►
Description
In this lecture, Valentin will talk to us about a long standing problem within the Nix community, documentation. Valentin will cover how we are flattening the learning curve, and how you can help.
Special thanks to the NLnet foundation, the European Commission, the NixOS Foundation, and Tweag for making this event a reality!
The continued discussion for the lecture series is happening over here:
https://matrix.to/#/#son2022-lectures:matrix.org
More information about the Summer of Nix can be found on the website:
https://summer.nixos.org
A
Yes,
it
should
okay
cool
so
welcome
again
to
another
lecture
of
the
summer
of
next
2022
public
lecture
series.
Today,
we'll
have
valentin
talk
about
flattening
the
learning
curve
and
I
think
of
some
especially
some
experienced
sneakers
and
people
that
go
a
little
bit
back
into
history
for
next
documentation
and
the
learning
curve.
A
If
you
want
to
ask
your
questions
directly
into
the
matrix
channel,
that's
also
fine,
because
after
the
lecture,
valentine
will
still
try
and
see.
If
you
can
answer
some
of
those
questions
on
matrix,
so
know
that
the
questions
this
time
will
be
answered
on
matrix,
just
as
with
our
main
stock
to
keep
the
flow.
However,
we
have
included
a
very
small
portion
of
q,
a
from
some
internal
colleagues
and
yes
without
further
ado
I'll
I'll
start.
A
A
B
Yeah
exactly
so,
my
name
is
valentin.
I
work
for
tweak.
B
I
joined
tweak
beginning
of
this
year
and
the
question
was:
what
could
I
do
as
long
as
I
don't
have
a
client
project
that
I
can
work
on
and
tweak
has
this
internal
mix
team,
where
we
develop
the
product
and
help
onboard
our
own
co-workers
working
with
next,
because
it's
a
daily,
stable
tool
in
the
company
and
well,
I
joined
tweak
as
a
mixer,
because
I,
when
I
discovered
the
project
that's
a
couple
of
years
ago,
I
finally
found
something
I
was
looking
for
for
a
long
time,
a
tool
to
make
computers
work
reliably
in
the
long
run
and
if
you
know
nyx
once
you
have
built
something,
it
will
work
reliably
until
you
change
it
and
not
change.
B
A
B
You
would
work
with
multiple
machines
and
well
the
learning
curve.
Anyone
who
has
tried
it.
It
really
is
proverbially
steep.
It
takes
a
lot
of
time
to
really
truly
grasp
the
thing
well
enough
to
be
confident
using
it.
So
I
probably
took
like
two
years
before
I
was
really
confident
packaging,
a
new
software
that
was
not
yet
in
mixed
packages,
so.
B
Well
and
when
I
observed
new
colleagues
joining
tweak,
who
did
not
know
nicks
at
all,
I
quickly
found
that
we
have
to
do
something
about
it
and
I
wondered
what
would
that
be
and
luckily-
or
I'm
really
happy
that
that
tweak
leadership
allowed
me
to
spend
some
time
on
this
doing
it
properly.
That's
the
other
thing
that
I
also
observed
when
I
was
spending
that
time
on
my
own
and
my
free
time.
B
It
simply
turns
out
the
problem
is
usually
too
big
to
attack
it
on
your
own,
like
if
you
spend
a
couple
of
hours
here
and
there
you
may
maybe
make
notes
on
your
on
your
own
or
maybe
fix
a
little
typo
here
and
there,
but
really
changing
anything
about
the
big
state
of
affairs.
The
big
picture
is
essentially
impossible:
the
entire
documentation
ecosystem
and
how
things
are
interconnected
and
how
they're
presented
it
feels
a
lot
like
a
jungle
I
would
say,
and
so
you
would
yeah
so
you
would.
B
You
essentially
need
to
get
everything
of
that
in
your
head
and
then
navigate
through
it
more
or
less
and
well
so
so
I
started
essentially.
So
my
presentation
is
a
little
bit
ad
hoc
here,
so.
A
This
is
my:
I
will
actually
pull
in
your
documentation,
so
yeah
valentin
will
be
talking
about
the
documentation
of
nexus.
If
that
isn't
obvious
just
yet,
but
I
look
forward
to
watching
it
valentin.
A
B
Yes,
please
interrupt
me
at
any
time
because
well,
it's
just
like
a
little
chit
chat
about
what's
been
going
on
in
the
past
month,.
B
Okay,
so
essentially,
this
is
just
my
internal
project
pitch
a
little
bit
made
up
for
public
viewing.
This
is
also
the
same
thing
that
I
submitted
to
the
summer
of
nix,
because
it's
all
the
same
concern
like
mix
is
really
hard
to
get
into,
and
maybe,
if
you
have
been
around
2019
or
so
you
have
seen
this
little
post
on
discourse
which
has
adapted
has
been
adapted
to
make
s,
so
I
started
figuring
out
what
is
actually
the
problem.
B
Why
is
it
so
hard
to
learn
to
use
nyx,
and
I
started
with
a
little
survey
on
my
own
intuition,
so
one
one
thing
that
is
very
very
prolific
is
ian
henry's
what
he
calls
a
let's
play,
but
for
obscure
software
documentation.
I
think
it
already
says
it
all
if
you
read
it
as
like
40
parts
or
something
it's
essentially
mocking
how
complicated
nyx
is
set
up
and
how
confusing
the
documentation
is
often
written,
and
I
actually
talked
to
him
and
asked
him
well.
Why
don't
we
change
something
about
it?
B
That
was
a
year
ago
or
so,
and
the
thing
is:
if
you
went
through
all
these
notes,
it
would
result
in
like
hundreds
and
hundreds
of
issues
or
pull
requests
against
the
documentation.
As
I
said,
this
is
nothing
a
regular
person
can
do
on
their
own
or
even
in
the
spare
time.
So
what
would
be
the
most
valuable
thing,
the
most
important
thing
that
we
could
attack
immediately,
then
I
started
just
doing
an
overview,
so
what
kind
of
data
do
we
have?
B
We
had
this
nyx
survey
in
the
beginning
of
the
year,
which
essentially
told
us
that
a
lot
of
people
use
nixos,
roughly
half
of
the
survey
participants
and
documentation
is
one
of
the
most
impressive
problems
that
we
have.
I
think
ron
summarize
it
as
documentation,
documentation,
documentation.
B
I
also
went
through
these
survey
results
because
there's
a
lot
of
them
and
helped
kind
of
sorting
out
all
the
all
the
numbers
just
to
get
an
idea
of.
Where
do
people
get
stuck
most,
but
that's
still
not
enough
because
well,
if
you,
if
you
say
well,
flakes,
are
really
complicated,
understand
how
they,
I
don't
understand,
how
they
work
or
if
people
say
well,
I
want
to
use
nixos,
but
it's
really
hard
to
get
into,
because
the
nick
language
is
hard
to
learn
that
doesn't
really
help.
B
So
I
started
doing
user
studies
and
observed
new
comers
to
tweak
and
some
friends
who
haven't
seen
seen
knicks
yet
to
see
where
they
stumble
exactly
and
at
the
same
time
I
had
to
know
what
kind
of
resources
are
they
even
in
the
ecosystem.
So
I
did
a
little
survey
of
resources
on
my
own
and
wondered
okay.
What
do
we
have?
B
There's
one
thing
called
the
documentation
system,
which
kind
of
categorizes
the
types
of
documents
from
which
people
can
learn
how
something
works,
and
I
try
to
just
figure
out
what
kind
of
categories
do
we
have
in
the
ecosystem
right
now
and
did
a
little
survey?
It's
not
entirely
complete
and
meanwhile
I
have
found
a
little
bit
more,
but
mostly
we
have
a
lot
of
tutorials
for
the
nyx
language.
B
B
We
have
a
bunch
of
how-to
guides
to
solve
actual
problems,
mostly
those
from
nick's
death
and
john
ringer,
who
has
been
doing
the
talk
last
week
also
has
some
great
introductions
and
also
how
to
use
nixo
s,
which
is
mostly
in
relation
to
deploying
stuff
and
doing
some
configurations,
but
then
the
two
other
interesting
things
is
what
are
the
what's
the
reference
material
like?
How
do
we
like,
where
are
all
the
important
aspects
of
nics,
listed
down
to
every
detail,
and
this
is
fairly
clearly
structured?
B
But
if
we
dive
deeper,
it's
actually
both
instructions
and
tutorials
and
a
little
bit
of
explanation
here
and
there
it's
not
really
clear
cut,
but
why
nyx
is
that
way
and
how
it
really
works
internally
is
only
really
explained
in
elko's
page
thesis,
and
this
kind
of
make
me
made
me
wonder,
is
may
that
be
the
reason
why
people
struggle
understanding
nick
so
much
because
that
was
I
mean
that
is
anecdotal
still,
because
when
I
started
using
mix
the
first
thing
that
I've
actually
read
was
the
thesis
and
by
some
incident
exactly
in
that
same
year,
disappeared
from
the
front
page
of
or
from
one
of
the
pages
of
the
nixa's
website.
B
And
since
then,
I've
constantly
had
people
complain
like
nix
is
too
hard
to
understand
it's
hard
to
learn.
I
don't
know
how
it
works.
I
only
copy
paste
code
and
well
not
all
of
the
people
are
crazy
enough
to
actually
read
a
200
page
via
phd
thesis,
but
the
most
important
thing
in
this
thesis
is
really
how
mix
works
in
principle,
and
that
is
just
that.
The
file
system
in
nyx
is
considered
to
be
used
like
the
memory
system
in
a
programming
language.
B
If
we
stop
thinking
in
terms
of
files
which
have
changing
contents,
which
may
just
not
be
there
or
have
been
have
be
something
else
than
we
expected
and
just
pretend
that
the
dependencies
that
we
have
for
building
our
software
work
just
like
pointers
in
a
computer
program,
if
we
nail
them
down
more
properly,
instead
of
just
changing
and
writing
onto
them
and
moving
them
around
freely.
Without
all
that
much
discipline,
then
we
would
solve
a
ton
of
problems
that
we
had
in
the
past.
B
I
don't
know
40
years
of
open
source
software,
so
I
kind
of
set
myself
on
a
well
tefan
called
a
crusade
to
help
people
learn
mix
in
a
more
principled
way
and
okay.
I
said
I
already
started
doing
these
surveys
and,
while
doing
these
surveys
I
found
out
and
also
talking
to
people
a
lot
who
had
a
lot
more
experience
than
mix
than
I
did.
I
found
out
that.
B
For
example,
it
doesn't
really
matter
to
people
all
that
much
how
how
nicks
works
internally,
whether
it's
a
functional
program,
because
most
people
who
are
interested
in
using
x
don't
really
care
all
that
much
about
functional
programming,
they're,
mostly
practical
developers,
and
also
they
don't
have
time
to
go
through
all
these
concepts
that
I
try
to
lay
out
when
documenting
the
underlying
architecture,
like
it's
a
ton
of
stuff
that
you
have
to
read
and
understand,
and
even
if
it's
sorted
out
a
little
bit
better,
it's
still
a
jungle,
it's
still
complicated
because
it's
a
well
it's
a
system
that
essentially
tries
to
reinvent
how
we
would
build
and
deploy
software.
B
So
what
people
really
care
about
is
well.
How
do
I
get
things
to
run
and
that's
that
usually
amounts
to
just
typing
nick's
shell
dash
p
and
then
finally
right
package
name.
So
the
question
was:
how
would
we
approach
how?
How
would
we
approach
people
who
want
to
get
things
done
with
nics
and
still
help
them
understand
it
well
enough
such
that
they
can
actually
make
use
of
it
on
their
own
at
some
point
and
navigate
the
jungle
of
documentation
that
we
have,
and
this
is
when
things
kind
of
turned
around.
B
So
my
first
idea
was
to
well.
Why
don't
we
just
write
a
book
like
the
next
book,
which
will
explain
how
nyx
works
fundamentally
and
then
show
how
you
would
use
these
fundamental
concepts
to
build
up
something
like
a
package
but
in
a
more
concise
way,
in
a
more
user-oriented
way
than
we
had,
for
example,
with
nyx
pills,
which
people
read
as
one
of
their
first
documents
when
they
get
introduced
into
nyx
and
they'd
get
totally
confused
about
the
details,
because
they
don't
usually
matter.
B
So.
I
wondered
what
if
we
did
like
a
principled
introduction
to
nyx
and
exercises
and
questionnaires
and
test
it
with
people
who
really
try
to
use
nics
and
when
it's
actually
started
trying
to
test
it
with
people
who
want
to
use
nyx.
It
turns
out
they
don't
care
all
that
much
and
when
I
did
my
survey
of
existing
material
of
course
stumbled
over
nix.dev.
B
I
don't
I
mean
I
didn't
stumble
over
it,
it's
one
of
the
most
prominent
resources
that
we
have,
because
it
has
such
well-designed,
well-written,
really
concise,
to-the-point
tutorials,
that
I
thought
these
should
actually
be
the
core
of
whatever
that
book
would
eventually
become,
and
the
other
thing
was
that
we
have
this
large
collection
of
tutorials
that
are
compiled
in
the
nexus
wiki,
and
both
resources
are
actually
unofficial
right.
B
So
domin
has
this
next.dev
and
had
is
going
for
many
years
now,
and
the
nexus
vk
as
well
is
at
least
I
don't
know
five
years
old,
and
none
of
these
reasons
are
really
official
and
the
question
was:
how
would
we
bring
this
combined
knowledge
together
to
to
make
most
use
of
it?
B
It's
essentially
a
little
bit
like
the
same
thing
as
mixed
packages
where
mixed
packages,
as
john
said,
is
this
huge
collection
of
of
organizational
and
procedural
knowledge
of
how
to
build
packages
and
how
to
organize
all
these
things
together,
and
what
we
probably
need
is
something
like
this:
just
for
documentation,
where
we
combine
all
the
insights
into
how
how
people
actually
use
nyx
and
how
they
what
they
need
to
make
most
effective
use
of
it
and
how
to
approach
it
most
effectively,
and
this
is
why
I
thought:
well,
let's
just
bring
all
of
these
people
together
and
make
something
like
a
knicks
documentation
team.
B
This
is
the
nexus
wiki.
What
would
you
call
these
people?
Someone?
Somebody
gave
me
a
link
to
this
thing
where
they
call
themselves
documentarians
and
then
I
said
well,
let's
just
become
mixed.
Documentarians.
B
What
would
that
entail?
We
would
bring
together
all
the
information
that
we
have
in
the
ecosystem.
All
the
people
who
write
the
interesting
articles
and
the
interesting
tutorials
and
the
reference
material
sit
down
together
and
figure
out
what
would
a
new
way
of
learning
nics?
What
would
a
path
on
this
journey
into
mastering
the
next
ecosystem
look
like,
so
I
did
a
bunch
of
drafts
and
proposals.
What
that
could
be
one.
One
of
them
would
be
okay,
so
assuming
that
this
is
how
you
land
on
the
front
page
and
see
some
examples.
B
What
we
already
have
is
that
the
website
points
you
to
nickel
dev
tutorials,
but
it
doesn't
really
say
that
they
are
nick.dev,
but
we
should
do
that
more
explicitly
and
then
not
just
start
with
the
most
simple
use
cases
such
as
finding
and
running
packages,
but
also
explain
the
main
features
of
the
ecosystem,
as
it
is
right
now
right
now,
there's
no
real
explanation
of
the
difference
between
imperative
and
declarative
package
management
right
now.
B
We
only
have
components
of
these
things
and,
as
I
said,
one
person
cannot
really
change
the
big
picture,
there's
just
too
many
moving
parts,
and
this
is
why
this
documentation
team
that
we
founded
recently
two
months
ago.
B
I
wrote
a
little
announcement,
but
really
the
point
is
easing
nyx
learning
to
the
point
that
we
have
pathways
into
this
jungle
and
pave
more
structured
road
systems,
because
the
jungle
won't
go
away.
The,
for
example,
the
dichotomy
between
imperative
and
declarative
package
mansion.
It
doesn't
go
away
because
that's
an
integral
part
of
nix.
What
also
won't
go
away
is
the
old
cli
and
the
new
cli,
the
non-flakes
world
and
the
flakes
world,
but
we
have
to
help
people
navigate
these
complexities.
B
One
of
the
distinguishing
features
of
nyx
is
strong
backwards,
compatibility
simply
by
the
virtue
of
being
able
to
reproduce
age-old
configurations,
so
people
will
have
to
navigate
the
options
that
nyx
has
offered
during
its
course
of
history,
and
this
is
what
essentially
I'm
working
right
now,
together
with
people
from
summer
of
nix
and
domin
and
mike
92
from
nixon's
weekend.
Everyone
who's
really
interested,
for
example,
in
phoenicial
sylvan
has
also
been
joining
the
documentation
team,
while
working
on
the
next
packages,
architecture,
effort
just
to
figure
out
strategies.
B
And
one
of
the
questions
that
have
been
raised
very
quickly
is
well
okay,
so
I
found
something
out
there.
I
want
to
contribute.
Where
would
I
put
that-
and
it
turns
out
it's
more
so
documentation
is
more
than
just
writing
things
down
how
software
works
or
how
to
use
that
software.
B
It's
really
one
level
higher
is
really.
How
do
we
communicate
in
the
community,
because
so
the
software
that
nyx
is
it's
it's
it's.
We
have
the
next
packagement
and
we
have
this
huge
collection
of
packages
and
mix
packages,
but
then
we
have
also
all
these
peripheral
tools,
such
as
niv
and
home
manager,
which
is
not
an
official
mix
project.
B
B
Deer
ant,
which
is
kind
of
entangled
into
nyx
with
this
next
shell
feature
all
that
software
is
really
useless
if
it's
not
accessible
to
those
people
who
need
it
most.
Those
people
who
cannot
write
it
on
their
own.
So
we
need
to
communicate
to
these
people
outside
of
code,
and
if
we
want
to
help
people
get
into
that
ecosystem,
we
have
to
tell
them
where
to
look
for
it,
and
so
it
turns
out
this
documentation,
effort
kind
of
got
into
a
meta
crusade
in
well.
How
do
we
even
organize
our
community
knowledge?
B
So
what
I
did
as
a
first
draft
in
this
documentation
team
is
just
figuring
out
what
how
how
do
things
work
right
now
and
in
my
impression,
the
discourse
of
the
community
is
essentially
the
central
hub
for
all
the
information
and
a
vision
for
what
we
could
do
in
the
future
is
start
at
discourse,
lead
people
to
discourse
and
from
there
help
them
discover
all
the
in
sorry
when
they
want
to
contribute
help
them
find
places
to
add
their
knowledge
such
that
they
can,
it
can
be
discovered
more
easily.
B
So
so,
really,
if
you
want
to
get
engaged
into
documentation,
what
you
should
first
do
is
engage
in
the
community,
as
most
people
have
done
so
far,
but
my
goal
for
the
future,
maybe
for
the
next
one
or
two
years,
is
to
go
away
from
the
best
way
to
learn.
Nicks
is
to
talk
to
people
like
to
do
sit
together,
and
someone
shows
you
how
to
do
it.
Someone
shows
you.
The
relevant
links
is
that
someone
can
actually
do
it
on
their
own
by
just
reading
the
relevant
material
and
discovering.
B
Are
we
still
working
on
this
because
well
just
gaining
the
overview
is
kind
of
a
full-time
job
and
then
writing
it
down
properly
is
still
even
harder,
as
anyone
will
know,
who's
ever
written
an
academic
paper.
So
the
question
is:
what
would
that
look
like?
One
thing
that
I
would
like
to
do
is
just
draw
more
overviews
like
these.
I
mean
this
is
already
kind
of
a
complex
graph,
but
I
came
up
with
it
in
an
afternoon
just
thinking
about
okay,
what
are
those
components
in
the
next
ecosystem
anyway?
B
B
Then
there's
nif
this
external
tool,
which
helps
pinning
nyx
packages
versions
or
external
things
and
so
forth
and
so
forth,
and
I
hope
that
it
will,
if
we
place
these
things
at
the
right
positions,
and
if
you
break
them
down
well
enough,
not
to
look
as
complicated
as
this
thing
is,
it
will
help
people
to
gain
a
first
impression
of
what
will
expect
them.
B
What
I
also
think
is
that
we
should
kind
of
change
the
way
we
teach
nyx
in
those
tutorials
by
setting
expectations
more
clearly,
for
example,
while
working
on
the
next
language
tutorial,
I
just
noticed.
Well,
yes,
the
next
language
is
a
simple
language,
but,
depending
on
your
previous
knowledge,
it
will
just
take
a
day
or
two
to
grasp
the
fundamentals
and
then,
as
somebody
points
out
learning
the
nyx
language
is
not
just
the
language
itself.
It's
also
learning
the
standard
library,
so
the
built-ins
and
also
the
library
from
nyx
packages.
B
But
then
again
you
have
the
nyx
os
module
system
and
all
the
particular
language
ecosystems
that
all
they
have
their
own
idioms
and
their
own
ways
of
doing
things.
B
So,
just
setting
these
expectations
up
front,
I
hope,
will
help
people
well
plan
the
effort
that
is
required
to
go
on
that
path
and
still
we
need
more
actual
user
studies
where
we
diligently
document
or
observe
neutrally
what
people
really
struggle
with
and
what
are
their
actual
needs,
and
this
is
something
that
just
needs
time,
and
I
hope
that
people
will
come
up
and
help
with
this.
B
So
already,
I've
seen
a
lot
of
contributions
from
the
summer
of
nick's
participants
who
have
been
also
struggling
with
some
of
the
aspects
of
learning
mix
and
what
I
would
like
everyone
to
do
when
they
onboard
into
this
ecosystem
is
to
follow.
Essentially
the
follow
the
boy
scout
rule.
B
But
then
we
have
more
broad
issues
which
cannot
be
done
with
right
by
contributions,
and
I
just
did
a
little
overview
the
other
day
to
to
get
an
impression
of
what
that
may
be.
So
what
we're
still
missing,
for
example,
is
a
non-technical
introduction
of
why
nyx
is
even
relevant
and
how
it
works
in
non-technical
terms.
So
we
have
this
marketing
page
on
on
on
nexus.org,
but
really
it's
it's
still
fairly
technical.
B
We
don't
have
a
learning
roadmap,
as
as
I
try
to
exemplify
here
like
if
you
want
to
tell
someone
what
is
nix,
you
would
have
to
kind
of
guide
them
through
that
thing,
and
we
don't
really
have
that.
Yet
we
currently
have
a
mix
architecture,
a
part
of
a
documentation
for
the
nyx
architecture,
as
I've
shown
in
the
beginning,
I'm
working
on
a
next
language
tutorial
and
then
so
I've
outlined
some
of
these
things.
B
B
Polish,
not
just
the
software
surface,
but
the
path
that
people
take
when
they
learn
about
and
how
to
use
nics,
and
I
hope
that
people
will
onboard
with
helping
out
with
the
little
details,
but
I
also
really
hope
that
things
like
the
recent
developments
and
the
next
foundation
and
the
commercial
interest
that
is
rising
in
the
next
ecosystem
will
help
financing
the
effort
that
is
required
to
do
that.
B
So
what
right
now
is
really
exceptional
that
I
can
take
multiple
weeks
more
than
two
months
of
almost
full-time
work
on
for
doing
this,
and
it
seems
right
now,
I'm
mostly
the
only
one
on
the
planet.
It
seems,
and
I
hope
that
we
can
manage
to
have
a
more
consistent
work
that
is
going
into
documentation
and
helping
people
learn.
We
have
that
on
the
development
side.
In
some
sense
I
mean
elco
is
working
for
determinant
systems.
Teopan
is
working
for
tweak.
B
A
A
So
first
question
that
came
in
is
from
solana.
I
tried
to
convert
a
few
friends
to
next
or
well:
convert
not
convert
and
like
introduce
weather,
they
replied
me.
They
replied
that
they
already
had
the
packages
in
their
linux
distribution
that
they
needed.
So
they
didn't
see
the
point
of
using
nick's
expected.
It
adds
another
layer
of
complexity
and
then
what
kind
of
argument
in
favor
of
next
you
use
in
these
situations
and
perhaps
also
adds
a
little
bit
to
that
question.
Can
we
document
these
advantages.
B
Well,
I
think
these
advantages
have
been
documented
or
told
a
million
times
already,
and
I
don't
really
see
the
point
of
repeating
them
except
for
well
it's
a
thing
that
you
don't
have
to
do
again
once
it
works.
So
that's
I
think
my
main
argument
for
converting
to
nix.
But
the
thing
is
it's
a
trade-off.
If
you
have
a
large
system,
you
will
have
large
migration
costs.
So,
for
example,
if
you
have
a
thousand
packages
that
you
already
manage
in
your
whatever
you
have,
converting
them
to
nyx
will
take
ages
and
it's
natural.
B
So
the
thing
is,
you
will
have
to
to
figure
out
what's
worthwhile
for
you
and
nyx
does
have.
Of
course,
nyx
does
have
migration
costs.
B
B
Just
like
any
other
system
in
the
world,
the
question
is:
how
long
do
you
want
to
maintain
this
system
and
what
next
promise
is
essentially
that
you'll
learn?
Long-Term
maintenance
costs
will
hopefully
go
to
zero.
A
That
makes
sense
then
atella
asked.
Is
there
a
plan
to
create
a
specification
for
the
nyx
model?
Ilko's
thesis
was
written
in
2003
and
not
many
know
what
has
changed
since
since
then,.
B
Yeah,
essentially,
that's
part
of
documenting
nyx
architecture.
What
exactly
the
next
model
is,
I'm
not
sure
about
the
terminology
here
yet,
but
essentially,
I
suppose
it's
the
principle
on
how
on
which
nix
operates,
and
I
think
that
should
be
in
the
next
documentation
itself.
It's
how
nix
works
at
an
abstract
level,
where
it's
a
different
thing
from
the
implementation.
Obviously,
because
that's
source
code.
A
Makes
sense
and
then
another
question
from
attila:
where
can
one
follow
the
progress
of
document
of
the
documentation
team
and
to
see
where
to
help
and
what
to
help
with
this
course,
perhaps.
B
Yes,
absolutely
this
course,
so
we
post
the
documentation,
team
meetings
on
the
discourse,
documentation,
team,
category
or
documentation
category
has
been
has
been
renamed
recently
and
from
there.
Everything
should
be
linked
all
these
things
that
I've
just
shown
can
be
accessed
on
the
internet,
and
if
you
want,
we
can
post
them
to
this
video,
and
that
is
what
I
also
encourage.
B
People
to
do
is,
if
you
want
to,
if
you
want
to
tell
people
things,
link
all
the
relevant
resources
such
that
people
can
follow
them
and
the
more
essentially
the
more
links
we
have
the
easier
it
should
be
to
discover
these
information
in
pieces.
A
Yes,
denseland
asked
when
blocked
with
a
nick's
syntax
question.
It
seems
most
people,
including
myself,
often
find
the
answer
on
github
guests.
It
seems
the
documentation
lacks
real
world
examples
that
are
easy
to
understand
and
absorb.
How
do
you
think
this
could
be
improved
in
the
official
documentation.
B
Good
question
so
yeah
that's
a
problem
that
are
also
perceived,
so
we
have
many
bits
of
insightful
information
dispersed
all
over
the
place.
We
have
parts
and
initials
wiki.
I
just
recently
discovered
a
little
syntax
overview
in
the
nyx
packages,
no
actually
in
the
nixo
s
manual
and
then
the
next
language.
Sorry,
the
mix
manual
with
the
nyx
language
section
also
has
an
entire
specification,
which
is
also
kind
of
incomplete.
B
I
think
we
should
move
all
the
reference
style
documentation
into
the
next
manual
and
either
have
a
kind
of
cheat
sheet
on
the
wiki
or
on
on
the
manual
itself
and
then
have
a
collection
of
cookbooks
or
guides
in
one
central
place.
There
should
be
one
place
right
now:
it's
dispersed
and
that's
one
of
the
efforts
that's
ongoing
and
where
anyone
can
easily
help
is
just
finding
these
instances
and
moving
that
information
to
a
more
relevant
place.
B
B
Exactly
but
you
have
to
take
into
account
that
it's
just
such
a
large
system
that
no
single
person
can
do
it
and
that's
the
whole
value
of
the
community,
that
we
can
do
that
together,
that
anyone
who
stumbles
upon
something
that
that's
missing
or
broken
or
confusing
just
help
out
just
just
change
that
little
bit.
It
will
help
enormously
for
all
the
people
to
come.
A
A
B
A
Maybe
also
I'll
extend
it
a
little
bit.
Is
there
something
that
we
can
help
people
understand
the
difference
between
nix,
the
language
next
to
package
manager,
next
packages,
the
package,
collection,
etc.
B
B
Ideally
in
my
view,
we
should
put
that
on
the
front
page
of
the
website
and
really
explain
that
distinction,
but
that's
a
tough
nut
to
crack,
because
we
have
so
many
things
depending
on
that
front
page,
and
so
that's
still
an
ongoing
discussion.
If
you
follow
the
meeting
notes,
you
will
see
that
we
have
already
talked
about
this,
and
I
want
this
to
change.
I
want
this
to
improve,
but
right
now
I
don't
have
a
final
answer.
The
other
thing:
what
can
we
do
with
the
next
language
itself?
B
B
A
B
Yeah,
it's
absolutely
confusing.
I
want
this
to
change.
As
I
said,
it's
a
hard,
it's
a
tough
nut
to
crack.
So
ideally,
we
would
just
have
nicks.org
as
a
domain,
and
we've
talked
about
this,
but
it's
really
expensive
to
buy,
and
this
is
a
foundation
topic
that
the
foundation
board
has
to
has
to
attack
this.
A
B
Yes,
the
migration
is
unfortunately,
not
entirely
complete,
and
this
is
where
help
is
highly
appreciated,
so
it
would
be.
It
would
be
a
lot
of
help
if
we
could
just
finish
this.
Finally,
yeah
okay.
A
B
Honestly,
I
haven't
been
working
on
that
yet
because
I
think
the
most
important
issue
that
we
have
is,
as
I
try
to
outline,
is
to
get
people
to
understand
how
to
navigate
the
documentation
in
the
first
place,
so
they
can
learn
more
autonomously
but
yeah
sure
if
you
care
about
it
enough,
please
look
into
it.
I
would
be
really
happy
to
help
get
change,
requests
merged.
A
Okay,
I
also
had
a
question
of
myself
and
that
plays
into
the
the
low
hanging
fruit
that
I
just
mentioned.
That
attila
said,
like
the
mark
done
versus
dog
book
thingy,
do
you
think
there's
other
low-hanging
fruit
that
people
can
very
easily
jump
on
to
without
too
much
fear
of
being
judged?
I
guess.
B
Yeah,
there's
a
bunch
so
in
that
little
graph
that
I've
shown
how
to
add
information
and
insights
into
them
into
the
relevant
documents.
It's
a
whole
post
that
I
wrote
for
the
summer
of
nix
participants,
but
which
really
applies
to
anyone
who
wants
to
contribute
to
next.
A
B
A
bunch
of
long-standing
issues
that
we
can
easily
approach
with
without
much
commitment,
so
one
of
them
is
helping
categorize
the
mixers
wiki.
So
we
have
like
50
to
100
orphan
pages,
which
don't
have
a
category
which
are
essentially
not
findable.
Unless
you
know
the
terms
to
google
for
there's
a
few
things
that
you
can
always
do
is.
B
B
If
you
see
something
in
the
vp
that
is
really
polished
and
it
just
works,
it
should
be
in
the
manual
probably
so
if
it's
kind
of
this
is
how
it's
done-
and
this
is
the
interface
that
exists
put
it
into
the
manual,
for
example,
we
have
still
haven't
detected
because
there's
so
much
to
do,
there's
always
much
more
work
than
we
have
time.
There's
no
proper
explanation
of
what
nixos
modules
are
in
the
nixos
manual.
It's
in
the
mix
as
wiki
just
move
that
article
into
the
mixer
s
manual.
A
That
would
make
very
much
sense,
actually
perhaps
more
of
a
personal
question.
I
have
had
to
write
documentation
in
the
past
as
well,
and
I
always
got
a
little
bit
scared
of
publishing
the
documentation.
If
that
makes
sense,
because
I
was
afraid
of
it
not
being
well
worth
in
the
north
or
if
there
were
too
much
spelling
mistakes
or
just
language
or
grammar
mistakes.
Is
this
something
you
think
people
should
fear
or
is
it
just
publish
it?
And
people
will
it'll
evolve
over
time.
B
Yeah,
I
think
language
and
grammar
are
really
the
least
things
to
worry
about,
although
there's
like
two
aspects
to
it,
if
you
make
a
pro
request
and
it's
barely
understandable,
barely
legible,
that
will
probably
not
entice
people
to
review
it,
so
it
should
be
at
some
some
mainly
readable
level,
but
you
can
always
use
online
tools
like
language
tool
to
fix
that,
for
you
or
just
ask
someone
to
to
review
it
first,
ask
your
friends:
do
you
understand
it,
but
other
than
that
small
mistakes
will
always
helpfully
be.
B
There
will
always
be
people
helpful
in
fixing
those,
and
also
what
I
also
observed
a
lot
of
times
is
that
if
people
actually
introduce
their
misconceptions
into
their
pull
requests,
there's
usually
someone
who
will
say
wait
a
minute.
This
is
not
how
it
works.
It's
actually
that
way,
and
then
these
things
will
evolve.
B
Naturally-
and
the
other
thing
is
yes,
please
just
add
your
contributions
as
you
see
fit,
for
example,
it
should
be
super
easy
to
contribute
to
the
nexus
if
you
just
find
something
and
don't
be
afraid
of
it
being
incomplete
or
not
well
worded.
What
you
can
always
do
to
make
sure
is
just
ask
someone
who
doesn't
know
and
observe
them
interacting
with
that
thing.
B
A
That
makes
sense,
and
then
we
had
a
little
talk
before
the
lecture
actually
started,
and
you
had
mentioned
a
book
that
you
didn't
mention
during
your
talk
that
the.
B
B
It's
called
how
learning
works,
and
this
is
what
I'm
mostly
basing
my
documentation
work
on,
because
it
it's
something
like
a
meta
survey
of
the
past
hundred
years
of
learning
science
and
very
concisely
and
aptly
explains
how
people
construct
their
internal
knowledge
graph,
as
I
would
call
it
and
I
highly
recommend
reading
it.
If
you
want
to
try
helping
people
acquire
some
skill
or
knowledge,
because
it's
I
mean
it's
super
helpfully
written.
I
wish
I
could
write
as
focusedly
and
concisely
as
they
do
in
this
book.
A
So
I'd
like
to
thank
you
valentin,
for
these
amazing
talk
and
insights
on
the
nexus
x,
documentation
and
trying
to
flatten
the
learning
curve
of
nyx,
and
then
I'd
also
like
to
thank
our
colleagues
and
the
sidelines
as
solan
and
atila
for
just
pumping
in
the
questions
and
making
them
come
in.
So
yes,
thank
you
again.
Valentin
yeah.
B
B
A
Terribly
sorry
for
that,
so
yeah,
what
I
said
is,
since
valentine
isn't
her
life.
Currently
he
will
answer
questions
eventually
in
the
matrix
channel.
You
can
check
that
over
below
here,
he'll
either
answer
somewhere
this
evening,
or
perhaps
tomorrow,
most
probably
tomorrow,
but
he'll.
He
study
to
try
and
answer
them.