►
From YouTube: 2020-05-05 .NET 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).
A
A
A
A
A
A
The
biggest
update
base
will
be
starting
working
on
replacing
open,
telemetry
concepts
of
span
with
activity.
That's
the
biggest
update,
but
before
we
go
into
that,
we'll
just
address
the
open
agenda.
We
have
about
using
evasions
and
infer
improvements.
I
think
really.
What
was
it
you
who
mentioned
infra
improvements?
Is
there
anything
specific
which
you
wanted
to
discuss
or
it's
a
generic
open
item?
Yes,.
C
Oh,
so
there
are
two
things:
one
is
a
like
when
we
start
open
projects.
At
that
time,
people
explored
github
actions
and
at
that
time
did
hub
actions
was
still
a
preview,
not
very
stable,
so
we
decided
to
go
with
Chaves
CI
and
our
github
action
has
become
much
more
stable
and
it
brings
more
functionality.
I
wonder
if
this
is
something
we
should
consider,
because
currently
open
family
tonight
has
DJs
using
the
Eiger
pipeline,
which
is
not
like
something
that
a
lot
of
people
use,
especially
for
this
community
project.
C
A
C
A
C
A
D
A
D
B
C
A
little
bit
different,
so
github
is
more
advertised
as
a
as
a
neutral
thing,
so
Microsoft
edge
high
to
dump
like
this
company
specific
thing
to
github,
but
for
the
other
pipeline
it
has
a
lot
of
like
integration
with
Microsoft's
own
technology,
so
or
since
the
earth
project
I
think,
might
be
a
good
thing
for
this
overall
product
to
either
move
like
I'm.
Finally,
Cesar
Chavez,
circus,
yeah
or
github
actions,
but
for
other
pipeline
I
do
have
some
minor
concerns.
C
A
Now
it's
Kayla
wait.
The
way
will
find
me
how
yes,
unless
we
switch
to
a
passion
action
sometime
in
the
future.
Yes,
so
right
now,
the
instructions
remain
the
same.
The
VM
file
contains
the
see
a
setup,
or
it
also
contains
the
framework.
We
are
installing
whether
it
is
wrong
or
Linux
windows.
All
those
things
are
public
environment,
so
anyone
can
see
it
modified
in
their
peer.
A
A
A
Okay,
so
what
I
do
is,
let's
give
it
another
few
weeks
before
we
do
invest
time
in
shifting
it
away
from
the
current
way,
because
it's
mostly
working
except
few
glitches
here
and
there
yeah.
So,
let's
use
the
energy
to
filter
and
like,
as
we
are
getting
closer
to
beta,
then
we'll
have
revisit,
and
then
we
can
evaluate
should
be
moved
to
github
actions
or
not
yeah.
C
C
So
previously,
when
I
was
working
as
a
maintainer
in
Python,
we
ended
up
moving
or
the
vendor
specific
thing
outside
the
repo
and
only
export
her
we
left
inside
the
main
repository
was
zip,
King
eager
and
the
open
telemetry
protocol
and
the
reason
we
have
that
is
because
they
were
all
open
source
project.
Even
the
back
end
is
open
sourced
and
and
this
more
like
a
new
hosting.
Why
do
I
adopted
by
the
industry?
So
in
this
repo
I'm
saying
that
we
have
application
insights
like
stab
and
stackdriver?
D
A
C
C
So
there's
a
central
repository:
if
you
go
to
the
open
climate,
Dow
IO,
there's
a
link
where
you
can
get
all
the
registered
exporters
and
that's
a
recommendation.
So
in
general,
open
country
gives
customer
I
its
vendor
the
flexibility
either.
You
can
put
that
as
a
as
the
country
mikage
or
you
can
create
your
own
thing
and
you
don't
even
have
to
make
that
open
source.
E
C
D
Eat
it's
like
the
way
that
I
always
saw
this
vendor
specific
packages
were
while
we're
developing
just
to
make
sure
things
were
working,
although
having
zip,
King
and
Jager
were
probably
enough,
but
I
mean
if
it's
slowing
down
right
now,
it's
like
it
makes
total
sense
to
pull
it
out,
either
to
add
a
tag
or
just
fork
it
if
to
a
contrib
and
then
leave
it
there
until
like
white,
could
pull
out
the
light
steps
thing.
I
assume.
No,
no
he's
he's
not
at
the
call
paper
he's
active
on
the
dot
net,
SDK
and
third.
C
Some
changes
on
my
cans
and
I'm
not
sure
if
the
they
can
like
continue
work
on
this
or
not.
So
this
is
something
I
need
to
followup
with
Ted
twisted
so
Teddy
it
he
all
four
lights:
tab:
okay
and
I,
like
he
reached
out
to
me
yesterday
and
were
falling
out
home
like
what's
the
next
fad.
So
currently
we
have
some
funding
issue
here.
A
Okay,
so
the
action
item
is
like
I'll
post
him
get
your
channel
to
see
if
we
have
all
four
light
strip
and
struck
Trevor
and
we'll
be
care
east
a
week,
if
not
we'll
just
give
it
and
it's
still
in
the
history.
So
if
anyone
wants
to
like
resurrected
somewhere
else,
they
can
use
the
github
history
and
find
the
actual
code
and
place
it
somewhere
else.
Yeah.
A
Thank
you.
We
have
two
topics
to
discuss
today.
It's
more
like
an
update,
I
been
telling
like
for
the
last
two
weeks
and
we
going
to
shift
from
open
telemetry
rather
say
we
are
going
to
leverage
the
BCL
library
instead
of
creating
our
own
span
trade
surpluses,
and
there
was
a
blog,
a
couple
of
maybe
yeah
three
weeks
Bank,
which
contained
like
brief
idea
on
what
what
is
what
exactly
is
going
to
happen
and
there
was
a
link
to
the
actual
proposal
from
the
dotnet
team
I.
A
So
this
week,
I
am
going
to
actually
start
working
on
this
so,
like
I,
have
a
very
basic
PR
sent
yesterday.
If
anyone
had
to
look
at
it,
so
essentially
it's
going
to
pretty
much
delete.
Like
my
goal
would
be
to
remove
anything
which
has
tracer
or
span
in
it
and
replace
them
with
activity
and
activity
sources,
it's
going
to
be
like
fairly
huge
change,
because
it's
going
to
affect
everything
from
a
PA
SDK
to
exporters,
so
part
of
the
lease
we
want.
A
Exporters
to
be
moved
out
is
because
I
won't
have
the
energy
to
fix
each
and
every
exporter
that
that's
what
Reilly
was
earlier
alluding
to
when
he
wanted
to
remove
the
exporters
so
I
like
try
to
break
it
down
into
like
single
symbol
PL,
so
most
likely
we
won't
go.
I
can't
remove
everything
in
one.
Should
we
add
activity
as
like
to
live
along
with
spans
for
initial
piers,
and
once
we
get
the
shape
and
form
correct
thing
like
how
a
walking
way,
then
I
will
go
ahead
and
start
removing
spams
and
tracers.
A
This
would
almost
mean
that
we
won't
have
the
notion
of
an
open,
telemetry.
Dotnet
APA
like
there
won't
be
a
nougat
package.
There
will
be
a
package
for
matrix
and
other
purposes,
but
for
tracing
there
will
be
no
open,
delimit
or
APA
package.
So
if
someone
asked
like
how
do
we
instrument
it,
we
just
tell
them
you,
don't
it
be
CL?
So
that's
the
goal,
so
we
won't
be
shipping.
Anything
called
tracer.
Ip.
E
A
So
I,
like
posting
information,
how
that
transition
is
going
to
look
like
and
yeah
one
of
the
biggest
confusion
which
everyone
has
is.
Is
it
going
to
only
work
for
dotnet,
5
and
above?
But
the
answer
right
now
is
you
know
it's
going
to
work
with
everything
else
as
well,
because
the
package
which
were
depending
on
a
system
dot
diagnostic
diagnostic
source,
it's
a
out-of-band
package.
So
even
if
you
are
on
dotnet
framework
for
25
or
dotnet
core
2.0,
even
older,
you
can
just
install
this
package
and
you'll
be
good
to
go
and
open
telemetry.
A
The
SDK
which
we
ship
from
this
repo
we'll
be
taking
a
dependency
on
the
newest
version,
which
would
be
5.0
of
the
system
diagnostic
package.
Unfortunately,
that
package
is
not
yet
available
anywhere
in
public,
so
Riley
and
I
got
a
private
build
from
dot
netting.
Hopefully
it
will
be
like
somewhere
in
my
gaiter
nougat,
oh
yeah,.
C
I
can
share
more
details
on
that
so
sustain
the
general
direction
from
Donna
team.
Is
they
agreed
that
donut
in
the
future
would
try
to
be
compliant
to
open,
climb
trees
back
so
with
that
spirit
in
the
dollar,
at
five
timeframe,
we
agree
that
donut
will
change
its
tracing
API
to
be
compliant
with
open,
telemetry,
tracing
API,
and
so
like
customers,
don't
need
to
get
confused
about.
C
Should
they
spend
or
activity,
don't
let
him
will
make
changes
to
activity
to
make
sure
activity
can
come
where
dispensing
area
we
want,
and
with
that
like
activity,
contacts
will
be
spending
comebacks
activity
listener
will
be
something
similar
to
spam,
processor
and
they're.
Making
a
changes,
so
serial
should
already
have
that,
like
link
to
the
Donatists
pack
change
like
his
issue,
so
alright
him
I
made
a
change
and-
and
we
have
done
testing
inside
Microsoft
with
many
different
teams.
C
So
far,
we've
seen
like
like
two
percent
of
the
gap
is
not
a
hundred
percent
compliant
open,
telemetry
and
we're
still
making
huge.
So
meanwhile,
donut
is
releasing
the
preview
version
4,
which
is
targeting
made
the
14th
so
by
end
of
May
the
foldings.
We
should
be
able
to
have
the
public
NuGet
package,
which
is
marked
as
done
at
5
preview,
for
which
has
the
system
agnostic
source
assembly
that
is
going
to
work
with
all
the
current
connect
framework
and
core
versions.
So
for
now,
in
order
to
make
progress
were
running
short
on
time.
C
We
take
this
private,
build
in
that
package
from
the
the
BCR
core
developer
and
incorporate
that
into
open
time
to
just
to
make
progress
without
us
having
to
remove
any
existing
API
Leeman,
the
sidecar
so
user,
you
can
either
use
like
existing
spend
or
you
can
create
activity
in
the
new
API
and
when
it
got
more
like
proved
week,
I
started
to
remove
spam
and
and
like
online
with
this
I
think.
So.
This
would
help
us
to
have
a
better
story
for
tracing
API
and
for
logging
were
working
with
a
donut
PCL
team.
C
The
first
immediate
goal
is
for
all
the
logs
and
made
his
role
event,
source
and
ilogger.
We
want
to
have
the
ability
to
access
the
trace,
ID
and
span
ID,
so
this
is
committed
in
donĆt,
like
preview,
probably
like
five
milestone,
and
the
next
step
is
for
a
logger.
We
do
have
a
requirement
to
make
it
easier,
and
you
know,
like
a
logger,
has
his
own
log
scopes
won't
line
up
better
with
day
with
the
activity
like
scope
here.
So
that's
one
thing
we're
trying
to
address
in
the
log
initiated.
C
C
Telemetry,
matrix
API
is
still
evolving,
so
that
cannot
take
a
bad
on
that
before
it
got
like
most
likely
in
shape,
so
the
decision
from
Donna
and
him
is
for
now
it's
not
in
scope
for
done
at
five
release
and
for
us
in
order
to
make
progress,
we
have
to
shape
the
current
matrix,
HDI,
SDK
and
Internet
six
timeframe.
When
we
have
energy,
the
goal
is
for
Don
at
him
to
implement
the
compatible
matrix,
API
surface
aligned
with
open
telemetry.
So
we
can
start
moon
and
any
questions.
I.
E
I'm
gonna
try
to
look
more
at
these
III,
really
just
look
at
the
PR
that
sigil
just
put
out
yesterday.
I
think
I
I
find
the
idea
very
good,
because
it's
the
wrong
time
so
I
think
makes
it
easier
for
people
instrumental
I.
I
do
have
a
feeling
that
people
that
are
familiar
with
open
telemetry,
we
kind
of
made
lack
they
P
eyes
that
they're
I
use
it.
You
mean
dotnet,
but
the
fact
that
is
something
that's
incorporated
on
the
VCL
makes
very
very
attractive.
E
C
Yeah
I
I
share
the
same
feeling,
and
also
there
there's
some
subtle
differences,
because,
when
donut
designing
the
new
activity,
the
needs
will
balance
between
two
things.
One
is
they
want
to
align
with
open
telemetry
another
one?
Is
they
have
dyslexia,
accurate,
a
customers
who
are
depending
on
the
existing
interface
and
behavior?
They
try
to
do
the
best
balance
there
I
believe
we
we
solved
90%
of
the
problems
and
for
the
rest,
10%
is
not
blocking.
C
For
example,
like
the
ability
to
add
links
after
exactly
got
created,
the
ability
to
have
the
GRP
say,
status
called
other
lessons
we
can
easily
add,
but
we're
not
trying
to
address
them
right
now,
because
we've
also
hit
the
the
preview
for
milestone.
So
those
are
things
I
believe
with
within.
Like
two
weeks,
we
can
cover
90%
of
the
the
spec
and
with
the
next
preview
release
we
can
cover
the
rest
5%
and
for
the
remaining
five
will
need
was
the
the
best
and,
in
general,
I'm
optimistic
that
we
can
have
a
good
answer.
Yeah.
A
Like
we
can
post
like
what
exactly
would
be
like
the
difference
between
what
open
delimit
respect
calls
for
and
what
is
offered
by
activity
like
a
couple
of
the
points
we
already
said
and
there
there
may
be
like
other
subtle
things
as
well,
so
we
can
like
have
an
issue
describing
what
exactly
is
the
difference
between
these
two?
So
people
are
really
aware
of
any
shortcomings.
Yeah.
C
D
C
C
We
explore
this,
and
currently
it's
not
from
the
donor
team's
commitment,
but
in
general
I
think
this
is
doable
freedom.
We
have
a
span
class
like
wrapping
the
activity
or
like
encourage
from
activity.
So
it's
quite
something
in
nature
and
we
figure
out.
This
is
a
problem
that
is
going
to
take
longer
time
than
what
we
can
achieve
in
the
preview
for
milestones,
we
decided
to
part
that
that's
part
of
the
5%
remaining
question.
I
have
yeah.
D
C
A
A
We're
thinking
like
instead
of
calling
point
three
point
four.
We
may
actually
change
it
to
like
alpha
one
two:
three,
because
this
is
like
whatever
we
are
going
to
ship
now
is
going
to
be
like
incompatible
with
whatever
we
shipped
so
far,
because
what
we
ship
has
trace
or
not
span,
but
what
we
are
going
to
ship
will
be
activity
door
start.
So
we
can
next
start
with
alpha
one.
Two,
three
and
map
like
do
a
do:
a
alpha
release
whenever
dotnet
releases
a
preview
version.
A
F
I
think
I
think
it'll
be
good
to
get
community's
perspective
on
this.
We
were
thinking
more
often
instead
of
T
version,
0.3
0.4,
which
do
not
have
tentative
dates
to
dates
associated
with
them.
You
kind
of
wanted
to
distribute
the
work
over
each
month
and
call
it
alpha
2.0
alpha
3.0,
until
we
really
until
we
get
to
a
actual
full
beta
stage,
does
anyone.
C
C
A
C
F
A
It's
via
like
beyond
the
beta
already,
because
every
other
language
like
Python
and
go
or
they
call
themselves
beta.
So
the
first
thing
we
need
to
do
is
like
go
through
all
the
current
issues
and
produce
the
first
ETA
I,
don't
know
whether
we
need
to
take
further
split
it
into
like
point
two
point:
three
point:
four
I
think
the
first
version
we
should
ship,
then,
would
be
the
beta
which
would
contain
all
the
activity,
changes
and
implementation
of
everything.
So
far,
SPECT
yeah.
C
I
think
the
the
scope
when
I
will
count
like
Python
I.
Remember
the
scope
was
to
have
a
tracing
API
and
the
metrics
EDI
implemented,
based
on
a
spec
latest
us
back
and
the
ice
DK
plus
the
exporter
for
eager
zip,
King
and
Prometheus
and
optionally.
We
can
have
the
bridge
for
open
Tracy,
so
these
are
the
beta
school
I
can
I
can
double
check
and
get
the
exact
definition.
Sin
like
putting
a
meet
knows.
A
Okay,
so
is
a
conversion
that
we
will
be
like
the
next
next
thing.
Actually,
you
get
which
we
ship
from
this
repo
will
be
called
a
beta
it,
and
it
should
be
depending
on
dotnet
preview,
and
it
should
implement
all
the
specs
which
other
risky
case
have
implemented
when
they
called
themselves.
Peta
I.
C
Think
yes,
so
our
beta
designation
should
align
with
open
document
definition
and
before
that
beta,
do
we
intend
to
release
another
alpha.
I
think
that's
open
for
discussion
and
if
it's
possible,
I
would
prefer
to
have
a
at
least
one
alpha.
Really
awesome
like
a
new
release.
Beta
beta
has
a
different
semantics.
In
general,
we
don't
try
to
do
like
a
dramatic
API
change
and
we
want
to
consolidate
more
feedback
from
customers.
So
we're
telling
customer
or
non-critical
business
scenario
you
can
use
beta.
This
is
a
different
message
from
alpha.
C
F
C
So
I
think
how
about
this
like
when
we
have
this
like
new
activity
thing
introduced
and
once
done
at
him,
they
have
this
preview
for
released.
So
we
don't
have
this
binary
inside
our
repo.
We
take
a
direct
dependency
on
the
on
the
public.
You
get
feed,
then
we
can
call
it
alpha
release
to
let
people
like
another
out
forest.
C
Let
you
go:
try
the
metrics
tracing
API
and
and
after
that,
if
they
get
enough
signal
that
we
should
move
forward,
we
can
clean
up
the
existing
expense,
then
processors
or
things,
and
then
we
do
like
quick
scan
through
all
the
existing
search
API
surface,
with
open
climate
respect
to
see
how
probably
are
from
big
and
then
move
from
there
towards
the
beta
release.
Okay,
that
make
sense
yeah,
so
the
first
alpha
release
won't
be
having
both
the
pivot.
C
He
spent
surface
at
the
tracing
API
and
we
add
the
documentation
to
explain
to
people
what's
the
direction,
we're
moving
and
and
probably
having
a
rough
version
of
the
like.
The
comparison
like
how
like
like
how
many
things
we
are
covering
and
what
are
the
the
gap
width
of
grandest
and
closing
those
gaps
as
much
as
possible
towards
beta
okay,.
A
E
A
A
Let's
see
like
if
I
have
like
something
suitable
into
like
smaller
chunks,
I'll
post,
an
issue
and
let
market
as
on
design.
So
you
can
pick
it.
Anyone
can
pick
it
basically
yeah
yeah.
It
has
coming
just
from
folks
to
do
matrix
for
having
a
message
and
I
didn't
get
any
response
yet
still
open
things
to
be
done
in
matrix
one.
A
One
of
the
reasons
why
I
want
to
go
really
slow
on
matrix
is
like
AP
itself
is
changing
pending
one
spec
to
be
much
like
it
could
get
much
anytime
today
of
this
week.
So
as
soon
as
we
like
that
spec
is
merged,
we
will
have
a
clearly
breaking
change
in
the
matrix
API
itself.
So
if
someone
is
planning
to
write,
auto
corrector
like
system
counter
or
anything,
it
may
be
better
to
wait
until
this
a
PHA.
A
Otherwise,
it
will
be
like
lot
of
free
reverb
for
the
person
who
is
actually
making
me
pay
change,
yeah
and
like
just
another,
update
on
metric.
Since
you
are
on
that.
Looking
for
views,
that's
the
next
major
thing
in
coming
in
views.
I
already
prototype
at
the
specification,
which
was
posted
in
a
hotel
a
few
months
back.
If
anyone
wants
to
like
play
with
it,
I
can
it's
a
private
branch.
I
can
just
share
the
prior
branch.
It's
it
should
just
work,
but
it's
it's
not
it's
like
a
prototype.
A
The
spec
itself
is
not
there,
so
we're
just
assuming
this
is
going
to
be
the
spec.
It
will
help
like.
If
anyone
is
like
actively
interested
in
following
the
matrix
discussions,
it
may
help
how
to
see
the
property.
We
have
alternate
for
implementation
in
Python
s.
Will
Python
also
did
the
prototype
sorry
delay
tonight
yeah.
So
if
anyone
asked
I
would
share
it,
otherwise,
I'll
just
keep
right
on
that.
C
Thank
you.
Another
thing
I
want
to
get
people's
idea
here
so
in
open
telemetry
like
the
community
like
we
started,
a
new
meeting
series,
which
is
a
the
maintenance
meeting
and
currently
like
like
seems
like
certainly
Mike
they're
leaving,
so
we
don't
have
maintainer
representing
Donette
and
I,
wonder
like
if
they
can
have
like
co-op
Paulo.
C
At
least
I
know
your
approvers
right
now,
but
you
can
represent
the
maintainer
since
the
maintainers
are
not
here
and
and
okay,
so
I
I
come
over
you,
the
the
meeting
series
and
this
special
situation
so
probably
like
I
need
to
communicate
with
Ted
to
see
what's
the
guidance.
Another
thing
is
like
on
the
first
maintainer
meeting,
which
happened
yesterday,
I
heard
from
Google
that
they're
willing
to
donate
60
interns
to
this
project.
So
currently,
as
a
maintainer
for
the
SIP
has
passed.
Project
I
got
eight
interns.
C
One
is
able
to
join
this
week
like
this
Friday.
Another
one
is
going
to
join
next
week
and
there's
smore
who's
joining
like
late
May,
so
on
the
donut
project.
If
we
are
willing
to
take
interns
from
Google
I
want
you
guys
to
think
about
the
following
things.
So,
first,
are
you
willing
to
become
the
mentor
like
Google
requires
a
one-to-one
mentor,
so
you
have
to
dedicate
the
time
for
the
intern
they
might
reach
out
to
you
and
get
her
or
through
email,
and
you
need
to
have
a
specific
topic.
C
You
want
intern
to
be
able
to
achieve,
and
their
time
frame
should
be.
Let
me
check
probably
like
three
to
four
months,
so
this
is
a
like
a
long
period
of
time,
so
you
got
to
think
about
what
was
the
reasonable
project
you
assigned
to
them
and
the
interest
will
walk
in
pairs.
So
if
you
want
to
be
a
mentor
for
intern,
you
got
to
be
a
mentor
for
a
pair
of
interest
to
interns.
They
will
be
pairing
up
and
work
on
the
same
project
that
you
give
to
them.
C
So
these
are
the
things
to
put
into
consideration.
Currently,
I
don't
know
if
we
were
waiting
to
take
interns
and,
if,
like
you
guys,
are
committed,
but
if
you
do
have
the
like
strong
interest,
this
post
data
channel
our
forward
CTO
and
Paolo
the
maintainers
meeting.
So
you
can
join
and
next
time,
yeah.
E
D
One
one
thing
before
I
like
now
that
I
have
a
gap,
I
think
Sergey
is
the
only
person
with
access
to
new
get.
They
did
create
an
org
I
think
we,
this
is
done.
There's
an
org
and
the
org
also
owns
the
package,
but
then
someone
has
to
get
access
from
Sergey
as
well
like
whoever's,
gonna,
be
the
new
maintainer
assume
to
be
able
to
call
the
packages.
Look
at
the
Lord
yeah.
A
A
C
A
E
I
had
some
conversation
last
week
with
Sergei
about
the
instrumentation
engine
and
the
intercept
exchange.
I
extension
I
find
a
very
good
approach.
I
have
a
concern
about
the
timelines
for
that
to
really
get
ready,
but
I'm
gonna
try
to
have
some
conversation
with
Sergei
to
get
some
of
that
cold
out.
So
we
can
kind
of
experiment
and
do
some
tests
you
kind
of
well
they
that
that
approach.