►
From YouTube: IPFS Weekly Call 2019-07-22
Description
Presentation: wab.ca and wabac.js
Speaker: Ilya Kreymer - https://github.com/ikreymer
Notes
* Ilya runs webrecorder.io
* Hosted web service that records traffic for archiving as the user browsers around
* Part of a grant funded non-profit called Rhizome
* Wab.ac is completely client side player for WARC files using service workers
* Any WARC file can be played back completely in the browser
* Can play wab.ac inside Internet Archive Wayback machine
* Also support .har files (built into the browser) - create in Chrome devtools
Questions
* Who is funding? Webrecorder.io grant from Mellon Foundation - files hosted on WebRecorder.io (2 year grant)
* Data can be stored anywhere and rendered completely in the browser
* Challenges public vs. private? Webrecorder.io is private by default. Possible for somebody to archive private social media
* Client side archiving is not possible yet in a browser without an extension
* Ibnesayeed: WARCreate is another tool for creating WARC files … missing piece is discovery. Opportunity to integrate IPFS for discovery. - Think it would be ideal to integrate with a separate discovery layer. Integration with js-ipfs would make a lot of sense.
* Ibneysayeed was in D.C. at the Web Packaging workshop … pages could run as the original domain.
* All the position papers from the ESCAPE workshop: https://www.iab.org/activities/workshops/escape-workshop/
* Signed exchanges are good for verification and trust
* Marcin: If anyone wants to follow, I am gathering notes about different ways of Mirroring Web to IPFS: https://github.com/ipfs/in-web-browsers/issues/94 (started as ideas for our browser extension)
A
B
B
So
the
first
thing
I'll
be
showing
is,
is
a
recorder,
so
I'll
just
serve
and
clear
record
at
I/o,
and
the
idea
with
my
recorder
is
that
basically
a
tool
for
anyone
to
use
to
archive
browser
traffic
as
you're
using
it,
and
so
that
basic
idea
is
that
anything
that
you
enter
into
the
browser
is
captured.
And
then
you
could
you
play
it.
So,
for
example,
just
something
interesting
like,
for
example,
I
work
for
a
nonprofit
called
rhizome
and
could
archive
rhizomes
Instagram
Edith,
it's
behind
for
the
URL
and.
B
Hopefully,
my
network
is
decent
enough
right
now,
so
it's
loading
this
Instagram
page,
there's
a
size
counter
here
and
that
is
being
recorded
and
so
the
way
that
this
works
is
that
it's
basically
a
proxy.
So
there's
a
it's
going
through
a
hosted
service,
that's
capturing
network
traffic
and
then
serving
it
back
and
as
it's
doing
guy,
that's
for
writing
it
to
work
files
and
so
I.
Essentially
what
it's
doing
is
it's
recording
exactly
what
I?
B
What
I
have
you
here
so
as
I
kind
of
rounds
around
you
can
see
the
size
counter
here
doing
up
and
I
can
stop
at
any
time,
and
so
this
is
kind
of
a
a
user
driven
archiving
tool.
However,
at
this
point
this
this
is
using
a
hosted
service.
So
it's
someone
centralized
right
now,
but
it's
all
open
source,
and
so
anyone
can
run
it
on
their
own,
so
I
got
I
can,
for
example,
I
think,
there's
a
video
here,
so
I'll
click
on
this
and
play
this
video
and
then
so.
B
I'll,
stop
this
now
and
then
amusing
and
anonymously,
and
then
I
can
also
click
on
this
and
now
I'm
in
the
leaflet
mode,
and
so
now
it's
playing
back
what
I've
recorded
now,
what
I
mean
here
it
says
browsing
and
what
it
has
captured
here
is
is
essentially
what
I've
browsed
so
so
so
far.
So,
for
example,
I
think
I
should
be
able
to
click
on
this
video
and
played
back
again,
but
I
won't
be
able
to
browse
things
that
I
haven't
yet
brows.
B
So
at
some
point
it
will
stop
so,
for
example,
I
haven't
scrolled
all
the
way
down,
and
so
this
is
also
still
using
the
web
record
IO
service,
and
this
is
a
project
I've
been
working
on
for
a
while
and
it's
part
of
a
part
of
a
nonprofit
called
rhizome
and
we're
in
London.
What
I
also
wanted
to
show
today
is
sort
of
a
really
new
idea
that
is
now
possible.
B
I
think
is
worth
cool
and
I
shared
a
vacuum,
and
that
is
so
essentially
the
way
this
works
is
kind
of
like
the
Wayback
Machine,
it's
playing
back
URLs
that
are
served
from
from
this
service,
but
it's
also
possible
to
actually
present
web
archives
entirely
in
the
browser,
and
so
do
that
I'll
go
ahead
and
download
this.
Whatever
chord
now
creates
file
format
called
wark,
I
was
just
a
standard
web
archiving
format
and.
B
Think
I
should
even
be
able
to
play
this
video
that's
been
captured
here,
and
you
know
this
is
all
this
is
basically
just
a
static.
If
I
ship,
refresh
without
service
workers,
you'll
see
that
it's
just
a
runny
one
github
pages
live
I,
do
it
a
normal
refresh,
it
will
wave
back
in
the
browser
and
so
right
now.
This
isn't
yet
integrated
with
the
idea.
That's
anyway,
but
I
think
that
this
provides
a
potential
to
serve
high
fidelity
web
archived
content
from
IP,
invest,
buying
doing
the
web
archive
replay
entirely
client-side
and
yeah.
B
What
it
do
is
that
it
can
so
in
turn
our
kite,
for
example,
it's
actually
possible
since
it's
just
a
static
page,
it's
possible
to
archive
this,
this
page
itself
into
the
internet,
wayback
machine
and
doing
that
it's
possible
to
provide
an
alternative
replay
mechanism
to
what
interact
I
provide.
So,
for
example,
this
is
this
site,
archived
in
the
internal
comms
way
back
machine,
and
it's
actually
kind
of
running
as
a
serviceworker
inside
of
there
and
with
that
I
can
actually
play
back.
B
So,
for
example,
I
can
play
back
a
version
of
cnn.com
from
an
archives
wayback
machine
using
this
system
inside
inside
of
their
way
playback
machine.
So
this
is
kind
of
a
way
back
to
see
an
inside
of
the
wayback
machine,
if
you
will,
which
I
think
the
reason
I
do
want
to
do,
that
is
to
provide
improved,
improve
replay
fidelity.
B
So
there's
I
that
I
think
probably
a
lot
of
different
applications
and
also
the
system
supports
hard
files
which
are
sort
of
built
into
the
browser.
So
I
want
to
show
another
so
before
I
use
web
report
that
il,
but
actually
you
don't
even
have
to
use
liver
for
that
I
owe
to
create
the
web
archive.
You
can
actually
also
use
hard
files,
although
they're
not
as
server
example
I
can
just
using
dev
tools.
I
can
kind
of
do
the
same
thing.
B
Like
this,
and
then
what
I
can
do,
it
is
when
this
isn't
problem
of
this
is
that's
not
very
user
friendly
I
can
right-click
and
I
do
can
do
save
all
his
heart
content
and
I
can
save
this
har
file
and
Daniel
I
should
be
able
to
collect
C
and
I
should
also
be
able
to
open
this
har
file
as
well
far
format
is,
is
a
JSON
based
format.
So
it's
kind
of
it's.
It's
not
very
ideal
for
sleep,
a.
B
B
Well,
maybe
maybe
the
heart
is
in
plate.
Well,
one
thing's,
maybe
maybe
further
this
the
heart.
Doesn't
it
doesn't
always
work
at
the
heart
format?
Unfortunately,
it
would
be
nice
if
browsers
provided
a
way
to
choose
save
dynamic
sites
directly,
but
we
don't
have
that
and
that's
why
we
feel
that
reporter
for
the,
but
at
least
with
a
replay
side,
we
can
so
I'll
just
go
back
milking
again
and
another
example
is
that
God
is
here.
Oh.
B
So,
for
example,
I
can
open
this
other
work
file,
and
here
I
can
should
be
able
to
view
this
Twitter
when
I
replay
also
entirely
rendered
in
the
browser
and
yeah.
So
this
is
basically
rendering
that
here
with
C
and
one
final
example
is
with
the
current
setup,
it's
actually
possible
to
to,
for
example,
link
to
a
work
file.
That's
stored
somewhere
and
have
way
back
to
G
is
open
it
directly
and
so.
B
I'll,
just
up
sharing
my
screen
from
out
and
yeah,
so,
for
example,
this
link
here,
actually,
if
you
load
that
it
will
load
this
work
file
locally
for
you
and
then
render
that
redirect
to
that
URL,
and
this
should
work
in
those
browsers
and
then
you'll
have
a
leaf
layafette
Twitter
page
rendered
locally,
and
so
the
idea
is
that
right
now
is
just
loading
from
from
the
web.
But
of
course
it
can
also
load
from
idea
thing
us
and
from
from
any
other
data
source
and
just
a
quick,
well
dude
create
a
quick
slide.
B
I
forgot
about
that.
So
to
summarize
basically
yeah.
This
is
a
surrealist
web
archive
we
play
designed
to
work
with
working
hard
files
or
really
get
to
be
extended
to
do
anything.
There's
no
integration
with
our
TFS
yet
and
I
haven't
done
anything
with
how
to
store
more
files
in
IVFs.
But
my
colleague
Southwood
has
done
some
work
with
that
with
the
ID
WB
system,
and
so
probably
that
format
could
be
used
to
store
work,
data
and
kind
of
expand
on
that
Wow,
yeah
and.
C
D
I'm
curious,
you
mentioned
this
is
kind
of
centralized
today,
so
who
is
backing
up
all
of
all
of
this
content?
Who
is
kind
of
funding
the
the
central
infrastructure
to
to
keep
all
of
these
archive
files
around?
Is
it
really
just
people
downloading
it
on
their
own
machines,
or
is
there
kind
of
a
coordinated
effort
going
on
here.
B
And
of
course,
that's
not
necessarily
the
way
to
go
so
I'm
kind
of
exploring
ways
that
that
we
can
not
be
relying
on
it
on
Amazon
for
the
web
record
that
is
or
this.
But
at
the
moment
that's
where
that's,
where
it's
running
and
we're
using
the
two-year
grant
from
the
mellon
Foundation
to
to
run
web
reporter.
But
the
new
thing
that
I
showed
with
the
browser-based
replay
that
essentially
allows
for
data
to
be
stored
anywhere
and
render
the
entire
Linda
browser.
B
C
E
I
see
on
the
site,
you
support,
you
know
I'd
like
to
focus
on
being
logged
in
so
that's
always
been
a
challenge
for
archivists.
Have
you
had
any
challenges
and
problems
around
public
versus
private
content
or
people
accidentally
archiving
and
exposing
private
content?
Is
there
a
takedown
process
or
there.
B
Is
and
we
haven't
really
had
had
specific
takedown
requests,
I
think
most
users
choose
to
so,
unlike,
for
example,
in
your
archive
that
their
goal
is
to
kind
of
a
kind
of
a
public
archive
of
the
web.
With
my
reporter,
it's
sort
of
aimed
at
users
making
their
own
personal
archives
or
content
that
is
important
to
them,
and
so
the
data
is
private
by
default
web
before
it
does
allow
users
to
log
into
to
whatever
sites
that
they
want
to
archive,
and
that's
sort
of
part
of
that.
B
One
of
the
features
is
that
you
can
log
in
to
two
sites,
but
of
course
there
is
kind
of
a
so,
for
example,
though
they're
all
work,
Dana's,
not
isn't
that
main
public,
but
the
except
for
the
owner
of
it
of
the
site,
so
does
the
owner
of
the
account,
and
so
the
yeah.
It's
essentially
whatever
the
users
choose
to
me.
B
Public
I
think
they
can
archive
their
private
data,
and
so
that
is
definitely
something
I
thought
about
what
happens
if,
if
someone
archives
that
private
social
media
feed
of
and
chooses
to
make
the
public
we
haven't
luckily
had
that
happen,
but
that
was
something
that
I
thought
about
and
have
a
process
for
that.
If
that
does
happen,.
A
So
yeah
I'm
really
excited
about
the
potential
here
for
sort
of
like
peer-to-peer
archiving
so
circles,
the
friends
capturing
the
websites
they
want-
and
this
is
just
really
useful
for,
like
I,
tried
out
the
system
just
before
I
hop
on
the
plane
flying
back
from
Barcelona
and
I
archived
a
few
sites
I
was
able
to.
You
know,
look
at
them
on
the
flame.
B
And
yet
the
because,
of
course,
how
how
browsers
work
to
actually
do
the
archiving?
It
is
not
really
possible
entirely
with
Service
Worker.
So
it
would
be
great
if
it
was
like.
You
really
need
an
extension,
probably
or
Weber
for
that
iOS
enjoy
you
to
do
that.
To
do
the
entirely
client-side
would
require
an
extension,
but
luckily,
with
the
with
the
replay
I
think
that
the
Service
Worker
approached
kind
of
changes,
the
access
part.
C
So
laboratoire
create
Matt
Kelley's
tool
is,
is
there
even
when
you
are,
and
you
don't
have
access
to
something
like
the
recorder?
It's
not
as
robust
as
well
recorder.
Obviously,
but
it
does
what
har
files
do,
except
it
stores
things
in
what
format
you
will
have
to
install
a
browser
extension.
Obviously,
I
think
I
think
what
Jim
was
suggesting
is
we're
like
a
group
of
friends,
archiving
stuff
and
then
collaborating
and
sharing
the
the
missing
piece
here
is
discovery.
So,
yes,
I
can
job
sites
on
my
machine.
C
B
C
C
C
So
it's
like
what
you
do
in
your
replay
system,
like
the
old
web
today,
where
internal
browser
sees
the
web
past
web
as
if
it
is
interacting
with
those
domains
directly
right,
because
you
have
a
proxy
layer,
it
kind
of
eliminates
the
need
of
that
proxy,
which
is
cool.
But
there
are
some
situations
where
there
are
issues
in
the
current
is
back,
and
so
I
was
talking
to
folks
at
Hoover
on
that
and
we
sorted
a
few
things
out,
and
hopefully
there
will
be
some
changes
accordingly.
C
B
The
other
part
of
this,
of
course,
is
an
interrogation
and
Trust
I,
think
which,
essentially
with
this
you
know,
anyone
could
create
a
work
file
and
and
render
it
and
say
that
you
know
this
appeared
on
Twitter
of
this
particular
date,
and
so
I
think
that
yeah,
that
that
becomes
even
more
important
to
have
that
I,
don't
know
if
we
were
to
support
design
exchanges,
for
example,
I,
don't
know
if
they
work
or
if
there
is
some
way
to
do
with
client-side.
To
have
that
kind
of
verification.
I
think
this
it
will
become
increased.
C
Were
we
were
more
stuck
on
the
possibility
of
trying
to
extend
the
trust
of
those
signed
exchanges
signed
exchanges
are
fine,
they
will
work
for
us
for
a
few
hours
few
days,
but
we
are
talking
about
not
decades
but
centuries
to
come.
How
long
are
they
going
to
last
and
the
trust
will
remain
like
yeah?
This
is
indeed
came,
so
we
discussed
some
ideas
around
that
to
keep
their
boundary
extending
in
you
know,
while
the
technology
walls,
you
know
the
similar
infrastructure
changes
and
things
come
and
go,
how
can
we
make
it
last
longer?