►
From YouTube: The Maintainers: Nicholas Blumhardt, Autofac and Serilog
Description
In this episode of The Maintainers, Nicholas Blumhardt talks about his history with Open Source working on Autofac and Serilog.
*** NOTE *** I realized it is misspelled in the video. Sorry about that.
You can follow Nicholas on twitter: https://twitter.com/nblumhardt
A
A
A
They
tend
to
come
back
and
answer
questions,
maybe
more
than
once
or
submit
more
than
one
more
than
one
patch.
So
I've
kind
of
got
a
rule
of
thumb.
Now
on
the
projects
I
work
on
where
once
people
have
made
a
couple
of
contributions
to
the
code
base,
then
they're
automatically
part
of
the
maintaining
kind
of
body
for
that
project.
A
So
cerelog
began
as
a
kitchen
sink
kind
of
a
project
and
originally
to
make
it
a
viable
project.
We
really
had
to
supply
everything
that
somebody
could
need
to
use
it
in
any
environment.
The
core
of
the
idea
that
got
serilog
launched
a
much
more
event-focused
framework
that
deals
with
structured
data
more
than
the
traditional
kind
of
text
format
that
logs
have
been
kind
of
historically
bound
to,
but
turning
that
into
something
that
people
could
use
in
an
enterprise
setting
meant
building
everything
from
the
configuration
system
all
the
way
through
to
different
sinks.
A
Nobody
was
going
to
come
along
and
volunteer
to
do
that
work,
and
so
we
naturally
took
that
into
the
stereolog
organization
to
begin
with,
and
then
fast
forward,
maybe
three
or
four
years,
and
suddenly
we
have
dozens
and
dozens
of
sub
projects
in
all
of
these
different
technology
areas.
It's
very
difficult
to
find
somebody
who
has
the
skills
to
work
on,
say
the
serilog
console
logging
provider
as
well
as
couchdb
as
well
as
mongodb
as
well.
So
we
reached
a
point.
A
A
couple
of
years
ago,
where
we
started
to
just
to
stop,
including
any
new
work
in
the
organization
and
we'd
instead
provide
support
for
people
through
their
github
projects
or
through
stack
overflow
and
try
and
encourage
people
who
needed
a
different
component
for
cerrilogue
to
to
build
that
and
maintain
it
themselves.
That's
been
quite
successful
in
that
a
lot
of
the
the
popular
pieces
of
the
stereolog
ecosystem
now
are
not
under
the
cerelog
organizational
umbrella.
A
A
I
view
the
open
source
that
we're
mostly
doing
in
the
kind
of
nuget.net
world
as
being
for
a
different
set
of
goals,
and
I
don't
know
if
they're
not
always
mutually
exclusive,
but
I
think
that
there's
a
lot
of
joy
in
working
on
like
on
building
your
own
tools
and
I
think
that
that
motivates
a
lot
of
non-ideological
kind
of
open
source.
So,
while
I
really
admire
and
value
a
lot
of
projects
that
are
aiming
for
software
freedom,
I
think
that
day
to
day
the
open
source
that
I
work
on
most
of
us
do.
A
That
really
just
comes
from
that.
That
desire
to
make
our
own
tools
better,
and
I
think
that
libraries
are
kind
of
a
special
case
of
developer
tooling,
where,
if
you
do
invest
in
them
and
you
collaborate
on
them,
then
you
know
the
whole
experience
of
doing
development.
In.Net
becomes
more
enjoyable.
A
So
I
think
I
don't
think
that
those
necessarily
have
to
be
kind
of
competing
goals
for
open
source,
whether
it's
kind
of
free
and
for
the
public
good
or
whether
it's
just
a
matter
of
kind
of
expediency.
I
don't
I
don't.
I
definitely
don't
feel
bad
putting
time
into
things
they're,
not.
I
guess
humanitarian
or
philanthropic
projects
that
way.
A
One
of
the
things
I
think
that
we
all
really
like
about
the
microsoft
ecosystem
is
that
there's
so
much
in
the
box.
This
and
there's
so
much
kind
of
maintain,
supported,
documented
code
like
that,
and
the
only
other
ecosystem
that
I
have
to
kind
of
compare
these
days
is
the
rust
world,
which
I
spend
a
lot
of
time
in
through
my
work.
A
Oh
and
increasingly,
I
think,
through
interests
in
open
source
and
the
the
approach
that
rust
has
taken
has
been
more
to
to
resist,
adding
anything
into
the
standard
kind
of
library
and
force
the
the
kind
of
main
components
of
rust
to
be
delivered
by
the
ecosystem
and
the
community.
I
think
by
doing
that
early
it
really
shaped
people's
expectations,
so
just
like
in
in
node
project.
A
It's
not
really
that
surprising
to
start
a
rust
project
and
install
10
kind
of
community
packages,
and
it's
not
really
bloat
in
the
sense
and
that
you're
not
you're,
not
really
getting
any
more
code
than
if
the
mozilla
or
the
rust
team
at
mozilla
or
rust.
Core
team
otherwise
had
built
those
things,
but
you
kind
of
see
where
they
come
from
a
lot,
and
you
accept,
I
think,
a
lot
more
diversity
and
kind
of
sources
for
code.
A
If
you've
got
an
idea
for
an
open
source
project
and
success
for
your,
I
idea-
I
guess,
is
that
everybody's,
using
your
code,
you're,
probably
going
to
find
yourself
disappointed
in
a
way
because
once
once
something
becomes
popular,
it
invites
kind
of
not
imitation.
So
much
as
I
think
people
absorb
the
ideas
that
are
in
it
and
then
they
find
other
ways
to
express
those
ideas
and
different
projects.
A
So
I
I
learned,
through
the
project
through
the
auto
fact,
project
a
lot
of
the
core
ideas
of
auto
fact
that
were
kind
of
even
controversial
at
the
time
that
the
container
should
always
hang
on
to
resources
to
dispose
them
the
idea
of
scoped
lifetimes
using
lambdas
to
activate
components.
Those
things
are
so
run-of-the-million.net
ioc
containers
now
that
I
think
you'd
be
hard-pressed
to
find
anyone
who
could
pinpoint
those
back
to
autofac
in
the
early
days
and
yet
autofact
feels
really
successful
to
me,
because
those
ideas
became
prevalent
and
serilog
as
well.
A
I
think
the
fun
the
fundamental
goals
of
the
early
project
are
very
much
achieved.
Now.
That.Net
has
got
a
message:
template
based
structured
logging
implementation.
End
log
has
also
implemented
the
the
same
style
of
structured
logging,
which
came
along
a
couple
of
years
ago.
If
you
really
love
the
idea
and
it's
worth
implementing
it,
so
others
can
kind
of
learn
that
idea.
A
Open
source
is
great,
I
think
any
other
kind
of
goal
where
you
really
want
the
world
to
use.
Your
code
is
just
destined
for
frustration,
because
you
don't
realize
so
much
yourself
that
even
say
cereal.
So
much
so
much
of
the
the
intellectual
kind
of
shape
of
cereal
comes
from
log
for
net,
which
comes
from
log
for
j,
which
comes
from
probably
something
that
somebody
spent
nights
and
weekends
on
implementing
in
fortran
in
1978.
A
So
I
had
the
experience
of
taking
auto
fact
all
the
way,
through
from
being
a
completely
new
project
to
kind
of
being
totally
mature
and
me
disengaging
from
it
the
eventual
kind
of
payoff.
I
suppose,
if
you
look
at
how
I
could
find
the
time
to
work
on
on
auto
fact
for
free
was
that
sometime
around
about
the
well.
I
guess
it
became
the
end
rapidly,
but
once
once
auto
fact
had
become
popular
that
opened
the
door
to
go
and
work
on
some
other
bcl
components
on
the.net
team
and
microsoft.
A
So
I
moved
on
from
auto
fac
around
about
the
time
that
I
moved
to
the
states
and
and
worked
for
microsoft
there.
But
you
know,
I
think
I
considered
open
source
sort
of
self-justifying
from
that
experience,
stereolog's
a
little
different.
I
think
that,
initially
from
the
very
beginning
as
well,
I
was
using
serilog
in
my
day-to-day
work
as
a
consultant
being
able
to
just
kind
of
show
people
the
insides
of
their
running
systems
without
all
that
kind
of
extra
extra
work,
that
comes
with,
say,
event,
sourcing
or
other
forms
of
like
event-based
instrumentation.
A
That
was
the
payoff.
I
think
that
made
building
serilog
in
the
beginning,
sustainable
for
me
and
now
stereolog
is,
is
in
part,
I
think,
maintained
by
companies
who
are
working
in
the
the
same
space
where,
for
example,
the
the
primary
maintainer
of
the
splunk
sink
for
stereo
log
works
for
splunk
and
the
primary
maintainers.
I
think
of
the
elastic
sink
actually
a
couple
of
the
maintainers
they're
from
the
community,
but
we
have
people
who
work
for
elastic
who
work
in
the
elastic
sink.
A
I
started
another
project,
maybe
three
years
after
beginning
stereolog,
that
was
end
for
consuming
and
viewing
logs,
and
so
that,
as
it
grew
into
a
into
a
product
that
I
work
on
day
to
day
called
seek,
has
provided
a
lot
of
time
for
me
to
work
on
serilog
during
the
day
as
well.
So
I
think
long
term
having
a
healthy
kind
of
open
source
project
does
at
least
in.net
I
think-
requires
companies
to
kind
of
take
their
part
and
and
move
that
forward.
I
think
it's
worked
well
for
stereolog
for,
for
the
most.