►
From YouTube: 2019-10-02 PHP SIG
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).
C
C
B
C
B
A
F
B
G
C
B
H
I
was
mostly
in
the
camping
in
the
parks
instead
of
Utah,
so
I,
don't
think,
there's
really
any
living
things
there.
That's
thanks!
There's
rattlesnakes!
That's
a
good
point:
I
didn't
get
eaten
by
rattlesnake
or
stung
by
a
scorpion.
I
think
the
only
I
did
see
in
klm's
National
Park
on
the
way
from
Des
Grieux
I.
G
H
H
It's
a
fun
trip
by
the
way
you
might
want
to
update
the
you'd
also
sent.
There's
two
meeting
invites
for
this:
there's
the
one
on
the
Oklahoma
tree,
public
calendar
and
I.
Think
one
you
just
sent
to
people
who
you
knew
were
attending
the
one
to
the
people
who
you
knew
her
attending
has
a
hangouts
meeting,
which
is
where
a
few
of
us
were
until
we
realized
everybody
was
like
I
think.
B
B
C
B
B
E
E
E
J
Hi
everyone,
my
name
is
Emanuel
a
decade.
I
work
at
a
rich
Tech,
Inc
I
worked
in
open
science
and
since
pretty
much
day,
five
or
so
worked
in
go
in
Python
helps
out
with
the
agent
help
right
the
agent
and
also
help
you
some
open
census.
Php,
metrics
design
and
I've
worked
with
Morgan
and
few
other
folks
on
this
call.
K
L
M
B
M
B
D
You
next
I
am
Brian
Aikens
or
bacons,
as
I'm
known
in
the
world.
I
work
with
Bob
and
I
do
more
architectural
and
that
type
of
stuff
I
do
a
little
bit
of
hands-on
keyboard
with
open
census,
mostly
in
other
languages,
so
I'm
interested
in
PHP
getting
up
to
the
level
of
those
where
you're
gonna
be
mixed.
B
I
think
that's
every
I
think
that's
everyone.
Thank
you.
All
I
know
that's
one
of
the
things
that
introverts
tend
to
hate
the
most,
but
I
really
appreciate
everybody
handling
that
next
we're.
The
next
thing
I
had
on
the
agenda
was
to
talk
about
the
current
state
of
tracing
and
PHP
in
open
census.
We
just
opened
the
open,
telemetry
PHP
repository
last
week,
but
we
have
the
only
thing
that's
been
actively
added
to
it.
B
Is
it
read
me
talking
about
this
meeting
so
I
wanted
to
make
I
wanted
to
see
what
everybody's
a
opinion
was
not
all
at
once
please,
but
we
need
done.
I
wanted
to
talk
about
our
current
state
with
open
census
and
open
tracing,
and
then
after
we
do
that.
Maybe
we
could
start
talking
about
how
we
envision
creating
the
open,
telemetry,
libraries
I'm
sure
there
are
people
here
that
have
somewhat
opinions
on
this
and
what
Bob
do
you
them?
I.
F
D
F
C
Yeah
I
can
give
it
like
a
snapshot
of
the
current
state.
So
obviously,
this
project
came
from
the
desire
to
merge
open
tracing,
no
consensus
into
a
single
project,
which
we
had
a
lot
of
success
with
so
far,
mostly
what
we've
been
doing
is
trying
to
write
a
new
specification,
so
we
spent
the
summer
trying
to
get
across
language
specification
together,
while
kind
of
bootstrapping
up
all
of
these
different
language,
implementations
sort
of,
at
the
same
time
just
to
get
everyone
kind
of
moving
and
rowing
together.
C
That's
mostly
been
a
pretty
straightforward
work.
There's
some
areas
where
we
really
wanted
to
clean
things
up
in
particular,
metrics
was
an
area
where
we
decided
more
recently.
We
wanted
to
maybe
take
a
different
approach
than
what
had
been
in
census
prior
to
that,
so
that
was
bogged
in
through
two
and
a
Josh
from
light
step
who
were
mostly
holding
that
down.
C
So
we're
hoping
to
get
a
draft
of
the
spec
at
the
door
that
we're
calling
point
to
we're
hoping
to
get
that
out
at
the
end
of
this
week
and
asking
every
sig
to
try
to
make
a
release
of
their
software
that
matches
that
point
to
version
of
the
spec.
There's
a
bit
of
massaging
and
bike
setting
around.
Like
literally.
How
do
we
talk
about
versions
in
this
project?
That's
going
on
right
now,
but
for
the
most
part,
what
we're
looking
at
is
releasing
a
point.
C
Two
right
now
doing
a
bit
more
spec
work
that
we
already
know.
We
want
to
do
releasing
a
point,
three
getting
feedback
from
these
early
adopters
and
trying
to
actually
do
instrumentation,
releasing
a
point
four,
and
hopefully
that
will
be
enough,
we'll
see,
but
hopefully
we'll
be
able
to
after
a
point,
four
go
to
something:
that's
more
like
a
stable
beta
and
we're
hoping
to
hit
that
by
end
of
year,
at
least
for
some
initial
set
of
SIG's.
C
Now
it's
notable
that
you
know
some
of
these
groups,
PHP,
obviously
Ruby.
You
know
some
other
communities
they're
starting
later.
Obviously
you
all
are
going
to
be
on
a
slightly
different
timeframe,
but
that's
sort
of
where
the
overall
timeframe
for
the
project
is
right.
Now,
one
thing
we
want
to
really
make
sure
we
emphasize
is
it's
critical
that
there's
new
hotness
that
we're
developing
and
we're
excited
about
our
new,
cleaner
api's
and
everything
it
still
needs
to
be
backwards,
compatible
Hunter,
busines
backwards,
compatible
with
existing
instrumentation
for
open
tracing
and
open
census.
C
We
want
to
make
sure,
even
though
we're
building
something
new
that
we
might
like
better,
we
don't
want
to
create
some
kind
of
you
know
Python
3
moment
where
there's
like
this
chasm.
People
have
to
jump
in
order
to
hop
over
so
we'll
also
be
making
sure
over
the
next
couple
of
months
that
all
of
that
stuff
is
working
properly
before
we
announce
like
a
big
beta
and
really
try
to
get
people
in.
C
So
that's
sort
of
the
next
set
of
goals,
we're
getting
the
website
together
and
we'll,
hopefully,
have
all
this
information
up
in
a
more
public
like
condensed
form
for
people
to
kind
of
self-serve,
but
that's
the
current
state
of
the
project.
Are
there
any
questions
about
like
high
level
project
stuff?
At
this
point,
I'm
happy
to
clarify
any
details.
F
So
you
know,
if
there's
any
questions
than
that,
perhaps
you
know
maybe
example
that
Christian
Sharia
starts
and
how
they
started
on
the
iPhone
se
unless
folks
on
the
call,
Bob
etc
know
you
know
you
have
fairly
good
understanding
of
that,
and
that's
fine-
and
you
know
we
get
a
lot
done-
is
that
but
I
just
recollected.
Somebody
had
asked
this
question
so
I
figured
I'd
raise
it.
M
Sure
it's
I
think
there's
like
three
things
here.
What
is
it?
The
spec
describes
the
API,
so
I
don't
know
if
they
can
mention
for
doing
this
in
PHP
is
but
we've
just
got
two
separate
packages
in
Python
we've
got
open,
telemetry,
API
and
a
lemon
tree
SDK,
and
you
can
basically
just
follow
the
spec
writing
the
API
package
right.
So
it's
just
like
a
set
of
interfaces
that
the
SDK
implements.
So
there
are
some
caveats
to
that.
The
API
actually
includes
a
full
implementation
of
context.
M
M
Also,
we
don't
you
know
so
that
we're
going
for
full
backwards
compatibility
and
in
open
Salama
tree
with
open
tracing
and
open
census,
but
the
way
that
we're
gonna
do,
that
is
with
a
bridge
in
open
Salama
tree.
So
it
actually
doesn't
need
to
be
backwards
compatible
right
from
the
beginning,
so
we
actually
haven't
written
the
bridges
yet
in
in
the
Python
repos.
M
This
is
what
we're
working
on
now,
but
basically
the
the
goal
is
we
also
switch
to
Python
3,
which
is
a
you
know,
a
breaking
change
and
open
Salama
tree
yeah,
good
good
for
some
bad
for
others,
and
so
this
means
that
we're
gonna
have
two
releases
of
open
census,
so
we're
gonna
have
one
that
is
still
backwards
compatible
of
two
seven
and
that
will
it
will
be
compatible
with
the
bridge
that
we're
writing
in
open,
telemetry
and
then
we'll
have
a
last
release
which
drops
support
for
two
seven
goes
straight
to
three
six
and
then
basically,
we
just
hollow
out
the
open
census,
library
and
everything
calls
through
to
like
the
new,
open,
telemetry
library
and
that
will
actually
depend
on
the
SDK
because
an
open
census.
M
M
H
H
H
C
I
can
start
I've
been
a
little
slow
to
reach
out
to
people,
so
apologies
for
that,
but
I
can
certainly
reach
out
to
the
open
tracing
community.
There's
you
know,
people
like
you
know:
Jose
Carlos
and
whatnot
who've
been
working
on
PP
tracing
and
zipped
in
and
open
tracing
and
I'm
sure
be
willing
to
contribute
yeah.
As
far
as
like
big
shops,
yeah
I
can
do
some
research
there
I
believe
like
box
is
like
a
big
PHP
shop.
C
H
C
M
C
F
Does
anybody
have
any
other
questions
related
to
the
sig?
One
of
the
things
that
this
exact
over
here
do
is
you
know
they
meet
I
think
once
a
week
at
least
one
of
the
frequency
that
is
conducive
to
order,
the
city
is
like
I,
think
ruby
is
one
of
the
the
last
ones
that
started
before
PHP,
so
I
know
they
have
something
to
be
district
Japan,
so
they
meet
very
different
times.
One
of
you
guys
ain't.
F
Let
me
good
morning
evening,
but
in
terms
of-
and
this
may
be
towards
Bob
and
Greg,
you
know
you
get
some.
He
knows
that.
Do
you
foresee
maybe
one
way
to
sort
of
get
the
conversation
a
little
bit
more.
You
can
leverage
the
folks
on
the
phones.
Count
is
at
this
point
in
time.
Do
foresee
any
areas
of
challenge
where
you
see
in
terms
of
starting
the
same
or
you
know
the
next
step
should
get
started
for
the
next
steps.
Maybe
that's
in
there.
H
Couldn't
some
of
the
other
things
when
we
started
some
of
the
things
just
started
with
either
the
like?
They
copied
the
code
over
from
either
open
tracing
or
open
census?
I
wonder
if
we
just
want
to
do
the
same
thing
here,
just
to
start
with
like
copy
over
like
the
open
sentence,
PHP
code,
because
it
works
and
then
start
jamming
on
that
yeah.
C
C
F
F
M
So
we
didn't
actually
copy
the
open
set
this
code
over.
We
just
started
from
scratch
with
the
API,
so
like
the
the
spec
mostly
describes
the
like
the
API
package,
and
so
we
just
wrote
that,
starting
with
the
tracer
and
span
creation
and
like
from
there
added
everything
we
needed
to
make
like
a
couple.
M
Basic
tests
work
so
like
the
spec
doesn't
really
respect
ascribes
the
API,
but
it
doesn't
like
fully
describe
the
behavior
of
the
library,
because
that's
that's
up
to
the
implementation,
but
if
you've
worked
with
open
census,
you
basically
know
how
tracing
works.
Right
like
you
need
to
be
able
to
create
spans
either
like
from
an
application
or
from
a
library
that
doesn't
depend
on
the
SDK.
M
The
Java
repo
is
also
like
the
the
farthest
ahead
of
any
of
the
languages,
so
like
often
we'd
switch
back
and
forth
between
the
spec
and
the
Java
implementation,
but
they
were
also
written
at
the
same
time.
So
if
you
go
and
like
look
at
the
history
of
the
spec
repo
you'll
see
that
a
lot
of
it
is
Java
specific
and
so
a
lot
of
like
the
early
work
that
we
did,
and
you
know
the
other
like
non
Java
languages
did
as
we'd
write,
something
you
know.
We'd
say
well.
M
This
doesn't
make
sense
in
our
language
I
and
say
you
wouldn't
always
use
the
Builder
pattern
or,
like
you,
don't
have
protected
like
class
members
like
this,
so
you
wouldn't
you
wouldn't
bother
putting
this
in
the
spec
and
I
suspect
you'll
do
the
same
thing
in
PHP
right
so
like
as
we're
implementing
it.
We're
also
still
defining
the
spec,
and
you
see
that
tracing
is
like
pretty
well
described
right
now,
but
metrics
is
still
like
very
much
up
in
the
air
so
as
as
you're
writing
this
following
the
spec.
M
If
they're
something
that
doesn't
make
sense,
you
know
don't
just
follow
the
spec
blindly
like.
Please
also
go
change
the
spec.
If
you
can
see
like
just
look
through
the
pull
requests
and
Python.
If
you
see
anything
here,
that's
like
surprising,
you
can
just
like
yeah.
Send
me
a
message
like
one-on-one
later
and
I
can
tell
you
about
any
of
them,
but
we
basically
just
started
from
the
span
creation,
API,
the
tracer
API
and
then
like
built
it
out
from
there.
So
we're
also
not
done
yet
right.
M
We
have
an
alpha
that
we
just
released
yesterday
and
the
day
before.
Yesterday
now-
and
it
includes
like
a
basic
metrics
implementation
and
a
largely
working
like
tracer,
but
it
doesn't
have
any
export
or
anything
yet
so
you
know,
basically,
every
implementation,
except
for
Java
is
still
trying
to
catch
up
with
the
spec,
so
I
expect
the
spec
is
gonna,
be
changing
a
lot
like
while
we're
a
mobile
writing
this,
but.
C
And
something
I've
seen
in
other
groups,
including
Java,
was
yeah.
The
first
implementation
people
broke
was
more
like
a
mock
or
fake
or
in
memory.
Sdk
right,
like
you,
know
something
that
they
can
use
to
test
the
API
and
then
once
they
got
that
yeah
working
to
a
point
that
they're
happy
with
then
started
porting
over
the
SDK
implementation,
or
you
know,
writing
it
from
scratch.
While
looking
at
the
Oh
consensus,
implementation,
essentially
or
whatever
existing
implementation,
I
know,
oh
I'm,
forgetting
your
name
nakusoo.
C
K
G
B
K
B
B
Know
that
we
were
looking
at
I
was
looking
about
the
last
week
and
there's
if
we
could
just
take
your
repository
and
essentially
create
a
PRI
using
your
repository
as
the
starting
point
for
this
repository
I.
Think
I
think
that
gets
us
from
zero
to
something
rolling
and
then
we
can
take
all
because
it
could
take
it
from
there.
C
C
Maybe
just
in
particular
a
review
relative
to
the
spec
that
might
be
where
a
place
where
maybe
some
of
us
who
have
been
working
on
this
from
other
languages.
Maybe
you
don't
know
PHP
but
know
the
spec
could
do
a
pass
at
that
and
just
see
if
there's
anything
in
there
where
we're
like.
Oh
that's,
it's
not
quite
what
the
spec
says.
Something
like
that,
but
you
know
that
sounds
like
a
great
place
to
start.
I
About
the
stats
in
PHP,
the
design
involved
that
sidecar
to
go
and
then
HP
would
write
the
metrics
over
used
to
being
socket
and
then
that
would
poured
it
on
to
the
host
C
engine.
M
But
one
thing:
that's
that's
very
different
about
like
metrics
as
its
described
right
now
and
specs
are
actually
is
not
even
in
specs,
yet
is
still
in
Josh's
like
PR
in
specs.
This
is
that
we're
not
we're
not
handling
aggregation
in
views,
the
way
that
we
didn't
open
sizes,
so
you
know
so
tell
me
if
this
matches
what
you're
doing
in
PHP
but
like
we
have
in
open
census.
Python
this
view
layer,
and
then
you
register,
a
bunch
of
metrics
and
a
metric
doesn't
have
any
aggregation.
M
Just
has
a
type,
and
then
you
have
a
bunch
of
types
that
apply
to
certain
or
you
have
a
bunch
of
aggregations
and
fly
to
certain
types
of
metrics.
So
if
you
have
like
a
inch
measurement,
you
can,
you
know
aggregated
into
a
histogram
or
whatever,
and
then
we
export
those
aggregated
measurements
right
now
in
open
telemetry
there
is,
there
is
no
solution
for
aggregation,
so
all
we
do
is
collect
metrics
and
then
like.
M
There
are
three
classes
of
these
they're
like
counters,
which
are
I,
think
effectively
gauges
and
open
census
and
then
they're
gauges,
which
are
kind
of
with
some
other
properties.
And
then
there's
like
an
observer
gauge,
which
you
can
attach
to
like
some
callback
and
then
there
are
raw
measurements
which
collects
all
of
the
information
about
the
like
the
thing
that
you're
you're
recording
and
then
those
presumably
get
aggregated
later.
M
But
we
don't,
we
describe
how
we're
actually
going
to
aggregate
those
yet
so
those
metrics
I
think
makes
sense
to
if
all
you're
doing
is
collecting
those
and
not
a
you're
getting
them
in
the
client.
I
think
it
makes
sense
to
just
send
all
of
those
over
to
the
agent
and
then
plan
to
do
the
aggregation
in
the
agent
and
it's
probably
the
first
implementation
that
will
actually
have
the
aggregation
later
written.
So
I
think
that's
that's
still
like
a
really
good
plan.
If
you're,
okay,
with
having
this
go,
sidecar.
C
Think
the
idea
with
the
collector
is,
you
could
run
it
remote
or
local.
I
mean
I,
don't
think
the
collector
particularly
cares
where
it
runs.
In
fact
they
had
two
names
for
a
collector
agent.
That
mostly
meant
whether
it
was
a
sidecar
or
not.
But
the
really.
The
only
difference
there
is
from
I'm
aware,
is
that
you
know
whether
or
not
you've
enabled
collecting
of
machine
metrics,
but
what
you
said
actually
this
gets
into
where
maybe
PHP
is
slightly
different
right,
because
PHP
is
usually
a
sort
of
hosted
language
right.
C
It's
like
request
level
context
and
there's
usually
some
kind
of
server,
that's
embedded
in
it.
That's
like
managing
a
you
know,
pull
of
PHP
processes
and
there's
like
a
variety
of
ways
of
doing
that.
A
certain
amount
of
like
state
and
contexts
are
kept
not
in
PHP
directly
but
in
C++
somewhere,
and
it's
been
a
long
time
since
I've
touched
PHP
now
five
point
three
forever,
but
I'm
curious.
C
J
Perhaps
I
could
help
under
Greg's
question
Greg
that
extra
sidecar
that
has
that
uses
UNIX
domain
sockets.
Our
plan
is
to
move
that
action
to
the
open
sentence
agent.
We
just
had
it
work
Jenny,
but
the
idea
will,
if
we
move
that
to
the
open
sense,
it's
agent
slash
collector.
Now,
then,
all
PHP
doesn't
just
connect
to
that
receiver,
because
the
agent
has
this
concept,
which
you
write.
So
that's
a
way
we're
able
to
keep
in
except
15
traces
or
Jaeger
traces
and
metrics
from
Prometheus.
J
I
I
C
C
A
C
B
I
This
case,
that's
the
mention
folks
how
they've
run
each
should
be
over
there
and,
if
being
used
in
trading,
collect
Eric
losing
that
receiver
to
the
collectors.
Room
works
for
your
architecture,
so
it
seems
like
this
will
work
better
the
war
or
looking
the
same
thing,
and
you
know
fixing
up
the
same
paper
Road
here
so.
B
To
answer
your
question:
Greg,
we
don't.
Currently
we
have
a.
We
have
a
bunch
of
spikes
that
we've
done
to
try
and
figure
out
the
appropriate
methodology
for
implementing
open
census
in
our
monolith,
and
essentially
what
we've
got
now
is
just
using
the
zip
blizik
and
exporter,
and
that
pushes
things
up
to
stack
driver
currently
and
it's
pretty
crude
and
we
were
looking
for
a
better
solution
to
it.
B
Currently
because
I
mean
it's
just
been
a
lot
of
it,
but
we
just
haven't
really
been
able
to
put
enough
effort
into
it
yet,
but
I
think
that's.
That's
part
of
the
reason
we
wanted
to
get
involved
in
this
initiative
is
that
we
can
make
sure
that
we're
making
the
right
determinations
and
doing
the
right
things
as
we
move
forward.
C
C
C
The
chat
room
is
also
you
know
useful
way
to
like
talk
about
stuff,
but
generally,
if
it
starts
getting
into
the
weeds
of
like
you
know,
what
should
we
be
doing
or
is?
Should
it
be
this
approach
or
that
approach
trying
to
move
that
out,
till
I
could
get
hub
issue?
It's
been
helpful,
Chris
I,
don't
know
if
you
have
any
advice
about
like
backlog
management
that
you
found
helpful
or,
if,
like
we'd,
be
just
enough
for
people
to
kind
of
look
at
what's
going
on
in
Python
right
now,
do
you
feel
happy
with
like?
M
I
mean
it's
it's
pretty
project,
so
you
know
we
don't
have
a
lot
of
skeletons
in
the
closet
of
the
backlog.
Yet,
but
it's
you
know,
it's
also
the
case
that
our
backlog
has
mostly
just
grown
monotonically
as
we've.
You
know,
worked
on
the
project,
so
it
looks
basically
like
normal
project
development.
Right,
like
we
have
some
big
issues
which
we
just
tackle
in
PRS
and
then,
as
we
have
comments
on
the
PRS,
though
you
know,
suggest
big
changes,
we'll
just
put
them
in
an
issue
and
then
plan
to
handle
the
issue
later.
M
Yeah,
there's
a
there's,
a
milestones
doc
on
in
the
this
factory,
bow
I.
Think
I.
Think
Sergei
actually
turned
these
into
like
actual
github
milestones
now,
but
these
describe
the
the
things
that
they
expect
to
be
done
at
each
of
four
releases
in
the
spec
and
we're
just
trying
to
have
the
spec
like
implemented
up
through
each
one
of
those
milestones
for
each
one
of
our
releases.
So
we
just
released
zero
dot,
one
dot
zero,
and
that
includes
the
stuff
from
that
milestone.
So
this
is.
M
This
has
been
helpful
for
us
because
we've
just
said:
if
it's,
if
it's
something
that
needs
to
get
written
to
you
know,
say,
make
some
future
work
or
like
that.
We
split
out
from
a
larger
like
PR
if
it
doesn't
and
it
doesn't
actually
match
one
of
the
upcoming
milestones,
we'll
just
put
it
in
a
ticket
and
save
it
for
later.
But
of
course
this
is,
you
know
just
open
source
and
we
want
all
the
contributions
we
can
get.
M
So
if
anybody
feels
passionate
about
some
particular
issue,
it's
like
just
helpful
for
them
to
pick
it
up
right.
So
we
actually
had
this
happen
for
the
the
whisky
integration,
because
we
had
a
guy
Allen
Feldman
from
New
Relic,
who
saw
that
we
were
writing
separate,
Django
and
flask
integrations
and
just
said
no,
you
should
be
doing
this
at
the
wizzy
layer
and
then
wrote
an
integration,
and
that
was
great.
M
It's
not
on
any
milestone,
but
it's
allowed
us
to
test
a
bunch
of
stuff
and
was
really
helpful,
and
then
we
had
Microsoft
folks
come
and
write
the
azure
exporter,
even
though
it's
not
on
the
roadmap
really.
So
it's
a
mixed
bag
that
people
are,
you
know
passionate
about
something:
that's
often
what
gets
written,
but
otherwise,
we've
been
trying
to
follow
the
milestones
yeah.
C
And
just
an
FYI
with
those
milestones
we're
actually
trying
to
get
them.
We
just
pushed
this
out,
so
this
is
new,
but
getting
them
to.
If
they're
written
in
the
right
format,
they
can
directly
drive
a
status
diagram
that
we're
putting
up
on
the
website
so
actually
I.
Think.
Just
yesterday
we
pushed
out
a
new
version
of
the
website,
which
is
very
much
a
working,
so
people
are
interested
working
on
that.
That
would
be
hopeful,
but
I
just
paste
it
into
the
meeting
notes.
C
This
like
project
status,
will
probably
move
this
to
the
front
page
once
it's
like
ready
to
go.
But
this
thing
every
time
a
website
will
get
billed
will
update
itself
based
on
what
milestones
have
been
passed
by
each
language.
So
there's
instructions
for
how
to
how
to
add
PHP
to
this
progress
graph.
B
B
A
B
C
Curious
what
open
tracing
supports
on
this
front?
I?
Don't
actually
have
the
information
at
my
fingertips
but
yeah
there.
There
is
definitely
this
balance
between
and
it's
possible
to.
Maybe
do
this
and
like
a
like
a
separate
implementation
or
something
but
I
can
see
you
know
longer
term
if
there
are
a
lot
of
shops
that
are
still
on
legacy.
Php
versions-
and
we
want
this
to
become
like
a
big
open
standard-
will
probably
eventually
have
to
deal
with
that.
C
B
C
B
C
For
example,
we
were
really
trying
to
support
Java
six
for
a
long
time,
even
though
that
kind
of
was
annoying,
but
you
know
Java
at
least
has
this
like
long
legacy
history.
So
you
know
if
you
want
JPMorgan,
Chase
or
somebody
else
to
like
go
run
this
stuff.
You
need
to
have
like
support
for
them
and
I'm.
It's
just
I
literally,
don't
know
what
that
that
kind
of
stuff
would
look
like
pythons
an
example
of
this
I
know
you
guys,
Chris
were
debating
about
like
how
to
approach
Python
to
support
right.
M
So
it's
century
IO
like
in
particular,
has
a
bunch
of
customers
who
are
just
never
going
to
move
from
Python
to
and
because
we
said
well,
like
you
know
three
six
only
for
open
telemetry,
it
means
they
are
gonna,
have
to
either
use
open
census
everywhere.
Try
and
ship
to
libraries
depending
on,
like
you
know
what
version
of
Python
their
customer
supports
or
what
they
might
actually
do
is
implement
open,
telemetry
themself
like
maintain
their
own
fork,
which
is
backwards
compatible
to,
and
you
know
that's
not
great.
M
B
C
Yeah
well,
I
could
follow
up
with
Jose
Carlos,
because
yeah
I
think
that
might
be.
You
know
if
open
tracings
been
out
for
a
while.
So
we
could
just
see
like
where
support
has
been
for
that
and
like
what
yeah,
what
kind
of
like
support
requests
people
were
getting
there,
but
you
can
imagine
that
whatever,
if
we're
gonna
say
like
people
who
are
using
open
tracing
currently
are
gonna
be
able
to
like
come
on
to
open
telemetry.
C
B
F
C
I
would
suggest
meeting
once
a
week,
at
least
until
things
yeah
and
maybe
there's
an
action
item,
I'm,
not
sure
who
wants
to
take
a
sort
of
lead,
maintain
a
role,
but
someone
may
be
tapping
themselves
to
at
least
organize
the
backlog
not
be
like
dictator,
but
like
just
hold
down
the
responsibility
of
at
least
to
get
going
around
like
seeing
the
backlog
with
stuff.
Cuz
I
predict
it'll,
be
probably
a
chicken-egg
thing.
Otherwise,
okay,.
B
M
B
That
I
did
that
already:
okay,
yeah
I.
It
was
mostly
me
adding
things
to
chat,
so
I
figured
that
just
there
at
the
bottom
of
the
the
open,
tslot
metric,
HP,
cig
meeting
agenda
notes
talking
with
that
right.
That
was
included
in
the
meeting
button
that
I
sent
out
also
not
a
meeting,
but
my
guess
is
my
prescribed
plan.
Right
now
is
to
have
this
meeting
every
week
at
this
time
in
this
place,
so
that
we
don't
confuse
everybody
if
that
works
for
everyone.
M
B
I'm
also
gonna
try
and
be
active
in
the
Gator
chatroom.
If
y'all
want
to
like
how
to
add
Hawk
talks
about
anything
I
know.
The
last
thing
you
need
is
another
chat
room
to
talk
somewhere
about
that,
there's
a
there's,
an
open
to
limit
route,
PHP
theater
room
and
I'll
link
to
that
at
the
bottom
of
the
document.
So
if
you
have
any
questions
ad
hoc
that
you'd
like
to
talk
about
I,
think
that
might
be
a
really
good
forum
for
it.
Yeah.