►
From YouTube: Languages & Runtime: .NET Community Standup - Sept. 12th 2019 - .NET Conf, .NET 5, OpenFSharp
Description
Join members from the .NET teams for our community standup covering great community contributions for Framework, .NET Core, Languages, CLI, MSBuild, and more.
A
B
C
B
It
does
yeah,
so
we
shipped
three
oh
preview,
9,
not
this
Tuesday
but
took
two
Tuesday's
ago
and
kind
of
the
the
there's.
A
couple
big
messages
here.
One
is
that
we're
really
are
almost
on
this
release.
So
there's
no
new
features
listed
here
intentionally.
I
didn't
even
actually
ask
anyone
if
there
were
any,
as
was
anything
notable
to
talk
about
really
trying
to
push
this
message
that
we're
done
and
so
Donna
choreo
is
actually
gonna,
be
launching
at
Donna
Kampf,
which
is
J's
the
twelfth.
B
So
it's
actually
it's
11
days
from
now,
so
we
are
taking
the
last
few
bug.
Fixes
I
mean
it's.
It's
now
really
hard
to
get
fixes
into
the
product,
so
yeah
we're
pretty
we're
pretty
much
done
and
we
will
be
moving
on
to
three
one
which
we'll
talk
more
about
both
of
those
at
Netcom.
So
it's
awesome,
yeah.
A
B
A
B
We
should
talk
about
the
visual
studio,
support
thing
because
we've
gotten
to
feedback
both
from
the
community
and
from
also
our
own
team
that
we
don't
do
a
good
job
of
this.
So
that's
why
I
wrote
three
paragraphs
on
or
really
two
on
this,
which
is
so
Visual
Studio.
Has
this
practice
of
putting
it
updates
relatively
quickly?
B
So
that's
you
know
the
16-0
61-62
kind
of
thing
and
that's
great,
because
it
just
means
that
the
visual
studio
team
and
is
really
trying
to
add
new
value
to
the
product
and
you
know,
fix
any
bugs
and
reliability
issues
there.
So
that's
like
this
constant
improvement
so
thing,
but
as
the
dotnet
team,
it's
really
hard
for
us
to
like
support
all
of
those
versions.
So
we
have
picked
16.3
as
yeah
63
as
the
one
aligned
on
a
core
three
Oh
with.
B
So
if
you
want
to
use
tonic
or
three
Oh
both
now
and
in
the
future,
you
need
to
be
at
least
on
sixteen
three.
Some
people
have
said
to
us
like,
oh
I'm,
using
16
or
16
1
and
everything's.
Awesome
I
don't
see
any
problems,
well,
I'm
glad
that
that's
the
case,
but
there
definitely
are
fixes
and
new
small
features
in
Visual
Studio
16
3
that
are
required
to
make
the
naquadria
work
well
in
all
scenarios
and
do
anything
to.
D
Add
on
that
one
yeah
I
mean
like
one
of
the
fundamental
units
of
reference
framework
references,
that's
a
concept
that
is
not
understood
by
the
Visual
Studio
IDE
before
bs
16.3
so
like
you
can
still
have
stuff
populate,
but
like
that's
an
example
of
the
UI
showing
you
something
that
isn't
like
quite
right.
Yeah.
B
D
B
C
B
A
A
A
B
A
B
D
D
A
Published
have
done
it,
but
me
saying
it,
but
I
think
they
added
like
and
three
things
with
something.
So
that's
fine,
you
know
be
great
if
someone
help
me
up
with
that,
so
I
had
a
gift
to
post
yeah.
You
say
that,
but
last
last
week
a
funny
story.
By
the
way.
Last
week,
I
submitted
a
fixed
for
us
mode
and
I'm
thinking
to
myself
or.
B
We
have
the
regular,
which
is
there's
much
less
scrutiny
about
changes
that
go
in,
and
then
we
have
something
called
tell
mode,
which
basically
means
there's
increased
scrutiny.
All
you
have
to
do
is
send
an
email
to
a
group
to
say,
you're
checking
something,
and
you
don't
have
to
wait
on
the
answer.
It's
more
just
about
creating
a
visibility
on
the
type
of
changes
that
are
going
in
and
the
last
one
is
ask
where
you
have
to
send
the
same
email
and
you
have
to
wait
for
the
answer
and
then
obviously
the
questions
apply
higher.
C
A
B
B
C
C
C
C
B
B
A
So
I'm
going
to
me
to
my
manager
I'm
like
well,
where
should
I
submit
the
fetch
to
and
yeah
he's
like
released
to
you
I'm
like?
Can
you
show
me?
We
released
the
OS
and
he's
like
it's
right
hold
on
turns
of
it
lead
to
the
branch
accidentally.
So
you
have
all
these
branches
and
able
branch
protection
on
github.
This
avoids
I.
B
Do
want
to
talk
about
that
very
last
question
which
is
so
Donna
corte-real
looks
great.
Why
won't
I
be
able
to
use
and
I'm,
currently
supported
Windows
versions
like
Windows
Server,
2008
r2,
so
this
is
actually
my
fault,
so
one
of
the
things
I
do
is
look
at
our
support,
matrix
and
so
I
did
this
exercise
like
six
months
ago
or
maybe
even
longer,
and
the
thing
that
I
want
to
prevent
is
that
we
support
a
version,
an
operating
system
version,
whether
it
be
Windows
or
Linux
or
Mac.
B
That
is
going
to
go
to
support
relatively
soon
into
the
support
lifetime
of
a
particular
product,
because
it's
not
really
helping
the
customer
in
any
particular
way.
So
Windows,
Server,
2008
r2,
goes
out
of
support,
as
I
recall
in
like
January
or
February
or
March
of
next
year.
So
my
take
on
that
is
there's
only
this
like
small
window,
where
you
would
have
benefit
and
then
you
would
not
be
supported
by
our
team,
so
we're
not
doing
anything
to
block
there,
the
donek
or
three
from
working
in
any
place.
B
A
C
A
In
10
days,
there's
not
a
comp
is
just
a
virtual
conference,
meaning
it's
streamed
and
you
can
watch
it
life
of
this
page
and
also
on
I.
Think
a
channel
mine
as
well,
which
is
actually
hosting
that
so
there's
no
reason
not
to
attend.
It's
all.
You
know
live
and
free,
so
we
just
click
on
a
link
and
let
it
run
while
you're
typing
your
stuff.
A
There's
awesome
sessions.
We
have
a
lot
of
content.
Actually
it's
like
three
days
fully
packed.
You
know
we
will
have
a
keynote.
Obviously
we
talked
about
net.
You
know
past
present
future
anything,
and
then
we
also
have
you
know.
What's
even
c-sharp,
you
know
desktop
apps
DevOps
laser.
You
know
everything
you've
heard
about
is
covered
by
both
people
from
Microsoft,
as
well
as
community
members
who
understand
the
product
really
well.
So
it's
a
great
lineup
of
speakers.
In
fact,
we
have
a
lot
of
them.
C
B
C
B
C
C
B
B
C
B
B
B
C
C
B
C
C
C
These
ones
are
so
Visual
Studio,
2019,
community
and
enterprise
again
with
the
Installer
I
can
do
lots
of
side
by
side,
installs
cool,
so
these
squiggles,
we
know
about
green
squiggles,
being
warnings.
If
some
of
you
might
have
noticed,
three
dots
are
our
suggestions
so,
and
you
can
also
have
errors,
so
there's
always
like
different
levels
of
severity.
C
Is
that
different
code
fixes
and
right
back
drains
appear
in
your
document
and
you've
been
able
to
configure
quite
a
few
of
these
in
an
editor
config
did
I
J
go
through
what
editor
config
Asst,
probably
okay,
so
editor
config
is
how
you
can
basically
record
all
of
the
code
styles
and
options
and
severity
zat,
which
thing
appear
all
in
one
document.
It's
actually
a
file
that
lives
with
your
code.
It
can
be
managed
by
source
control.
So
here
it's
in
my
solution,
Explorer,
it's
pretty
great,
just
wanted
to
give
ya
one.
C
C
C
C
C
C
This
is
a
little
bit
buggy
and
jumps
around,
but
you
we
want
to
actually
be
able
to
configure
code
severity
x'
and
what
they're
appearing
at
in
the
editor
so
I'll
go
ahead
and
change
this
warning
to
an
error
and
when
it
reruns
so
now,
I'm
getting
a
red,
squiggle
and
I
was
able
to
do
that
right
in
my
editor
and
now,
if
I
go
down
to
my
editor,
can
fake
Muslim
I
can
actually
see
it
was
CS
0
to
19
and
they
set
the
severity
to
an
air.
So
normally
I
would
do
like.
C
Oh
I,
don't
know
you
can
change
it
back
to
a
warning
here
and
save
it,
and
then
we
code
analysis.
It
runs
that
so
now,
I,
basically
just
don't-
have
to
switch
back
and
forth
to
my
editor
config
and
when
I
change
the
setting.
It's
you
know
saved
in
that
file.
That
will
be
changed
across
all
of
memory
bugs,
so
you
can
do
that
with
multiple
ones.
Something
we're
also
so
excited
about.
C
This
is
that
it's
much
easier
to
change
the
severity
level
of
like
built-in
tools,
so
all
of
our
built-in
analyzers
and
warnings
we
didn't
have
a
way
to
configure
an
editor
config
because
they
just
shipped
ads.
This
is
the
default
project,
and
now
a
bunch
more
of
those
will
be
exposed
eventually,
when
the
ships,
which
is
not
now
so.
A
C
B
C
C
It
is
currently
an
extension,
so
it
our
config
language
service
extension,
which
is
really
cool,
I,
definitely
recommend
downloading.
That
I,
don't
think
I
have
it
on
this
install
but
yeah.
You
can
get
basic
help
with
setting
editor
config
things,
so
it
can
show
like
intellisense
completion.
That's
though,
that's
not
it,
but
this.
A
B
A
C
D
C
B
Trying
to
put
a
new
cross-platform
product
on
that
core
and
it
looks
like
we
will
constantly
be
having
to
jump
to
new
versions
of
Linux
or
risk
customer
versions
failing
or
falling
one
of
the
yeah.
So
we,
you
know,
the
Donna
Corp
project
is
now
like
five
years
in,
and
you
know
we
kind
of
started
it
like
publicly
five
years
in
and
we
started
it
a
little
bit
before
that.
B
So,
like
our
big
thing,
which
then
this
is
very
similar
to
I
was
talking
about
before
is
like
so,
for
example,
like
Debian
8
went
out
of
like
mainstream
support,
I
think
I
live
earlier
this
year
or
last
year.
I
can't
keep
all
these
dates
in
my
head,
and
so
that's
the
point
at
which
we
dropped
support
for
Debian
8
with
Donna
core,
because
if
a
customer
comes
to
us
and
says
like
oh,
my
app
is
is
busted.
Can
you
help
me
and
then
we
determine
that
it's
a
problem
in
Debian
8?
B
Then
we've
got
no.
That's
it's
like
no
way
to
affect
that.
Get
that
app
working
again
with
Debian
8,
because
there
would
be
a
fix
that
would
be
needed
and
Debian
8.
So
as
a
result,
we
kind
of
stay
closer
to
the
front
such
that
you
know
we
get
people
on
Debbie
and
9
and
Debian
10,
for
example,
so
that
they're
much
more
likely
to
have
a
good
experience.
So
that's
what
that's
all
about
you
want
to
take
the
one
about
done.
This
yeah.
A
And
what
I
will
take
right
after
this,
because
I
think
it's
a
nice
segue
to
the
next
topic?
There
was
one
question
about
the
designer
for
good
films
in
WPF,
so
we
will
not
ship
the
designer
as
our
TM
quality
of
16.3.
We
will
have
a
preview
I
think
all
they
are
still
working
through
the
motions
of
where
to
release
it
under
I
think
they
decided
not
to
use
the
vs
gallery
for
that,
but
they
were
probably
pronounce.
The
name
is
isomer
that
you
can
download.
A
B
D
And
the
wind
forms
one
is
even
bigger
than
1,
because
the
WPF
one
like
I
mean
had
a
huge
amount
of
work
for
the
scene.
We
did
that,
but
their
initial
architecture
was
like
easier
to
kind
of
adjust
and
make
it
work
out
of
proc.
They
already
had,
like
a
lot
of
you,
know,
resiliency
around
hey,
we're
still
searching
for
things
and
we're
trying
to
be
responsive
about
that.
What.
A
B
B
It
improv
with
yeah
also
we've
done
it
framework
everything
in
the
same
process.
The
funny
thing
about
that
is,
it
would
end
up
needing
a
bunch
of
the
same
work
because,
like
pointers,
don't
cross
like
VM
boundaries,
they're
completely,
two
different
universes,
so
you'd
end
up
having
a
bunch
of
the
same
work
to
do
and
the
the
architecture
would
just
be
fundamentally
worse
and
have
different
and
have
worse
runtime
characteristics.
So
it
would
just
be
all
bad.
So
we
didn't
do.
B
A
C
B
C
C
Such
a
good
point,
so
we
looked
a
lot
of
that
about
how
to
document
code
style
and
we
picked
editor
config
because
it
actually
existed
as
a
code
style,
recording
I,
guess
paired
him
before
like
outside
of
dotnet,
and
we
went
out
and
adopted
it.
So
editor
config
existed
for
multiple
other
languages,
so,
instead
of
creating
our
own
stack,
we
went
in
yeah
adopted
a
standard
that
we
already
recognized
it
that
way.
I
guess.
B
C
And
I
think
different
part
of
the
question
was:
are
all
of
these
severity
and
warnings
caught
by
msbuild
or
like
from
when
running
from
the
command
line
and
I'm
trying
to
think
where
we
are
with
that
we're
definitely
working
towards
it.
I
think
some
stuff.
We
can
catch
at
the
command
line
with
dotnet
format,
but
it
doesn't
cover
all
editor.
Config
rules
and
I
think
anything
that
you
promote
to
an
error
might
be
able
to
break
your
builds
or
still
working
on
it.
I
need
to
look
it
up.
C
B
A
The
older
ones,
so
the
one
question
was
when
you
start
in
center
two
point:
one
finalized,
maybe
Marshall
difference
between
the
net
current
standard.
Would
there
be
a
check
that
the
new
certain
areas
are
covered
in
the
center
two
point
one.
So
the
short
answer
is
not
in
send
a
two
point:
one
ships.
At
the
same
time,
it's
not
a
core
three
of
art
hams,
which
is
basically
the
same
thing
we
have
done
over
the
last
years.
So.
B
A
Days
or
whatever
yep
so
the
the
thing
that
is
always
confusing
to
people
is,
they
expect
them
to
be
exactly
the
same.
We
said
originally
don't
and
Senate
will
always
be
trailing
behind,
meaning
we
require
an
API
to
ship
stable
before
we
put
it
under
standard,
which
has
been
kind
of
the
case
for
earlier
versions
of
Don
and
standard
by
the
fact
that
we
really
didn't
really
add
new
API.
As
to
it.
We
just
standardized
already
existing.
A
What
a
trimeric
api's
with
two
one
is:
what's
the
first
Felisa
we
actually
have
net
new
API
is
that
didn't
exist
yet,
and
so
the
we
have
kind
of
cheated
into
one
in
the
sense
that
we
actually
did
add
brand
new
API
pseudonym,
where,
like
range
and
index,
and
some
of
the
other
core
of
ECL
concepts,
they
be
also
simultaneously
exposed
to
known
as
Dana.
But
we
haven't
done
this
across
the
board
for
simple
reasons,
because
every
time
you
had
an
API
other
people
have
to
do
actual
work
like
salmon.
A
A
B
I
think
the
the
super
short
version
is,
we
have
had
boxes
sorry
diagrams
very
similar
to
this
in
the
past,
and
they
were
all
based
on
dotnet
standard.
Providing
the
commonality,
as
you
know,
many
listeners
will
know.
Dinah
standard
is
just
a
spec,
so
it's
like
virtual
sharing,
if
you
will,
and
so
in
this
new
model,
we're
actually
using
the
same
base
class
library
implementations
across
all
the
different.
You
know
runtimes
and
frameworks
and
everything,
and
so
the
two
biggest
advantages
of
that
is
it's
the
same
code
everywhere.
So
it
will
behave
the
same
way.
B
You
know
that's
awesome
and
then
the
other
thing
is.
It
gives
our
team
less
frameworks
to
work
on.
So
it
means
we
get
to
focus
on
making
this
one
implementation
better
in
all
places
and
it
out
sorry
and
also
obviously,
the
community
actually
ends
in
some
ways.
Actually
that
that's
another
thing,
which
is
not
only
it's
a
community,
get
to
work
on
one
thing,
but
it's
also
way
less
confusing
I'm
sure,
there's
lots
of
folks
who
are
like.
B
A
This
is
basically
the
the
the
slide
that
hunter
showed
where,
on
the
left
hand,
side
you
have
the
current
world
2016
and
on
the
right
hand,
side
is
division.
4.5,
like
I,
think
I
sort
of
just
said.
We
have
too
many
boxes
right,
so
another
frame
that
I
think,
which
you
came
up
with
earlier,
was
that
dot
next,
a
node
is
basically
unifying
shape,
versus
the
hold
on
and
fighting
is
unifying
implementation
right,
we're
sort
of
having
three
different
kind
of
product
categories
built
in
different
repos,
with
different
build
tools
and
different
crazy.
A
We
just
have
one
now
you
see
on
the
right-hand
side
at
the
very
bottom,
this
small
written
thing,
dotnet,
standard
I,
think
we're
still
trying
to
figure
out
what
Donna
standard
means
in
this
new
world
when
we
actually
unified
implementation.
But
the
idea
is
that
you
know
in
the
new
world
if
we
had
an
API,
it's
instantaneously
available
everywhere.
It's
no
longer
this
editorial
aspect
of
okay.
Now
we
add
an
API
to
core,
then
afterwards
we
add
it
to
the
standard,
and
then,
after
that
we
add
it
to
other
platforms.
It's
it's.
A
C
B
A
One
thing
that
was
brought
up
is
that
project
files,
when
you
edit
them
and
you
screw
them
up
like
we,
don't
have
a
hard
time
opening
projects
afterwards.
I
mean
the
general
problem.
I
think
we've
msbuild
is
that
its
effect
I
mean
even
though
it
looks
declarative.
It
really
isn't.
It's
basically
a
script
that
runs
and
executes.
So
there's
ordering
and
other
concerns
to
keep
in
mind
versus
editor.
Config
is
basically
declarative,
but
yes,
there's
also
ordering
and
nesting
and
replacement,
but
it's
data-driven
rather
than
code
executes
and
does
something.
B
A
B
A
C
A
There's
a
few
libraries
that
we
have
built
specifically
targeting
that
Center.
Oh
there's,
a
good
chunk
of
things
that
we
have
in
the
Microsoft
extensions
area
that
will
be,
but,
like
asp
net
core
in
the
EF
core,
they
decided
to
break
with
that,
because
the
whole
point
of
both
of
these
projects
is
to
to
innovate
at
a
faster
pace
than
we
can
donate
framework,
and
we
basically
decided
on
a
framework
at
this
point
is
feature
complete.
A
We
will
make
bug
fixes
to
it,
and
some
of
the
back
fixes
might
be
in
feature
territory
like
adding
a
new
crypto
standard
or
whatever.
But
the
spirit
is
we
only
do
things
that
keep
your
existing
button
framework
abs
working,
we
don't
add,
new
features
to
it,
and
so
that
means
for
a
year.
Of
course,
they
want
to
take
advantage
of
new
things.
We
have
they
kind
of
have
to
be
on
things
that
are
basically
on
that
Center
one
which
excludes
don't
ever
yeah.
B
I
think
there's
another
answer
which
is
complimentary,
which
is
there's
this
tension
of.
So
if
you
were
the
EF
core
team,
so
everything
you
said
is
true.
But
if
you
were
the
EF
core
team,
then
you
have
the
choice
of
I
can
just
use
kind
of
lowest
common
denominator,
features
that
happen
to
exist
in
Donna
core
and
on
a
framework.
So
I
guess
that's
basically
done
its
standard.
B
That's
a
choice
we
made.
It's
like
your
explanation
is
much
more
backward-looking.
My
explanation
is
much
more
forward.
Looking
like
it's,
we
are,
we
are
entirely
focused
as
a
team
on
like
high
levels
of
innovation
across
the
entire
stack,
and
the
only
way
to
do
that
is
by
building
and
using
features
that
are
donna
core
into.
Third
literally,
is
another
way,
because
we
tried
the
whooping
thing,
building
kind
of
like
variations
of
features
like
span
out-of-band,
and
we
have
discovered
that
this
just
does
not
work
in
practice.
Yeah.
A
A
The
problem
is
that,
while
we
can
do
that,
the
fidelity
usually
isn't
there,
and
so
we
have
to
do
kind
of
Frank
and
designs
to
accommodate
this
whole
thing
and
don't
always
work
which
doesn't
always
work
and
you
know,
tends
to
create
a
mess
or
a
good
example
is
you
have
to
do
extension
methods
rather
than
adding
the
instance
methods
in
the
right
place,
and
that
prevents
us
from
making
things
virtual
and
blah
blah
blah.
So
there's
a
whole
bunch
of
trade-offs,
and
we
don't
want
that.
A
Somebody
says
well.
My
framework
is
dead.
Long
live
donut
framework.
Well,
so
this
is
a
phrase
that
I
want
to
comment
on
briefly,
because
people
keep
telling
me
this
all
the
time
that
we
should
just
stay
like
my
family
is
dead,
but
Microsoft
said
it
says
dead.
It
really
means
you
should
golf
go
off
of
it
because
we
basically
start
patching
it.
That
is
not
the
case.
Without
net
frame,
I
got
a
very
cool
ship
of
the
US,
as
it
had
always
been.
It
will
be
patched
as
part
of
the
operating
system.
A
D
It's
the
super
shirt
and
I
would
encourage
pretty
much
anyone
here
to
open
up
their
search
engine
and
search
for
the
phrase.
Dotnet
core
is
the
future
of
net.
We
have
our
glorious
director,
Scott
wrote
a
fancy
blog
post
that
basically
says
don''t
framework
is,
is
stable,
secure,
it's
getting
all
this
sort
of
stuff,
but
it's
not
where
you
want
to
build
new
stuff,
and
it's
not
where
new
things
are
going.
All
future
best
parts
are
going
into
da
net
core
and
in
the
future,
gone
i5,
which
is
what
donna
core
is
evolving
into
right.
C
D
B
Like
to
address
the
one
question
someone
or
brave
Cobra
too
is
asking:
are
we
getting
Linux
GUI
supported
of
the
Vox
at
some
point
in
time?
I
actually
had
a
similar
question
from
someone
else
recently.
My
question
back
to
you
is
if
we
were
like
there's
a
few
different
toolkits
that
are
in
available
on
Linux
gtk
and.
B
B
Great
question
there's
been
some
amount
of
discussion
about
that
for
the
foreseeable
future.
We
intend
to
use
the
mono
at
runtime
and
really,
the
reason
is
is
two
part
is
or
it's
really
one
thing,
which
is
the
monorun
time
has
been
optimized
for
that
scenario,
and
you
know
we
kind
of
optimized
course
alar
for
something
else,
mostly
cloud
oriented
workloads,
and
so
it
would
be
a
significant
project
to
go.
B
Make
core
CLR
work
as
well
as
mono
like
fitting
into
small
small
devices
like
you
know,
watch
OS,
for
example,
because
they've
done
a
bunch
of
work
for
that.
So
that's
that's
not
on
the
road
map
currently
definitely
could
happen
in
the
future,
but
it's
not
something
we're
taking
on
right
now:
yeah
QT!
That's.
A
Right
yeah
brave
Cobra,
says
he
wants
both
I
mean
like
that.
That's
the
general
problem
that
we
have
of
investment
they'd
be
like
it
would
be
nice
if
we
could
do
everything
at
the
same
time
and
Microsoft
sometimes
looks
like
this
ginormous
company,
but
like
each
team,
has
only
so
much
funding
right.
So
we
have
to
pick
our
battles
in
order
to
also
be
able
to
productize
stuff.
You
can't
just
chase
the
latest
things
all
the
time
and
then
get
nothing
finished
well
and
once.
B
A
If
somebody
says,
wouldn't
laser
an
electron,
be
a
good
option
for
that
for
crosstalk
from
UI.
There's
certainly
are
great
options
for
that.
I
think
that
the
biggest
difference
between
blazer,
slash,
electron
or
anything
like
that
is
it's
crazy.
Coming
from
a
bap
mentality
about
HTML
and
JavaScript
and
I
think
the
question
is
that
that
is
certainly
a
way
of
doing
it,
and
it's
probably
appealing
for
basically
everybody
who
does
web.
A
If
you
build,
let's
say
small
devices
like
you
know,
Android
iOS
watch
I
mean
even
though
they're
no
longer
that
small,
that's
the
trans
out,
but
there
is
a
slightly
different
approach
for
many
of
these
things,
and
the
question
really
is
is,
like
you
know,
maybe
we
have
only
one
option
which
is
blazer
based
or
electron
based.
Maybe
we
have
multiple
options
and
it
goes
kind
of
back
to
what
rich
said
like
what
customer
bases
do
we
have,
and
even
today
I
mean
you
see
there.
A
A
A
A
D
C
D
A
B
B
And
I
said:
well:
we're
gonna
talk
about
it
anyway.
So
I
think
this
is
a
great
question.
So
I
thought
it
might
be
useful
to
talk
about
how
our
team
approaches
remote
workers
from
both
policy
and
practical
standpoint,
so
I
think
the
super
short
version
is
this,
which
is
policy
wise,
we're
definitely
open
to
remote
workers
and
we
actually
have
a
bunch
of
them.
B
So
that's
the
practical
part
but
they're
pretty
much
almost
all
developers
so
because
we
have
developers
and
program
managers,
so
I
would
say,
like
90,
plus
percent
of
the
remote
workers
on
the.net
team
are
developers,
the
only
place
where
that's
different
is
Scott
Hanselman
team,
all
remote
workers
I
think
all
and
you
know
it's
because
they,
the
people
that
they're
working
they're
the
community
team.
So
the
people
are
working
without
the
community,
so
it
doesn't
really
matter
where
they
are,
whereas
with
us
we're
all
on.
B
What's
the
opposite
of
remote
workers,
on-site
yeah,
I,
don't
know
yeah
on
so
we're
all
on
site
workers,
because
the
people
that
we
work
with
daily
are
primarily
the
developers
who
are-
and
you
know,
building
a
team
with
us,
and
so
we
have
not
cracked
the
nut
of
figuring
out
a
good
model
for
remote
PMS
that
work
with
developers
that
are
on
site
I.
Think.
A
That
it
kind
of
depends
on
engineering
too.
You
might
I
think
github
is
more
distributed
in
general,
which
no
means
on
the
github
side.
It
doesn't
really
matter
right
versus
on
our
side.
The
large
amount
of
workforces
in
Redmond
and
the
PM
wants
to
be
close
to
where
the
workforce
is
right.
If
your
workforce
is
distributed,
then
PM
can
also
be
distributed
by
the
way,
but
I
think
that
even
SPM,
it's
our
job,
is
mostly
influencing
people
and
building.
A
You
know
getting
people
basically
ratted
up
behind
an
idea
and
driving
that
idea,
and
that
is
much
easier
if
you
can
grab
people
in
the
hallway-
and
you
know,
hunt
them
down
and
actually
grab
them
for
an
hour
rather
than
doing
that
over
distributed
time
zones.
That
gets
really
really
really
good
you're
right,
but
we
do.
B
B
So
this
team
has
been
entirely
in
Redmond
for
the
longest
time,
but
we
actually
now
have
opportunities
in
Vancouver,
British,
Columbia
and
also
in
Prague
in
the
Czech
Republic
that
we
actually
are
building
those
are
actually
not
remote
teams,
they're,
actually,
other
on-site
teams
with
you
know
some,
like
you
know,
20
or
30
people,
each
I
actually
don't
know
the
numbers,
but
so
that's
actually
exciting.
Yeah.
C
A
A
A
Was
one
that
was
a
on
our
blog
post,
so
the
short
story
is
like:
if
you,
if
you
check
and
change
it
since
it's
a
repo,
sometimes
code
has
to
flow
across
various
repos
before
we
can
actually
get
a
build
out
of
the
SDK
or
the
runtime.
That
actually
has
those
changes
in
a
consumable
fashion
and,
of
course,
these
kind
of
you
know
code
flow
problems
also
cause,
then
the
build
to
take
sometimes
24
hours.
A
There
actually
varies,
you
know
trigger
through
various
parts
of
the
system,
and
that
makes
it
really
hard
for
us
to
both
share
sources
as
well
as
just
confusions
where
issues
should
be
filed,
and
you
know
we
have
sometimes
PRS
that
have
to
be
staged
and
like
coordinated
across
multiple
repos,
and
this
now
we
had
the
pointer
we
talked
about.
You
know
days
per
build
rather
than
builds
per
day,
which
is
kind
of
unfortunate,
and
so
this
whole
adversity,
embarrassing.
A
It
is,
but
also
our
product
is
not
small
right,
I
think
like
just
course,
lr+
core
effects,
I
think
it's
like
what
4
million
lines
of
code
or
something
so,
of
course,
that's
not
easy
to
coordinate
so
part
of
this.
This
challenge
here
that
Stephen
opened,
is
you
want
to
consolidate
or
repos
into
into
more
chunky
things
don't
want
to
mesh
everything
in
one
giant
repo,
because
that's
also
not
workable,
but
we
want
to
definitely
minimize
the
code
flow
and
the
time
delay
and
it
will
be
challenging.
A
So,
of
course,
we
need
to
think
about
what
happens
to
the
repos
that
we
already
have
what
happens
with
community
contributions
in
the
time
maybe
maybe
migrate.
How
do
we
make
sure
that
the
team
doesn't
go
offline
for
like
a
month,
that
sort
of
thing
and
so
I
think
that's
kind
of
a
part
of
it
will
keep
us
busy,
probably
after
three
one
ships
until
we
know
500
opens
you.
B
A
A
I'm
gonna
lipstick,
like
noticed
the
mana
repository,
but
as
in
although
the
mono
repo
is
it
mono,
repo,
ironically
ironically
I
guess
but
I
think
yeah
some
it's
generally
supportive,
but
everybody
has
questions
about
what
I
want.
What
happens
to
history
was
I
remain
being
listed
as
a
contributor
or
abysus
master
history.
I
assume
we're
not
doing
no
we're
not
we're.
Basically,.
A
To
clean
up
things,
removing
binaries
that
were
checked
in
some
of
the
infrastructure
PRS,
what
we
generally
maintain
history,
of
course
charge
will
change,
but
we
think
that's
not
a
problem.
So
there's
a
whole
bunch
of
discussion
here
like
if
you
care
about
repo
organ.
Is
it
organization?
That's
a
that's
a
good
one.
A
Yeah,
so
effort
is
asking
you
able
to
maintain
the
revision
history
so
basically
the
the
source
stories.
It
will
smash
all
history,
all
graphs
into
one
repo
and
then
probably
just
rebase,
some
stuff
out.
So
some
shots
will
change.
But
if
you
do
a
git
log,
like
your
commit
is
still
there.
You
know
the
description
is
still
there.
We
don't.
You
know
you're
still
listed
with
my
address.
We
are
still
a
contributor.
So
all
the
all,
the
claim
you
have
a
lines
of
code
written
will
remain
yeah.
B
C
B
A
So
there's
an
interesting
thing:
I
think
I
shared
this
on
Twitter.
If
you
want
to
search
for
it
like
there's
a
github
repo
where
they
use
shell
clones
and
apparently
github
on
get
up,
engineer,
responded
and
explain
what
cello
Clones
does
to
their
servers.
So
shallow
clones
basically
means
you
clone
the
repo
baby,
only
tech
that
many
commits
you
don't
take.
A
The
whole
history
review,
which
is
for
CI,
builds
naively
thinking
beneficial,
but
in
many
cases
it
can
actually
cause
more
load
on
the
server,
because
the
the
protocol
that
you
use
to
actually
retrieve
those
changes
not
as
optimized,
is
just
doing
a
bug
read,
and
so
basically,
you
still
figure
out
how
we
make
sure
that
git
clone
is
fast
on
local
machines.
The
probably
doesn't
matter
that
much
because
most
people
enlist
once
and
then
that's
it,
but
on
CI
machines.
A
A
Well,
I
mean
git
has
pros
and
cons.
Kit
is
not
optimized
for
large
content,
but
you
know:
Windows
has
done
some
work
with
the
good
virtual
file
system
that
they
contributed
back
so
I
think
I,
don't
know,
I,
don't
see
it
being
the
big
problem
here,
I
think
most
of
us
what's
the
problem.
So
we
just
how
do
we
maintain
operational
doing
the
migrations
yeah.
A
C
A
A
D
D
Yeah
open,
F,
sharp
as
opposed
to
closed
F,
sharp,
so
open,
F
sharp
is.
You
can
probably
tell
by
the
website
they're
open,
F
sharp
org.
It
is
a
wonderful
conference
coming
up
this
month
right
after
Donna
clock
I
will
be
attending
speak,
as
will
many
other
wonderful,
F,
sharp
developers.
This
is
a
community
conference.
That's
been
going
on
for
three
years
now
and
it's
based
out
of
San
Francisco.
D
It's
a
really
really
great
place.
There's
always
there's
like
an
easy
track
and
a
more
advanced
track,
and
this
year
we're
covering
topics
from
getting
started
with
F
sharp
to
principles
in
quantum
computing,
so
like
that's
by
and
plenty
of
data
science
stuff,
there's
going
to
be
a
really
cool
talk
on
using
F
sharp
to
talk
with,
like
the
intersection
of
someone
who
like
is
really
interested
in
operating
or
like
ham
radios
and
using
F
sharp
as
a
part
of
that.
So
like
there's,
there's
a
lot
of
really
interesting
stuff
there.