers
of
course,
who
want
to
at
least
do
these
kinds
of
searches
on
their
own
complete
instance,
and
we
have
customers
with
very
large
data
sets
as
well,
so
we'll
need
to
figure
it
out,
and
even
though
they've
come
to
dog
food
is
at
scale.
We
might
actually
need
elasticsearch
experts
on
board
to
manage
that
thing
as
something
that
makes.
E
Since
I
think
we
should
start
hiring
for
the
last
six
years,
experts
anyway,
because
we'll
need
that
anyway,
consider
if
it's
the
difference
between
the
right
and
the
read
load,
I.
Think
right
now
we're
super
read
heavy.
So
even
if
we
have
multiple
charts,
we
can
just
send
the
query
to
all
the
different
charts
that
will
increase
the
read
load,
but
our
lassic
search
is
really
good
at
dealing
with
to
read
load
I.
Think
for
us,
the
write
load
is
always
going
to
be
the
problem.
C
That
sounds
reasonable.
If
you
don't
mind,
I'll
jump
to
Dylan's
sub
question
and
yours
I
have
reconsidered,
allowing
people
to
configure
a
URL
to
their
own
elasticsearch.
We
do
actually
allow
that
it
doesn't
need
to
be
an
elastic
search,
bundled
with
omnibus
or
anything.
You
can
have
your
own
research
cluster
set
up
somewhere.
People
can
now
oh
you're,
saying
actually
only
get
that
of
come
level.
I'm
assuming
villain
for.
F
A
E
On
people
we
should
just
make
sure
get.
Lancome
has
good
yeah.
We
got
krenov
gotten
used
to
not
having
good,
searching
and
think
it's
acceptable.
Look
we're
most
people
use
it.
Storing
their
code
searching
to
code
is
a
very
basic
thing,
and
we
should
not
see
this
as
some
optional
thing,
we're
adding
it's
basically
a
service
that
is
down
right
now.
We've
gotten
used
to
it
because
it's
been
down
for
so
long,
but
we
have
like
the
the
normalization
of
deviance
I'm,
probably
messing
that
up,
but
anyway
we
shouldn't
get
used
to
this.
E
A
A
Looking
at
issues
to
see,
if,
like
a
library,
is
being
used
on
another
project
and
looking
at
their
issues
to
see
if
someone's
found
to
work
around
so
really
great
global
search
is
going
to
be
critical
if
we
want
to
actually
be
a
place
where
what
people
want
to
put
their
open-source
projects,
because
one
reason
you
put
them
on
a
SAS
platform,
isn't
it
people
can
find
them
so
Kenny
I
think
you've
got
the
next
question.
Yeah.
B
D
C
A
On
the
code
review
front,
they
do
a
really
great
job
of
assigning
reviewers,
which
is
slightly
different
to
approvals.
We've
put
a
lot
of
effort
into
having
great
approval
rules
for
policy
and
access
control
to
prevent
things
merging
without
the
right
people
reviewing
it,
but
the
actual
process
of
assigning
and
allocating
one
of
those
approvers
to
do
a
review
is
something
that
we
don't
do
well.
We've
got
reviewer
roulette
in
danger
and
we
really
should
work
to
make
those
the
first-class
feature.
The
way
github
is
done
with
some
of
those
features.
I
mean.
A
So,
and
one
of
the
reasons
for
that
is
that
having
one
ephemeral
for
each
security
issue
doesn't
necessarily
solve
the
release
problem,
because
you
still
need
a
stable
branch,
that's
secret,
and
so,
even
if
you
have
the
the
fix
on
an
ephemeral
for
at
point,
you
have
to
merge
that
into
a
release
branch
actually
deploy
it
and
it's
the
only
public.
The
only
sort
of
main
repo
you
have
is
the
public
one.
It's
difficult
to
actually
merge
it
into
a
stable
branch.
A
So
we're
sort
of
I
guess,
starting
with
a
workflow
that
addresses
the
gitlab
requirements
of
regular
cadence
of
security
releases
where
they
some
accumulate
many
fixes,
and
then
we
release
them
together,
but
will
need
to
mature
and
evolve
a
security
release
process
with
confidential
merge
requests
to
better
accommodate
smaller,
open
source
projects
and
ephemeral.
Forks
is
an
important
part
in
allowing
non
project
members
or
people
that
don't
have
access
to
that
security
fork
with
all
the
fixes
in
it.
A
So
a
security
researcher
that
contributes
a
report
might
want
to
do
the
fix,
but
in
doing
the
fix
we
shouldn't
give
them
access
to
every
other
vulnerability
report.
So
there's
some
interesting
differences
and
approaches
that
we
can
work
on
there.
Those
are
the
ones
that
immediately
come
to
mind:
snippets
Thank,
You
Gabriel.
You
pointed
that
out,
our
snippets
quite
different
to
have
github
is
implemented
gist's,
as
have
one
support
for
a
single
file
and
our
own
version
and
stored
in
the
database.
A
A
Think
this
meetings
meant
to
stop
at
25
past,
and
it's
now
26
passed
by
my
clock.
If
there's
any
more
questions
feel
free
to
reach
them
in
the
ask
them
in
the
create
slack,
Channel
and
I'll
do
my
best
to
answer
them
so
and
others
in
the
create
team
as
well
like
dower
and
Andre,
who
is
in
here.
Thank
you.
Everyone
have
a
great
day
thanks.