►
Description
Scott Hunter talks about Project K and the start of .NET Core. He gives an update on what's on the way in .NET 5, and shows some brand new demos.
Community Links: https://www.theurlist.com/aspnet-standup-2020-09-29
Featuring: Scott Hunter (@coolcsh)
#ASPNET #DotNet5
A
A
B
All
right
we're
live,
welcome,
scott.
B
B
So
this
is
a
great
time
to
be
talking.
A
B
Really
really
exciting
wow.
Well,
I
can't
wait
to
dig
into
all
the
stuff
you're
gonna
be
showing
us.
Let
me
start
off,
as
we
always
do
with
community
links.
So
let
me
jump
right
into
those,
as
always,
there's
a
ton
of
stuff
going
on
in
the
community.
I
had
some
links
already
together
and
then
I
asked
on
twitter
and
just
got
flooded
with
them.
B
I
tried
to
some
some
great
ones
that
are
blazer
specific
we're
moving
to
next
week,
because
next
week
we're
starting
our
first,
our
monthly
blazer
show
so
asp.net.
We
have
regular
shows
on
blazer
and
we
figured
why
not
just
make
it
a
monthly
thing.
So
sofia
is
going
to
be
co-hosting
that
with
me
and
we'll
be
starting
that
next
next
tuesday,
but
let's
dig
in
so
as
always.
B
These
links
are
shared
out
in
the
chat
they're
shared
out
in
the
description
I'll,
throw
them
in
the
in
the
chat
right
now
and
lots
of
good
stuff.
We'll
dig
right
in
here.
First
of
all,
patrick
has
this
end-to-pen
tool,
and
I
like
this-
he
regularly
does
this
he's
been
doing
this
for
years
and
years
where
he
goes
through
and
compares
apis
between
versions.
So
here's
just
looking
at
in.net
5
and
it's
kind
of
this-
is
a
deep
post.
You
can
see
from
the
scroll
bar.
He
goes
into.
B
You
know
real,
deep
detail.
Some
things
that
kind
of
jump
out
at
me
are
there's.
Definitely
like.
There's
system
system,
net
connections,
api
which
is
interesting,
there's
also,
of
course,
the
system
text,
json
serializer
a
lot
of
stuff
going
on
with
that
and
a
lot
of
important
things
for
perf,
especially
of
course
so
system,
net,
http,
json
extension
methods,
and
he
goes
into
a
lot
of
depth
on
all
these.
B
So
this
is
this:
is
a
meaty
post
and
a
lot
of
information
here,
but
definitely
worth
taking
some
time
to
to
look
at
and
then
at
the
end,
patrick,
you
know
he
links
over
and
and
he's
got
this
end
to
pen
tool
that
helps
you
do
this.
So
this
is
just
a
ton
information
here,
it's
really
great.
B
This
is
neat.
People
have
been
doing
some
neat
things
with
source
generators,
so
you
know
the
team
has
been
saying:
hey
some
things
that
we
expected
a
lot
of
interest
in
and
and
some
things
that
we
expected.
Some
slow
uptake
on
source
generators
was
one
where
they
thought
would
take
some
time
for
people
to
really
dig
into,
and
people
are
getting
into
these
already
really
in
depth.
So
here's
an
example
of
using
core
rt
and
json
source
generator.
B
So
the
the
nice
thing
with
this
is
source
generator
at
build
time
generating
from
json
and
then
the
end
result
is
super
tight.
You
know
super
fast,
startup
time
and
super
tiny
size.
So
this
is
really
kind
of
an
experiment
at
this
point,
but
really
neat
to
see
this.
You
know
taking
json
files
and
then
building
stuff,
and
then
you
know
compiling
tree
shake
all
that
stuff
and
get
these
tiny
things
so
tons
of
cool
experiments
and
things
that
people
are
doing
right
now
with
source
generators.
B
Okay,
a
few
cool
updates
from
the
abp
team,
so
they're
they're
doing
some
great
stuff.
One
and
they've
been
around
for
a
while
abp
is,
you
know,
come
over
from
asp.net
boilerplate
and
they
have
a
an
open
source
version
and
a
paid
commercial
version.
This
is
a
very
sophisticated
system
that
does
all
kinds
of
things
with
it's
an
opinionated
framework
and
it
really
helps
you
build.
You
know
structured
applications
quickly.
B
What's
really
cool
here,
they're
announcing
a
new
blazer
ui,
so
they've
actually
brought
over
the
developer,
I
believe
from
blazer
eyes
to
help
build
this.
So
this
this
it's
a
full
blazer
ui
that
builds
on
top
of
the
abp
system,
so
they
do
a
walkthrough
of
how
to
build
this
as
you'll
see
it's.
You
know
it's
very
structured,
everything's
set
up
and
broken
up
between
the
api
and
the
you
know
front
ends.
B
So
a
lot
of
great
you
know,
support
for
all
kinds
of
things
and
and
of
course,
this
blazer
client,
so
good
stuff,
here's
a
post
describing
it
and
another
thing
they're
announcing
is
they've,
got
some
detailed
documents
as
well.
So
here's
updated
documents
and
they've
really
put
a
lot
of
work
into
this.
B
I
think
the
and
I'm
sure
you'd
agree
scott.
The
key
benefit
of
the.net
ecosystem
is
having
this
great
blend
and-
and
you
know,
you've
got
these
the
open
source
frameworks
out
there
and
then
you've
also
got
these.
You
know
amazing
products
that
people
build
and
support
for
years
and
have
have
great.
You
know
you
can
get
support
from
them
and
components
etc.
So.
A
Yep
and
we're
going
to
show
some
of
that
today
as
well.
B
Very
cool:
here's
a
neat
one:
this
is
from
the
exceptionless
team
and
they've
upgraded
their
production,
they're
running
on.net
five.
Already
they
couldn't
wait
so
so
they're
they're
talking
about
you,
know
they're
already
seeing
here's
some
reasons
why
they've
jumped
in
so
they've
got.
You
know,
of
course,
the
performance
improvements
are
amazing,
also
taking
advantage
of
c-sharp
features
like
record
support
and
docker
improvements
as
well.
So
just
a
you
know,
cool
to
see
people
jumping
in
trying
it
out,
and
so
of
course,
performance
is
always
a
big
thing.
B
Speaking
of
performance,
james
newton,
king,
just
tweeted
this
last
night,
so
he's
built
out
an
example,
and
this
is
showing
a
comparison
with
grpc
just
compared
to
jason
and
he's
seeing
you
know:
70
percent,
smaller
network
usage
and
deserialization
10
times
faster.
So
this
is
cool.
He's
got
this.
You
know
live
site
where
you
can
try
it
out
and
just
you
know,
cool
to
see
it.
So
this
is
this
fun
to
see.
B
I
love
how
james
is
always
building
experiments
on
top
of
things,
and
you
know
so
he's
writing
the
code,
but
he's
also
building
these
cool
experiments
to
test
it
and
and
understand.
What's
going
on
all
right,
scott
brady
talked
about.
This
is
an
advantage
in
net
five
with
certificate
management.
B
So
pemkeys
is
something
that
apparently
has
had
spotty
or
apache
support
in
windows
and
net
in
the
past,
and
he
talks
about
you
know
with
net
core
31.
You
had
to
do
a
few
things
and
you
really
kind
of
had
to
understand
the
details
of
how
pem
worked
and
with.net
five.
It's
just
a
few
lines.
So
this
is
pretty
neat,
including
you
know,
loading
a
pem
certificate
couldn't
be
much
simpler,
so
it
shows
loading
the
pem
certificate
in
just
a
few
lines
and
then
using
that
in
kestrel.
B
So
this
is
this
is
neat.
You
know
just
pointing
to
the
path
and
key
path,
so
something
I
wasn't
aware
of,
but
this
is
great
to
know
to
see
that
support
and
then.
Finally,
as
I'm
wrapping
up
one
of
the
cool
things
that
that
you
talked
about
and
that
came
out
at
ignite
was
the
these
support
for
blazer
in
azure
static
web
apps.
So
that's
just
such
a
cool
combo
of
having
you
know.
B
A
Think
it's
hard
to
grasp.
You
know
the
when
this
is
one
of
the
things
that
I
I
I
recorded
this
for
bill
on
for
ignite,
as
you
said,
and
I
wasn't
actually
happy
with
the
recording
that
part
of
the
recording,
because
I
didn't
think
I
described
the
value
prop
of
this.
A
A
And
the
whole
point
of
the
static
web
apps
is
what,
if
your
website
doesn't
need
to
do
any
of
that,
and
so
in
this
particular
case,
because
your
app
is
a
is
a
blazer
wasm
app,
the
server
doesn't
need
to
do
anything,
and
so
the
static
web
app
costs.
You
almost
nothing,
because
all
it
really
is
doing
is
serving
static
files.
A
There's
no
net
code
running
in
the
in
the
you
know
in
the
cloud
in
that
scenario,
which
means
your
cost
to
host
the
app
is
super
cheap
and
then,
as
you
said,
what
what
that
sample
app
does
is
it
has
an
azure
function
here
there
to
return
to
return
some
of
the
results
and
once
again
the
cool
parts
of
azure
functions,
is
they
only
run
while,
while
they're,
while
you're
calling
them
so
once
again,
if
your
app
set
idle
the
whole
day,
your
cost
to
run
is
almost
zero,
and
so
I
mean
it's
hard
to
get
my
mind
wrapped
around
the
idea
that
my
server
for
the
most
part
doesn't
do
anything
at
all
other
than
serve
static
files.
A
It's
literally,
the
equivalent
of
you
know
hosting
your
app
in,
like
you
know,
a
storage
blob
and
putting
a
url
to
it.
The
difference
here
is
this
tech
also
supports
you,
know
github
actions
and
stuff
like
that.
So
you
know
you
build
all
your
code
and
get
repo,
and
then
you
know
the
actions
fired
over
to
the
static
site.
B
Yeah
yeah,
it's
amazing,
it's
really
it's
something
that
we've
kind
of
or
you
know
people
have
been
doing
a
little
bit
already.
You
know
even
like
the
site
live.net
that
serves
this
up.
We
have
back
end
functions
and
then
we
have
front
end
and
it's
it's
being
converted
to
blazer
noun
is
using
react
for
a
bit
and
stuff,
but
you
have
this
pattern
of
a
single
page
app
and
then
a
back-end
function,
like
you
said,
only
paying
for
what
you
use
and
then
a
lot
of
time
for
these
spa
applications.
B
It's
it's
static
files,
you're,
delivering
and
then
you're
calling
back
for
those
functions.
So
it's
it's
really
nice
to
have
a
supported.
You
know
documented
way
to
do
all
this
yeah,
yep
cool
and
then
niels
also
just
wrote
a
walk
through
on
on
doing
this
as
well.
So
you
know
I
showed
the
docs
and
then
cool
to
see
community
going
in
and
writing
up.
This
is
his
walk
through
and
how
how
it
worked
for
him.
So
you
know
just
neat
to
see
the
integration
nice
blog
post
too.
B
All
the
you
know
showing
pushing
it
out
to
github
and
all
this
yeah
yep
neil
niels
does
quality
blog
posts.
I
always
love
his,
so
a
lot
of
you
know
screenshots
and
just
you
know,
cool
to
see,
and
then
things
like
digging
into
the
github
actions
and
stuff
so
and
then
the
last
one
I
want
to
share
was
just
the
link
over.
So
if
people
want
to
see
your
talk
from
ignite
and
all
the
stuff
this,
this
is
a
link
to
that.
B
That's
all
I've
got
it's.
It's
always
exciting
to
see
the
community,
the
excitement
and
the
exploration,
the
things
people
are
trying
out
and-
and
you
know
digging
in
and
stuff
it's
I.
A
I'm
surprised
every
week
every
month
that
I
see
the
crazy
stuff
that
people
build
and,
and
it's
so
exciting.
Let's
talk
about
something
cool,
john
all
right.
So
so
a
couple
things
to
think
about
here
is
we
have.net
conf
coming
up
on
november,
the
10th?
A
That's
our
virtual
online
conference,
we're
gonna
launch
down
to
five
and
as
part
of
that
I'll
kind
of
leak
out.
So
now
we
can't
back
out
for
sure
is
we
plan
to
have
scott
guthrie
in
the
keynote
of
gotten
a
cough
with
me,
and
I,
as
we're
trying
to
prepare
for
this.
In
fact,
I'm
going
to
meet
with
scott
in
about
a
week
and
a
half
and
we're
going
to
go
over.
You
know
what
his
segment
of
the
keynote's
going
to
be.
A
What
we've
been
thinking
about
doing
for
that
segment
of
the
keynote
is
talking
about
some
of
the
points
in
time:
in.net
wow,
because
there's
there's
been
lots
of
points
and
times
that
you
know
that
I
can
think
of.
If
you
go
way
back
in
time
to
like
oh
wait,
that
was
you
know
the
shipping
mvc
for
the
first
time
doing
it
with
a
in
a
semi
open
source
way.
We
at
least
put
the
source
code
out
there.
You
know
that
was
that
was
a
milestone.
A
A
A
tools
didn't
show
up
until
2017.
I
thought
today.
Maybe
we
would
would
briefly
talk
about
the
history
of
of
how
net
core
came
to
be.
B
A
I
actually
remember,
I
think
you
and
I
were
at
a
italian
restaurant
in
redmond,
because
you
were
up
visiting
and
we're
having
dinner
together
and
I'm
telling
you
all
this
stuff,
and
I
think
you
probably
thought
I
was
I
was
on
drugs
or
something.
A
But
it
is,
it
is
a
funny
point
in
time,
so
I'm
going
to
show
some
slides
that
we
put
together
back
in
2013.,
interestingly
enough
right
after
we
had
open
sourced,
you
know
the
all
the
aesthetic
on
you
know.
A
Back
in
the
day
when
we
started
thinking
about
doing
a
net
core,
we
knew
that
we
had
to
go
run
it
by
both
scott
guthrie
and
jason
zander.
A
You
know
scott
and
jason
were
that
were
the
two
big
vps
at
the
time
that
you
had
to
go
run
these
ideas
by,
and
this
is
going
to
sound
crazy.
I
think
we
spent
over
half
a
year
preparing
the
deck
for
scott
and
jason,
and
I
think
we've
talked
about
it
on
the
understanding
way
back
in
the
past,
but
we
we
spent
all
this
time
working
on
this
deck,
to
go
and
sell
the
concept
of
doing
a
new.net.
A
That's
the
open
source.
Well,
I
shouldn't
say
actually
open
source.
It
wasn't
going
to
be
open
source
at
the
time
it
was.
We
were
going
to
go,
sell
the
new
composable.net
that
would
replace
the
the
tech
that
we
have
today
and
the
thinking
was
around
the
2012
era
we
started.
Looking
at
you
know
is
this:
this
is
damian
edwards
and
I
we
were
looking
at.
How
do
we
take
system
web?
A
That's
the
web
stack
that's
inside
of
the
dynamic
framework
ford
and
the
only
way
that
we
could
take
it
forward
was
basically
to
go.
Take
it
out
of
band
we'd
already
done
nbc
out
of
band
we've
done
web
api
out
of
band
we've
made
entity
framework
go
out
of
band,
meaning
we.
A
To
have
a
version
we
could,
we
could
make
breaking
changes
in
and
not
you
know,
destroy
the
ecosystem,
and
so
levi
broderick
on
the
engineering
team
helped
us
and
we
actually
got
system.web
running
outside
of
you
know
we
had.
We
could
run
a
private
copy
of
system.web
on
on.net
framework
and
so
daniel
super
excited.
We're
like
yeah
we're
gonna
go.
B
A
Use
them,
and
so
we
thought
we
had
this
path
forward.
We
built
the
predecessor
to
kestrel.
How
do
you,
how
do
you
host
this
out
of
band
system
web
in
the
product
and
then
one
day
we
we
tried
to
make
it
work
with
visual
studio
and
we
hit
a
weird
wall,
mainly
because
visual
studio
loads,
a
copy
of
system.web.
A
It
can't
load
too,
and
so
we
had
to
go.
Look
at
okay.
Well,
do
we
change
visual
studio,
so
it
could
load
a
different
version
of
system.web
yeah.
I
think
it
was
levi
and
he
came
back
and
told
us
we
had
to
make
over
a
thousand
changes
to
visual
studio
and
I
didn't
think
the
odds
of
of
us
going
into.
B
A
Room
and
telling,
if
we
were
going
to
make
a
thousand
changes
to
visual
studio,
to
support
that
event
that
it
felt
like
we'd,
hit
a
the
end
of
a
wall,
and
so
we
had
to
do
some
some
next
generation
thinking,
and
so
let
me
share
this.
A
Now
you
can
yep
okay,
so
what
we
did
is-
and
you
can
look
at
the
dates
of
this-
I
plan
to
talk
to
scott
about
this.
When
I,
when
I
talk
to
him
in
in
the
keynote.net
conf,
we
started
something
called
project
k,
aka,
cloud.net
and
also
to
realize
this
was
the
asp.net
team.
This
is
not.
We
were
not
the
dot-net
team,
we
were
the
as.net
team,
and
so
we
primarily
think
from
the
web,
and
I
just
want
to
be
very
clear.
A
We
were
not
part
of.net
time,
and
so
we
wrote
a
deck
and
the
deck
as
we
go
through
it.
It's
not
meant
to
be
a
negative
deck.
It
was
meant
to
be
something
that
would
change
your
thinking,
meaning
that
you
know
we've
been
shipping
a
product
for
a
long
time
and
you
get
close
to
the
thing
you're
shipping.
Sometimes
you
need
to
go
back
out
to
a
5
or
10
or
15
000
foot
view
and
go.
A
What
are
the?
What
are
all
the
challenges
and
the
problems
with
the
tech
that
we're
building,
and
so
this
is
a
deck
that
kind
of
talked
about.
This
is
the
initial
ideas
around
this
and
then
david
fowler.
I
think
we're
the
first
prototype
of
what
became
asp.net
core
and.net
core
that
we
then
showed
around
november
of
this
year
at
the
mvp
summit
of
2013.,
and
so
unless
again,
as
I
said,
don't
look
at
this
slide
as
being
a
negative
slide.net
framework,
an
amazing
piece
of
attack.
A
But
you
know
these
are
the
the
challenges
we
had
with
the.net
framework
number
one
is:
is
it
was
old
and
what
I
mean
old,
I
mean
it's
full
of
a
bunch
of.
If
you
notice
down
here,
contains
obsolete
server
components.
You
know
the
dotnet
framework.
When
we
started
this
journey,
there
was
like
two
versions
of
workflow
in
dotnet
framework
yuck
there
was
something
called
hailstorm
which
was
a
predecessor
to
like
you
know,
azure
active
directory.
A
There
was
there
was
lots
of
just
dead
code
inside
of
there,
but
if
you
think
about
it,
it
was
monolithic,
meaning
it
was
one
huge
thing:
it
was
an
os
component.
This
is
one
of
the
biggest
challenges
we
had.
It
was
the
part
of
windows.
If
you're
part
of
windows,
you
belong
in
windows
servicing.
You
belong
in.
That's
why
it
says
heavy
service
servicing
burden.
You
belong
in
compatibility.
A
Yeah
exactly
it
was
so
hard
to
get
to
work
together.
We
did
all
kinds
of
attempts
there
was.
There
was
back
in
the
net
3.5
time
frame.
We
did
something
called
red
bits.
There
was
parts
of.net
you
could
touch.
There
was
parts
of
that
that
you
could
not
touch
and
trying
to
solve
these
compatibility
burdens.
There
was
performance
issues,
those
are
historic.
I
mean
we've
talked
about
these
for
many
many
times.
A
You
know
when
we
built
asp.net
system
web,
we
meant
for
it
to
be
semi-compatible
with
active
server
pages,
and
so
we
baked
a
bunch
of
stuff.
In
that
you,
don't
you
don't
need
back
in
those
days,
you
probably
wear
this
john.
You
know
there
was
this
notion
of
modules
that
loaded
in
in
asp.net
and
we
defaulted
all
the
modules
in
whether
you
needed
those
capabilities
or
not.
We
defaulted
them
in
and
they
were.
A
It
was
hard
to
know
how
to
turn
them
off
yeah
and,
of
course,
we're
the
web
team,
and
so
we're
like.
Why
does
it?
Why
does
dot
net
contain
non
non-server
components?
You
know
why
does
your
web
app
need
to
have
winforms
or
wpf
or
any
of
those
things,
and
you
can
also
see
that
windows
phone
forked
nets
into
their
own
specific
skus.
A
A
And
so
the
idea
behind
project
k
and-
and
if
you
wonder
where
the
k
comes
from,
we
had
we
had
been
doing
a
project
called
katana,
and
it
was
a
way
that
we
were
kind
of
disentangling
some
of
the
bits
in
asp.net
at
the
time
and
so
project
k,
just
kind
of
took
the
it
was
meant
as
an
evolution
of
katana,
which
is
why
it's
that's
where
the
cake
came
from.
Interestingly
enough,
if
you
want
trivia,
the
microsoft
band
team
was
also
called
project
k.
A
We
only
know
that,
because
we
we
already
had
grabbed
the
aliases
inside
of
the
exchange
server
at
microsoft
and
they
came
to
us
asking
could
can
we
can
we
have
that
and
we're
like
no
we're
using
it,
but
what's
here
is
it
was
meant
to
be
a
lean,
composable
net
stack
that
provides
you
know,
modern
web
frameworks
and
cloud
scenarios,
and
so
you
can
see
this
a
small
core
clr
that
can
only
contain
fundamental
building
blocks
and
I'll
talk
as
we
go
through.
A
So
what
we've
changed
since
this
since
then
framework
to
be
an
ecosystem
of
packages?
We
already
moved
asp.net,
you
know
nbc
and
web
api
and
energy
framework
and
web
pages.
They
were
all
based
on
packages.
It
was
our
way
of
you
know
being
out
of.net
framework,
letting
us
move.
So
the
idea
was
everything
would
be
packages
in
this
world
packages.
You
know
in
framework
ad
crazy,
versioning
and
servicing
rules,
so
we
want
to
simplify
those
that
true
side
by
side.
That
was
an
important
part
of
this.
A
B
The
whole
side-by-side
thing
is
huge,
and
people
may
forget
about
that
now,
but
I
I
remember
early
on
as
a
dot-net
developer
going
and
I
went
to
some
training
on
net
2
like
dot-net,
you
know
net
framework
2-0.
I
was
really
excited
about
the
membership
and
some
of
these
other
features
and
then
I
came
back
and
I
was
working
at
a
financial
company.
I
said:
okay,
we
got
to
get
on
this.
B
A
I
I
same
thing
happened
to
me
john,
I
was
I
was
actually
working
at
a
company
and
and
dotnet
2
had
just
come
out
and
you
know
tons
of
new
features,
and
I
I
had
the
same
same
fight.
I
finally
finally
was
able
to
get
get
them
to
give
me
one
server
that.
B
A
I
could
play
with
dot
net
two
on
and
I
you
know
you
know
I
I
was
just
like
you,
john
I'm
like
telling
them
we
can
be
so
much
more
efficient.
These
apps
are
gonna,
be
so
much
better,
but
the
the
cycles
when
the
when
the
framework
is
kind
of
like
a
part
of
the
operating
system
just
makes
it
super
slow.
I
mean
most
companies
back
in
the
in
that
era,
only
upgraded
to
do
not
do
not
new.nets
when
they
upgraded
to
new
versions
of
windows.
So
you
had
to
wait.
A
You
know
for
a
new
windows
and
sometimes
those
windows
gaps
would
be
yeah,
and
you
can
see
here
here
were
some
of
the
ideas
that
we
had
for
customers.
You
know
modern
cutting
edge
web,
simple,
easy
developer,
experience,
surface
area,
meaning
that
you
know
you
don't
have
all
the
non-web
stuff
get
in
your
face.
That
is
the
the
side-by-side
comes
up
again.
A
B
Was
a
as
you
mentioned
before,
that
was
a
problem
before,
because
when
we
were
comparing
performance,
you're
loading
up
this
whole
giant.net
framework
that
could
do
all
kinds
of
stuff
like
manage.
Like
you
said,
there
were
bits
from
five
ten
years
ago
for
different
stacks
that
were
still
in
there
right.
So.
A
And
and
we've
done
dumb
stuff
too,
like
mvc
and
web
api.
Well,
how
do
they
work?
Well,
there's
there's
classes
called
controllers.
Well,
how
do
you
find
the
classes
that
are
called
controllers?
Will
you
load
all
the
assemblies
and
look
for
it
and
we
did
other
crazy
stuff
john?
It
used
to
be
that
when
you
would
boot
up
an
asp.net,
app
nvc
would
look
for
all
the
controllers
and
then
web
api
would
look
again.
A
We
didn't
think
about
hey,
that's
kind
of
a
common
thing.
Why
don't
we
just
do
it
once
and
share
it
across
all
the
all
the
stuff-
and
you
know
you
can
see
obviously
improved
throughput
performance
has,
you
know
been
a
big
thing
for
us
optimized
for
cloud
diagnostics.
That
was.
That
was
our
revelation
that
you
know
in
system
web.
There
wasn't
really
a
diagnostic
story
at
all.
There
was
a
web
page.
You
could
hit
on
a
web
form
application.
A
That
would
kind
of
show
you
the
the
page
date,
but
there
was
no
real
diagnostics,
you
know
in
there
and
there
was
no
good
way
to
turn
it
on
when
you
put
it
to
put
it
to
the
cloud
and
then,
of
course
this
was
all
before
there
was
big
big.
You
know,
azure
and
and
aws
and
google
cloud.
So
you
know
you
can
see
hosters
at
this
point.
You
know
it
wasn't.
Clouds
it
was
it
was.
A
B
And
that
was
always
a
frustration
too
the
same
we
went
from
talking
about
how
like
big
financial
companies,
but
also
hosters.
You
would
have
your
site
on
a
host
and
they're
running
an
old
version
of.net
and
you're,
like
logging,
a
support
ticket
like
when
are
you
going
to
put
the
new
thing
on
there?
You
know.
A
and
we'll
go
through
one
two,
one
one
or
more
of
these,
and
we
can.
We
can
jump
to
some
other
other
stuff.
I
thought
it's
just
kind
of
fun
to
look
at
this.
You
can
see,
obviously
the
the
primary
workloads
for
us
were.
You
know
web
pages
nbc
rest
with
web
api,
real-time
collaboration
that
was
signalr
and
being
able
to
use
the
azure
sdk.
A
You
know
in
nvc
web
pages
web
api
signal.
Our
code
should
not
be
super
hard
to
port.
Over
most
third-party
libraries
should
have
you
know
a
project
k
version,
and
this
is
where
you
start
seeing.
Parts
of
full.net
will
not
be
carried
forward
to
project
k.
This
is
realizing
that
there's
some
legacy
in
there
that
doesn't
need
to
come
forward.
A
A
This
is
something
that
we
still
hopefully
are
striving
to,
and
at
today
we
don't
want
to
bring
obsolete
stuff
into
you
know,
that's
why
people
ask
us:
why
did
you
not
bring
something
over
well
if
it's
something
that
you
would
if
you
would
build
an
app
today
with
it,
we
would
bring
it
if
you
would
only
build
an
app
that
you're
porting
from
from
10
years
ago,
that
doesn't
make
sense
to
come
over
and
you
can
see
that
it
should
just
be
packages
everywhere.
A
Making
the
the
rules
on
how
we
load
assembly
is
much
simpler.
You
know,
remember
all
the
assembly,
binding,
redirects
and
all
the
crazy
stuff
and
the
projects
back
then
john,
that
just
made
your
brain
hurt.
B
Oh
man,
yeah
and-
and
that
was
something
where
it
was-
you
could
show
off
at
work
by
knowing
how
to
do
all
the
binding
redirects
and
all
that
kind
of
stuff.
But
it
also
was
a
huge
headache
and
it
was
part
of
why
you
you
I
working
at
at
companies
would
be
like
hey
we're
going
to
start
deployment
friday
night
and
everyone
planned
to
you
know
you
may
have
to
stay
all
weekend
until
we
get
it
working
on
the
production
server.
You
never
knew
if
it
was
going
to
work.
A
Yeah,
so
you
can
see,
number
four
was
everything's
packages
and
the
number
five
was
basically
that
that
you
you
deploy.
Basically,
what
number
five
is
saying
is:
there's
no
gak,
which
was
the
global
assembly
cache
that
we
had
in.net
before
and
once
again,
as
you
said,
it
was
one
of
those
dangerous
things
that
if,
if
you
had
something
you
know
the
gap,
overrode
everything
and
so
by
putting
something
in
the
gap,
you
could
actually
break
an
application,
and
so
we
wanted
to
get
away
from
that
now.
A
B
B
A
I
remember
us
preparing
it:
wasn't
this
exact
deck,
but
I
was
working
on
this
for
a
long
time
and
a
few
of
us,
I
think,
was
dimitri
robson,
david
fowler
and
myself.
A
We
get
on
scott
guthrie
and
jason
xander's
calendar
and
we're
playing
we're
gonna
go
present
this
for
an
hour
and,
and
you
know,
we're
expecting
a
fight
and
we
roll
into.
I
think
it
was
scott
guthrie's
office
at
the
time,
and
we
start
start
this.
The
deck
and
we
get
through
like
three
or
four
slides
and
scott
and
jason
look
at
us
and
they
go
well.
Jason
goes
well.
This
is
this
doesn't
go
far
enough.
He
goes.
A
This
needs
to
be
fully
open
source
and
it
needs
to
be
cross-platform
and
remember.
This
is
2013.
This
is
not
the
microsoft
of
today
and
I
I
think
our
jaws
just
hit
the
floor.
We're
like
oh
wow.
I
mean
they
basically
threw
out
the
gauntlet
to
just
go
completely
bold
and
so
the
whole
meeting
lasted.
I
think
under
10
minutes,
wow
yeah.
They
knew
that.net
needed
to
make
a
transition,
and
so
there
was
there
was
no
no
convincing.
But
that's
that
is
a
a
fun
part
of
the
original
story.
B
B
A
It's
or
and
if
not,
I'm
pretty
sure,
that's
what
project
n
was.
It
was
dot
net
native
and
the
data
for
people
that
don't
know
what
that
is,
that's
in
what
ended
up
being
the
version
of.net
or
this
style
of.net,
that
is
in
uwp
applications
and
so
that's.net
that
we
actually
compile
to
native
code
and
and
so
it
doesn't
have
to
be
jitted
when
the
application
starts.
A
We
were
so
that
is
actually
covered
here
on
this
slide
here,
where
it
says
simple,
lean,
developer
experience.
That
would
be
an
example
of
where
the
project
system
kind
of
stuff
would
come
into
play
here.
Obviously,
you
know
we
went
through
a
a
whole
model
of
project
system
where
we,
you
know,
first
off
project
k
had
its
own
project
system.
That
was
not
in
visual
studio.
A
It
was
not
based
on
cs
project,
all
that
was
our
magical,
json
project
system,
and
so
one
of
the
one
of
the
key
things
of
this
was
which
I
don't
think
this
deck
talks
about.
A
It's
too
earth
too
early.
One
of
the
other
chances
of
this
was,
I
think,
probably
after
the
scott
and
jason
meeting,
once
we
decided
that
we
were
going
to
go
cross-platform
that
brought
the
requirement
that
you
might
not
be
in
visual
studio
to
build
your
app.
B
A
A
So
you
know
we
built
our
own
project
system
to
start
with
and
then
obviously
before
we
shipped,
you
know
version
one
of
of
the
of
the
tooling
in
in
march
of
2017,
we
moved
back
to
the
cs
proj
that
already
was
used
for
net,
but
we
moved
to
a
simplified
version
of
it,
which
is
yeah
the
other
day.
B
Because
those
cs
projects
like
you
were
saying
there
would
be
hundreds
of
lines
of
xml
and
you're
so
scared
to
break
something.
You
know
there
would
be
like
a
oh.
I
want
to
do
a
binding
redirect
or
I
need
to
add
a
module
or
a
handler.
But
if
I
do
this
or
or
edit
a
configuration,
you
know
a
connection
string
and
you're
like
well.
I
might.
A
Break
my
whole
app
exactly
or
yeah,
if
visual
studio
that
arrow,
when
that
happens,
you
know
it
won't
load
your
project,
the
whole
project,
the
files
disappear
and
you
basically,
you
can
only
add
that
one
file
and
you
just
have
to
hope
how
do
I
find
out
in
600
lines
of
xml,
where
I
forgot
to
close
something
or
I
put
a
wrong
wrong
character
somewhere
that
that
was
there.
But
I
I
don't
you
know
the
the
building
our
own
project
system
really
was
something
that
happened
after.
A
I
think
the
cross
platform
thing
came
to
play
it's
crazy
to
look
at
his
deck
and
not
see
about
the
open
source
or
cross
platform
which
are
such
big
tenants
of
how
we
build
you
know.net
today,.
A
The
biggest
thing
is
not
being
in
windows,
is
windows
brings
that
compatibility
requirement
and
it
also
brings
the
servicing
burden.
So
you
know
we
still
support
and
we
have
teams
that
work
on
fixing
bugs
and
net
3.5
today
and
I
think.net
3.5
shipped
shipped
in
2008
yep.
So.
A
Ago
and
we're
still,
you
know,
that's
that's
the
challenge
of
being
in
windows
once
you're
in
windows.
Your
lifetime-
I
was
just
talking
to
on
a
two
customers
in
the
last
week,
were
asking
about
the
support
of
active
server
pages
active
server
pages,
which
was
the
predecessor
to
asp.net
it
shipped
in
1996,
so
still
supported
today,
24
years
later,
yeah.
B
A
I
see
a
question
over
there.
What
was
the
rationale
behind
going
open
source?
Was
it
competition
from
node.js
and
I
would
actually
say
no,
it
was.
It
was
not
the
competition
from
node.js.
We
started
open
source
and
open
sourcing
parts
of
asp.net.
You
know
years
before
2008,
you
know
we
we
introduced
nbc
and
then
that
in
an
open
source
flavor
in
2012,
we
took
contributions
on
nbc.
A
What
what
caused
open
source
to
happen
was?
You
know,
as
my
team
traveled
around
the
world
talking
to
customers,
we
saw
this
slow
change,
starting
to
happen
where
companies
were
embracing
open
source.
I
remember
being
somewhere
talking
to
a
customer
and
they
said
we're.
Gonna
pay
for
operating
systems
and
that'll
be
the
only
close,
closed
source
component
that
we're
gonna
use
and
everything
else
will
be
open
source,
and
so
the
asp.net
asp.net
team
heard
more
and
more
of
that,
as
as
time
went
on
and
we
realized.
A
How
do
you
compete
against?
You
know
all
data,
because
we
you
know
mdc
was
was
more
going
after
ruby
on
rails,
but
it
was
open.
How
do
you
compete
against
open
when
you're
closed
and
the
reality
is
being
open?
Source
is
a
way
of
leveling.
The
the
playing
field
so
net
had
a
chance
to
be
something
that
somebody
would
consider.
B
Did
you
expect
at
the
time
like
there
was
there
were
parts
of
asp.net
like
you
said
that
were
open
source
on
codeplex
and
they
they
took
some
community
contributions,
but
it
was
primarily,
it
was
like
here's
our
source
code.
Did
you
expect
the
amount
of
community
pull
requests
and
community
involvement
in
project
k.
A
I
don't
think
it
did.
You
know
all
up
either
if
you
go
back
to
one
there's
a
slide
that
you
showed
earlier.
It
was
one
of
the
it
was
the
one
of
my
ignite,
slides
and
I'll
bring
it
up
my
machine
just
just
to
share,
because
I
have
it
here
too,.
A
This
was
the
slide
that
you
were
showing
john
earlier
yep,
and
you
know
on
this
line.
We
talked
about
being
one
of
the
top
30
highest
velocity
oss
projects,
and
so
we've
done
the
analysis
on
github
to
go,
find
out.
Beth
massey
has
an
amazing
slide
somewhere
john.
It
actually
shows
all
the
different
texts
and
it
shows
where
dot
net
you
know
compares
to
things
like
linux
and
kubernetes
and
stuff.
A
Like
that,
I
I
never
expected
that,
because
for
the
most
part
we
open
source
late,
you
know
one
of
the
things
that
that
that
I
regret
is
is
not
finding
a
way
to
get
us
open
source
faster,
because
if
you
look
at
the
package
ecosystem
around
java,
it's
it's
better
than
ours
and
that's
because
they
were
open
source
earlier,
and
I
mean
a
lot
earlier,
and
so
I
think
it
was
a
barrier
to
entry
to
our
market
and
it
was
a
barrier
to
entry
for
writing
packages
for
us
and
and
a
variety
of
other
things,
but
because
we
really
open
sourced
in
2014.
A
That
was
when
we
announced.net
core
was
coming
and
we
and
we
fully
open
sourced
all
the
bits
to
think
that
we
open
source
in
2014
and
we're
in
the
top
30
in
github.
Then
that
is
shocking
to
me.
Actually,
it
makes
me
very
happy
too.
It's
it's.
B
Yeah
and
it's
exciting
to
see,
then,
as
you
know,
people
are
digging
in
and
trying
out
things,
and
you
know
some
of
the
performance
gains
and
all
those
things
having
it
be,
a
collaborative
effort
with
the
community
and
and
seeing
it.
You
know,
including
like
the
early
things
where
everyone
was
community,
was
jumping
in
on
kestrel
and
then
also
seeing
people
giving
pull
requests
for
like
linux
transports,
and
you
know
like
these
advanced
certificate
management
and
other
things
that
are
operating
system
specific.
It
makes
a
huge
difference.
I.
A
Want
to
grab
one
of
these
questions.
There's
a
great
question
here.
The
one
that's
windows
has
a
certain
long
support
time
frame
which
is
beneficial
to
some
customers.
Did
this,
you
know.
How
do
we
think
about
this?
A
I
think
this
is
a
a
a
crazy
important
thing
to
understand
when
we
think
of
net
today
our
support
model
that
we
have
in.net
core
slash.net5.net6
is
designed
we
ship
every
year
and
then
every
other
year
is
a
long-term
support,
release
and
the
which
means
you
have
roughly
three
years
of
support
on
a
given
lts.
A
If
you,
if
you
pick
up
3.1
you're
good
for
three
years,
you
can
run.net
six
for
three
years
and
you
might
ask
you
know
that
windows
support
policy
of
supporting
stuff
for
10
15
20
years.
It's
great,
you
know
it
is.
A
It
is
amazing
that
you
can
go
build
an
application
and
not
touch
it
for
20
years,
and
it
still
runs-
and
we
did
not
do
that
in
the.net
core
world,
because
in
that
world,
when
I
was
when
we
were
working
on.net
framework,
a
good
percentage
of
our
team
worked
on
on
supporting
the
old
releases
mm-hmm
and.
A
B
A
B
A
Yeah,
that's
a
that's
a
great
question
and
it's
one
that
comes
up
all
the
time
and
in
many
cases
people
are.
You
know
I'll.
Give
you
a
great
example:
I'm
I'm
in
asia
in
january
of
this
year
and
I'm
sitting
with
a
customer
and
they're
they're
deciding
they're
they're
in
the
process
of
rewriting
their
platform
and
they're
gonna.
Do
it
on.net
core
and
I
I
think
they
turned
through
dot
net
core
3.1
and
they're.
Asking
me
the
question:
it's
like
yeah.
A
We
have
to
wait
to
you
know.net
fix
after
after
3.1,
but
during
that
conversation
they
said
something
really
funny
they're
like
yeah.
We
update
angular
every
six
weeks
and
I'm
like
you,
you
don't
want
to
update
net
for
three
or
four
years,
but
you're
willing
to
update
the
javascript
framework
for
your
client.
A
Six
weeks
and
I'm
like,
why
am
I
different
and
the
room
got
quiet
and
they
could
not
tell
me
why?
Because
here,
here's
here's
the
reality.
Maybe
you
don't
want
to
upgrade
because
you're
afraid
of
a
breaking
change
totally
get
that,
but
even
if
you
stay
on.net
core
3.1,
if
you
want
to
stay
supported,
you
need
to
update
your
dot
net
core
3.1.
Almost
every
month
we
we
patch.net.
A
That
means
we
skip
a
month
occasionally,
but
for
the
most
part,
there's
typically
a
servicing
update
every
month,
and
so
I
would
ask
a
customer:
are
you
actually
grabbing
those
servicing
updates
every
month?
Are
you
or
are
you
actually
sticking
with
the
version
that
came
in
in
december
and
sitting
on
it
for
a
couple
months?
A
If
you
call
our
support,
our
support
will
only
support
you
if
you're
on
the
latest
patch
version,
because
we
don't
go
and
patch
old
versions
in
the
history
we
only
so
if
you're
having
a
problem,
you
would
always
have
to
verify
that.
So
I
would
challenge
if
you're
a
net
customer,
you
should
be
updating
once
a
month
anyways
and
if
you're
updating
once
a
month.
The
move
from
a
three
to
a
five
should
not
be
a
big
deal
now.
A
We
as
as
the
net
team,
I
will,
I
will
be
very
honest
and
and
damien
would
tell
you
the
same
thing
for
versions.
One
two
and
three,
we
kind
of
changed.
The
way.net
was
put
together.
Every
single,
every
a.net
core
was
put
together
every
single
one
of
those.
Those
are
not
the
easiest
transitions,
but
I
do
believe
we're
maturing
now
and
as
we
mature,
we
actually
talk
about
this
during
planning
our
our
goal
without
n6
is,
it
should
take
any
dot
net
5
customer,
no
more
than
10
minutes
to
upgrade
the
rapto.insights.
A
As
I
said,
that's
a
that's
a
planning,
it's
a
it's
a
it's
a
goal,
I
can't
say
well,
but
it
is
something
that
that
we,
you
know
as
the
platform
occurs.
We
want
to
make
moving
from
version
to
version
easier
and
then
I
would
hope
you're
just
on
the
on
the
latest
monthly
version.
Every
month
and
the
bump
to
five
to
six
and
six
to
seven
is
not
a
big
deal.
Now
there
are
customers
that
do
have
very
valid
needs
to
have
a
long-term
support.
A
A
We
talked
to
some
medical
companies
that
do
mris
and
stuff
like
that,
and
and
they
don't
have
the
ability
to
go
in
and
and
and
patch
the
device
that
said.
Hopefully,
the
risk
of
those
devices
is
less
as
well.
If
they're
not
internet,
connected
or
you
know,
hopefully
you're-
not
even
a
web
browser
on
your
on
your
mri
machine.
B
Yeah
wow
yeah
and
then
definitely
like
a
lot
of
the
choices
that
you're
talking
about,
with
the
design
being
able
to
run
side
by
side
and
being
able
to
run
cloud
native,
and
all
these
sorts
of
things
make
it
easier
to
update
because
you're
not
worrying
about.
I
can
deploy
a
you
know:
net
5
and
a
net
31
version
of
my
application
on
the
same
server
and
not
worry
or
I
can
use
docker
or
I
can
you
know
all
those
sorts
of
improvements.
A
There
so
and
we'll
get
we'll
get,
but
we'll
get
better,
we'll
get
better
in
those
cases,
and
we
want
to
provide
more
tools,
you
know
as
well.
I
did
want
to
show
one
more
demo,
I'm
not
really
pretty,
but
I
want
to
show
a
demo
john
all
right,
so
you
were
showing
components
earlier
of
stuff,
and
this
is
actually
a
preview
of
something
I'm
going
to
show
dot
net
conf
and
I'm
going
to
show
it
in
the
context
of
one
control
vendor
I
want
to
preface
before
I
show
this.
A
We
have
lots
of
control
vendors,
I'm
not
trying
to
give
one
control
vendor
more
love
than
the
next,
but
on
this.net
core
3.1
journey.
A
You
know
we
brought
winforms
in
wpf
into
the
into
the
platforms,
and
we
did
that
you
know.
If
you
look
at
those
those
reasonings
in
project
k,
we
said
not
to
bring
obsolete
things
in,
we
don't
consider
wpf
and
winforms
to
be
obsolete.
You
know
we
have
millions
of
customers
using
them
every
day,
but
along
this
journey,
we've
been
slowly
and
I'll
switch
to
share
my
screen
again:
real,
quick
and
I'll
go
to
a
vs.
A
Hopefully,
you've
not
shown
this
before
this
is
so.
We've
been
working
on
the
on
the
designer
for
winforms
for
dot-net
core,
and
so
you
can
see.
We've
got
all
of
the
the
primary
controls
now
in
the
designer.
If
you're
on
the
latest
version
of
visual
studio
they're
all
there,
the
one
missing
piece
is:
how
does
a
third
party
control
vendor
get
their
stuff
into
the
into
the
platform?
So
if
I
go
look
at
my
cs
proj
over
here,
you
will
see
that
I'm
actually
referencing
minimize
much
stuff
here.
A
I
am
referencing
the
telerik
dotnet
5
controls
and
televic
has
now
got
their
controls
in
the
right
shape,
so
they
show
up
in
our
designer.
So
this,
if
I
click
on
it
here
and
bring
up
the
properties,
I
will
scroll
up
to
the
top.
A
You
can
see
that's
a
rad
check
checklist
box.
B
A
So
you
can
see
all
the
controls
here
so
right
up
in
this
in
the
toolbox,
and
then
I
I
had
drag
one
over
here,
put
it
on
the
form
and
if
I
go
to
the
properties
for
that
and
go
down
to
there's
columns,
I
want
to
put
some
columns
in,
but
you
can
see
all
the
all
the
ui
looks
up
and
does
all
the
right
stuff-
and
so
hopefully
we're
pretty
close
john
to
being
able
to
have
all
of
the
third
party
control
ecosystem
start
showing
up
directly
in
in
the
windfall
designer
which
will
end
the
journey
of
capabilities.
A
B
That's
such
a
cool,
you
know
like
there's
dot-net
is
such
an
interesting
ecosystem,
because
now
we've
got
this
open
source
cross
platform,
tons
of
community
built
things,
and
then
we
also
have
that
component
vendor
ecosystem,
which
is
important
for
businesses
that
want.
You
know
to
buy
something
with
a
support
contract
and
you
know
like
they
want
to
pay
something
or
maybe,
as
a
startup
at
the
beginning.
B
I
want
to
buy
some
components
and
then
later
I'm
going
to
write
my
own
or
whatever,
so
having
that
kind
of
selection
of
you
know,
the
the
broad
ecosystem
is
is
really
cool.
So
that's
exciting
to
see
that
wow.
A
I
can
I
can
show
one
more
that
you
probably
have
not
done
on
the
show.
So
let
me
just
load
up
something
here.
While.
B
B
B
You
know
the
asp.net
community
stand-up
started,
I
believe
in
2014.,
and
a
lot
of
the
early
shows
were
here's
a
sketch
on
a
whiteboard,
as
you
know,
damien
and
david
and
folks
were
kind
of
figuring
out
how
they
wanted
things
to
work,
and
now
we
have
three
shows
a
week
and
all
the
different
areas
and
stuff
so.
A
B
B
A
A
B
One
there's
yeah
and
this.
A
Is
this
is
the
this?
Is
a
net
core
3.1
application?
It's
kind
of
your
minimal
web
app.
You
know
it's
got
the
program,
it's
got
the
main
all
of
that
stuff.
This
is
the
same
one
using
net5
and
it's
using
top-level
programs.
So
there
is
no
main.
You
just
basically
write
the
code.
You
know
inline
in
the
page.
A
Let's
go
back
here,
here's
the
three
point,
one
one
and
you
know
it's
your.
This
is
a
a
lightweight
application
when
it
comes
to
the
amount
of
stuff
on
the
on
the
page.
Here
you
got
your
your
main.
You've
got
your
you
know,
class
programs,
you
have
to
have
a
you
know.
Every
dotnet
app
has
to
have
a
class
that
surrounds
the
main
code.
A
You've
got
the
main
code,
that's
the
entry
point
to
the
application
and
so
on,
and
then
here's
the
same
version
of
that
running
on
net
five
and
notice
that
the
main
the
program
is
gone,
the
main
class
is
gone
and
the
main
is
gone.
A
So
this
is
using
top
level
programs
and
you
can
see
we
can
knock
a
chunk
of
lines
of
code
off
of
here
and
you
get
to
a
an
api
that
has
more
of
the
footprint
of
something
that
maybe
in
something
like
node.js
would
look
like
kind
of
our
one
of
our
goals
to
make
it
like
that.
But
one
of
the
cool
parts
of
this
is
have
you
have
you
talked
about
the
linker
at
all.
B
David
showed
a
little
bit
with
the
linker
yeah.
He
showed
some
experiments,
it's
and
and
those
things
you
know.
I
think
it's
easy
to
look
and
go
like.
Oh,
it's
just
a
few
lines
of
code
we're
cutting
out,
but
the
idea
when
you've
got
these
micro
services
and
and
the
whole
service
code
is
just
a
few
lines.
B
A
Okay,
so
let's
let's
go
and
let
me
share
again
here-
and
this
is-
I
don't
know
what
david's
shown,
but
this
is
work
that
happened
that
david
did
after
the
microservice
version
of
the
show
is.
Is
this
is
the
three
one
application
and
for
folks
who
don't
know
it,
you
can
actually
right
click
on
one
of
these
types
of
applications.
A
You
can
do
publish,
but
I
don't
want
to
publish
to
the
cloud.
I
want
to
publish
it
to
a
folder
and
once
I
do
that,
I
have
the
ability
to
go
and
edit
some
of
the
settings
and
if
I
go
down
here,
hopefully
you've
shown
this
on
the
show
before
I
can
decide
between
being
self-contained
or
framework-dependent,
and
so
in
this
case
I've
decided.
I
want
to
be
self-contained,
then
I
can
choose
down
here.
A
A
That
means,
instead
of
having
a
pile
of
files,
we
just
have
a
a
single
file
and
then
there
is
this
trim
unused
assemblies
which
is
called
for
preview,
because
we
don't
have
good
tooling
to
prevent
you
from
hurting
yourself
with
this
today,
so
meaning
that
the
linker
could,
if
you
load
an
assembly
with
reflection-
or
you
know,
if
you
load
any
dynamic
code,
that
loads
an
assembly,
the
linker's
not
going
to
see
and
can
blow
it
out.
A
But
but
you
know,
if
I
save
that
I
can
now
click
the
publish
button
and
we'll
produce
a
single
trimmed
super
small
microservice
application.
And
let's
go
see
what
this
is
going
to
look
like.
B
Here
and,
and
so
some
advantages
of
that,
like
one
is,
if
you're
building
docker
images
or
whatever
that
you
get
more
density,
but
also
you
get
like
quicker,
startup
and
better
memory
as
well,
right,
yeah,.
A
Well,
it
depends
there's
other
options
there
there
was
ready
to
run,
but
that
ready-to-run
option
would
help
me
with
startup,
because
that
what
that
does
is
that
prevents
us
having
to
do
some
shooting
when
the
application
starts
up?
A
But
you
can
see
here
it's
it's
still
crunching,
because
it's
doing
a
link,
the
end
result
that's
going
to
be
about
40,
megs
and
in
dot
net.
5
we've
cut
that
down.
By
about
now
there
you
go
45
megs,
just
finished
it's
because
I
moved
it
to
x64.
I
think,
but
if,
if
I
switch
over
to
my.net
5
version
of
the
same
thing
and
we
do
the
exact
same
exercise.
A
We
will
get
a
smaller
file
and
this
is
work
that
david
did
it's
the
the
whole
team
is
done,
but
what
we
did
is
we
need
to
have.
We
need
to
go.
Make
our
frameworks
more
tell
the
leaker
linker
more
about
what
they
need,
and
so,
after
we
did,
the
microservice.net
conf
david,
went
and
attributed
a
bunch
of
the
framework
that
looks
all
the
same.
A
Yeah,
do
you
mix
and
I'll
do
this?
Do
the
same
thing
we
did
before
while
that
runs,
I
said
we
still
call
this.
I
would
call
it
you
know
experimental,
meaning
that,
and
you
can
see
the
end
results
here-
is
21
megs,
it's
still
producing
it.
A
So
in
a
second,
so
you
can
see
we
cut
it
down
by
about
half,
but
I
hope
in
the
dominant
six
journey,
we're
gonna
we're
gonna
do
a
lot
of
work
to
go,
make
our
libraries
we
want
to
provide
the
tools
for
library
authors
to
make
their
libraries
linker
friendly.
We
want
to
go,
do
more
what
david
did
to
hp.net
and
and
attribute
our
own
frameworks
to
make
them
more
linker
friendly?
So
this
ends
up
being
a
maybe
giving
you
a
an
analyzer
that
you
can
run.
A
That'll
show
you
where
you
might
have
wobbly
stuff,
you
know,
but
the
end
goal
is
basically
to.
A
Keep
making
you
know,
dominate
applications
smaller
and
lighter
and
faster
to
start
up,
and
once
we
do
that
step
it
it
it.
You
know,
I'm
sure
folks
have
talked
about
core
rt
before
doing
all
the
stuff
to
fix.
The
linker
will
help
the
core
t
efforts
as
well
and
make
more
net
run
on
on
things
like
the
quarantine
projects
and
stuff
like
that.
B
That's
cool
yeah.
We
haven't
really
talked
about
that,
so
that'd
be
something
to
go
into
in
more
detail
later
so
well,
that
was
cool.
I
I
shared
one.
You
know
it's
cool
people
are
excited
about
this.
It's
cool
to
see.
You
know
paulo
talking
about
as
a
long-time
java,
jvm
developer,
he's
evaluating
this
and
and
excited
about
the
future
of
dot
net.
So
there's.
A
Yeah
and
here
people
see
the
changes
we
made
in
the
platform
and
and
the
improvements
in
it,
we
didn't
even
show
that
you
know
the
our
grpc
implementation
is
one
of
the
fastest
ones
out
there.
I
think
you
know
james
has
tweeted
about
it
a
few
times.
The
only
thing
that
was
faster
than
us.
The
last
time
there
was
a
public
benchmark
mark
run
was
rust,
but
james
thinks
we've
done
enough
to
actually
pass
rust,
maybe
for
the.net
5
rtm.
A
It
will
not
we
one
of
the
one
of
the
core
tenants
of
of
of
dotnet.
You
know,
dotnet
five
is
primarily
an
evolutionary.net
core.
It's
it's.
You
could
almost
call
it.net
core
five,
but
we
mentioned
before.
We
don't
want
to
be
an
operating
system
component.
As
soon
as
it's
a
component
of
the
operating
system,
then
the
compatibility
version
problems
come
to
be
the
servicing
burdens
come
to
be.
You
know,
there
are
no
other
frameworks
that
are
part
of
the
windows
operating
system.
A
B
That
makes
sense
yeah
in
in
all
the
things
that
you
talked
about
in
your
slide
at
the
beginning
of
decoupling,
it
kind
of
giving
you
the
most
flexibility
and
everything.
This
is
a
an
interesting
question
here
on
the
the
naming
of
net
core
like
calling
it
dotnet
five.
You
shared
with
me
early
on
that
that
people
had,
you
know,
got
confused
by
seeing
dot,
dotnet
4.x
and
seeing
that
as
a
bigger
number
than
net
core
3.1
and
thinking
4.x
is
newer
and
better
than
3
dot.
So
yep.
A
I
I
do
think
that
naming
it
done
in
5
will
help
with
some
of
that,
because
people
will
go.
Oh,
it's
the
next
version
of
net.
It's
a
it's
a
little
bit
of
a
falsity,
though,
because
what
we
don't
want
is
people
and
we'll
make
sure
we
message
this
really
good
at
dotnetconf
and
on
the
blogs.
We
did
it
two
years
ago,
but
you
know.net
five
does
not
have
all
the
features
of
net
framework.
For
you
know
it's
still.
There
is
no
wcf.
There
is
no
workflow.
A
There
is
no
system
web,
but
we
thought
the
only
way
to
solve
the
confusion.
Problem
was
to
put
a
bigger
version
number
on
it
and
make
it
clearly
show
that
it's
the
future
platform,
whether
it
has
all
the
apis
or
not
wow,.
B
Yeah
yeah,
that
makes
sense
and
as
a
cto
or
whatever
sometimes
they're,
just
looking
at
the
numbers
and
it's
very
clear
like-
and
that
was
the
name
of
your
talk
at
ignite
was
dotnet5-
is
the
future
right,
so
cool?
Well,
those
are
those
are
amazing.
It
was
great
to
have
you
on
and
and
those
were
some
some
cool
demos
and
and
and
fun
to
look
at
the
kind
of
the
history
of
going
back
to
2013.
A
The
big
thing
for
me
is:
is
try,
try
net
five
preview
right
now
and
because
it's
the
last
window
to
give
us
feedback
on
it
and
shopadonicoff
we'll
talk
about
a
lot
in
we'll
go
in
full
detail
in
depth.
You
know
across
all
the
updates
and
changes
in
dining
five.com.
B
Awesome
all
right
well,
well,
thanks
a
bunch
scott
and
as
as
always
anytime,
you
want
to
pop
back
in
the
show.
We're
happy
to
you
know,
see
anything
we
we
can
from
you
this.
This
was
great.
So
thanks
for
having
me
john
all
right,
we'll
see
you
at
uh.net
conf
and
we'll
see
all
our
our
dot-net
community
stand-up
viewers
we're
on
now
every
tuesday
wednesday,
thursday
at
10
a.m.
Pacific.
So.