►
From YouTube: CI/CD UX Team Design Review | 14 January 2020
Description
Design Review session of CI/CD UX Team.
- Juan J. Ramirez | 00:31
xUnit Pipeline Screenshots
A
And
then,
just
for
the
record,
I'm
gonna
say
it
again:
I'm
gonna
be
presenting
something
that
I
already
presented,
but
this
time
is
gonna
be
a
like
the
groomed
down
designs
that
are
like
more
MBC
that
the
original
ones
and
also
repeat
some
small
aspects
from
the
original
one
that
I
present
so
okay.
So
let
me
share
my
screen.
A
So
I
don't
know
if
you
guys
remember
these,
but
I
think
I
showed
these
last
year,
maybe
like
last,
like
first
days
of
December
I
believe
so
basically,
what's
going
on
here
is
like
in
testing.
We
have
like
the
ability
of
knowledge
ability.
We
have
like
certain
sweet
certain
tests
that
go.
They
look
a
web
app
and
then
they
care
like
test
that
web
app
and
within
the
context
of
that,
sometimes
we
need
screenshots.
So
for
dad,
you
use
things
like
selenium.
A
You
use
like
like
a
webdriver
that
goes
on,
like
navigates,
into
like
an
actual
review
up
and
then
performs
all
these
actions
and
then
like.
If
you
give
the
instruction
to
the
script
to
take
a
screenshot
on
failure
or
like
on
certain
case
scenarios,
then
it
will
do
that
right.
So
because
this
is
kind
of
like
an
important
use
case
for
many
people
taking
the
screenshots,
then
we
decided
that
we
wanted
to
create
like
a
dedicated,
buford
screenshot.
So
what
I,
what
I
showed
last
year,
was
basically
okay.
A
There's
gonna,
be
these
kind
of
like
super
you
in
the
job
view
that
it's
gonna
show
you,
like.
All
the
tests
that
are
running
for
that
particular
job
and
then,
if
they
have
screenshots,
you
will
see
that
they
do
have
screenshots
and
you
can
expand
them
and
you
can
see
like
the
screenshot
and
you
can
see
the
details
of
the
screenshot
like
the
beaut
or
sites
had
in
which
took
the
screenshot,
plus
all
the
user.
Details
like
the
URI
of
the
like
of
the
review
up
in
which
the
at
which
the
screenshot
was
taken.
A
So
there's
a
lot
of
details
and,
of
course,
these
had
like
lot
of
functionality.
Functionality
going
on
here,
like
there
was
I,
get
the
ability
of
switching
between
sessions
like
filtering
by
status,
hiding
tasks
spanning
screenshots,
so
there
was
quite
a
bit
of
things
going
on
there
and
then
the
ability
of
seeing
the
raw,
the
raw
logs
for
that.
So
this
is
goodbye
like
we
had
to
groom
down.
These
I
mean
we.
We
we
needed
sort
like
very
smaller
and
there's
also.
A
Some
considerations
about
this
design
in
terms
of
like
where
it's
leaving
on
right
now,
so
these
leaves
at
the
job
level
and
after
for
further
analysis,
we
are
we
realize
that,
like.
If
you
have
this
at
the
job
level,
then
if
you
have
multiple
jobs
that
are
taking
screenshots,
then
it
becomes
very
painful
to
just
go
Alex,
inspect
all
those
jobs
and
see
all
those
screenshots,
and
that's
that's
something
that
can
definitely
happen.
B
A
So
the
first
thing
that
we
did
is
that
now
we
are
not
gonna,
have
this
view
of
the
job
level
anymore.
So
the
idea
is
that
now
this
view
is
gonna
live
at
the
pipeline
level,
and
this
is
better
because,
like
we
already
have
tests
and
tests
are
very
similar
to
what
screenshots
is
gonna
be
doing
so
in
tests,
we
basically
show
all
the
suites
that
run
jobs
and
we
we
show
the
jobs
or
like
the
actions
that
those
job
jobs.
A
Sorry
the
Sweden
job,
is
basically
the
same.
So
we
we
Schober,
although
all
the
jobs-
and
we
showed
the
actions
that
the
job
took
when
he
was
running
the
tests
we
show
if
it
fail
or
if
it
passed,
and
then
we
show
the
trace-
and
we
show
like
many
of
those
things,
so
it's
very
very
similar
to
like
what
we're
gonna
do
with
screenshots.
A
Basically,
this
means
that
option
because,
like
if
you
see,
for
instance,
this
example
that
we
have
here,
there's
like
like,
like
135
through
1458
tests,
so
like
that's
pretty
hard
to
navigate
like
if
you
want
to
find
the
screenshots
inside
those
tests,
it's
gonna
be
pretty
hard,
and
then
that
means
that
you
also
need
to
build
some
sort
of
functionality.
For
that
you
know
like
some
filtering
or
some
type
of
sub
that
allows
you
to
see
the
screenshots
and
then,
on
the
other
hand,
is
like
most
people.
A
They
just
want
to
see
the
screenshots
on
isolation
right.
They
want
to
see
like
show
me
all
the
things
that
fail
and
like
the
ones
that
had
screenshots
and
I
care
about
those
things,
I,
don't
wanna.
If
I
wanna
expect
the
tests
deeper,
then
I'll
just
go
to
the
test.
Stop
so
basically,
we
decided
that
we
were
gonna
use.
The
ship
like
create
like
a
new
tab
called
screenshots,
and
this
tab
is
basically
a
table
and
the
table
shows
the
name
of
the
action
the
suite
from
which
is
coming.
A
The
suite
is
basically
the
equivalent
to
the
job
the
status
of
the
action
on
a
start
time,
duration,
all
that
stuff,
the
action
that
it
took
in
terms
of
like
pioneers,
selector
and
all
that
stuff
and
then
the
screenshot
itself
right.
So
these
in
this
particular
case,
we're
saying
that
there's
like
five
screenshots
right.
So
all
those
screenshots
are
shown
here
in
context
and
then
there's
two
things
that
I'm
going
to
explain
here.
A
One
is
like
there's
the
ability
of
downloading
the
screenshots
or
like,
in
this
particular
case,
downloading
the
whole
artifact,
because
there's
two
things:
there's
a
screenshot
and
then
there's
the
actual
kind
of
like
roll
lock
that
you
can
download
as
well.
So
basically,
the
idea
here
is
that
these
will
don't
download
a
sheep
with
like
a
seeped
pile
with
the
screenshot
and
and
kind
of
a
crawl
lock
like
the
XML
xunit
test,
and
then
one
particular
aspect
about
these-
that
it's
pretty
important
is
that
just
screenshot
is
not
gonna.
A
Give
them
enough
context
of
what
happened.
I
mean
it
matters
a
lot,
because
you
can
see
that
something
is
wrong,
so
maybe
they
can
say,
like
you
know,
like
these
buttons
shifted,
like
five
pixels
to
the
left,
for
some
reason
right.
So,
like
not
that's
important,
but
now
they
need
to
understand
why
you
know
I
what
class.
What
was
the
test
that
failed
so
like,
for
instance,
the
test
was
testing
for
like
CSS
classes
right.
A
So
from
this
view,
they
cannot
tell
that
so
the
way
to
actually
figure
out
that
is
by
looking
at
the
trace,
there's
like
like
the
trace
that
block
trace
that
matters
for
that
particular
screenshot.
So
the
way
that
you
achieve
that
is
by
clicking
this
guy
so
like.
If
you
hover
over
this
guy,
you
will
see
like
spheal
trace
and
then
we
would
present
this
model
that
has
the
trace
right.
A
I
show
these
to
my
UX
body
yesterday
and
she
Alexis,
and
she
she
had
some
good
comments
about
like
the
fact
that
we
were
trying
to
do
it
like
these,
like
she
had
some
concerns
about
like
context
like.
Do
you
lose
context
too
much
if
you
click
these
and
then
like
you're
kind
of
like
oh
what
I'm
looking
at
you
know
what
what
does
this
thing
correspond
to
in
the
table?
Right
and
I
agree
with
her
I
think,
like
there's,
probably
some
things
that
we
could
do
to
improve.
A
That
I
mean
we
have
the
name
of
the
we
have
the
name
of
the
action
under,
but
one
thing
that
I
was
thinking
is
maybe
even
showing
the
screen
on
the
morale
as
well.
I
mean
like
that's
basically
redundant
but
like
I,
don't
think
it's
gonna
hurt
the
user
and
they
can
see
like
oh
here's,
the
screenshot,
and
here
you're
like
the
load
screenshot,
you
can
see
all
the
trace
so
yeah.
This
is
basically
so.
A
The
idea
now
is
just
like
show
this
table
always
with
the
screenshots
expanded,
because
it's
all
the
screenshots
that's
name
of
the
tab.
This
is
gonna
work
for
anything
that
it's
coming
from
an
ex.
You
need
report
so
like.
If
your
test
generates
an
X,
you
need
report
regardless,
if
it's
using
selenium
or
puppeteer,
whatever
like
it's
or.
However,
you
are
like
doing
these.
It's
always
gonna
read,
look
for
the
X
unit
report
and
then
it's
gonna
generate
this
view.
A
Out
of
that,
and
then
one
thing
that
I
also
did
here
is
like
showing
the
error
State
for
this,
because,
like
there's
like
one
kind
of
grace
condition
which
is
like
it
can
actually
like,
like
the
whole
thing
can
like
say
like
oh
I'm,
looking
at
X,
you
need
report.
I
see
screenshots,
but
they're
like
it
goes
on
a
characterize
to
retrieve
those
screenshots
and
they
failed
for
some
reason,
I
mean,
like
that's
unlikely,
but
it
could
happen
so
I
yeah
we
needed
to
collect
some
Tom
ever
stay
there.
For
that.
A
Yeah,
that's
pretty
much.
It
like
I,
also
like,
of
course,
I'm
looking
for
feedback
on
all
these
things,
but
I
also
was
thinking
about
like
an
older
I
have
been
thinking
about
one
particular
kind
of
like
think
about
this
design,
which
happens
a
lot
in
grid
lab,
actually,
which
is
the
fact
that,
for
instance,
we
tests
on
with
security
which
is
not
showing
up
here.
We
we
only
show
those
stops.
A
If
the
pipeline
has
those
things
configure
right
like
like,
if
it
has
security
things
going
on,
then
he
will
show
that
tab
right
and
I
was
wondering
if
that's
the
best
practice
right
like
do.
We
really
want
to
do
it
that
way,
because
I
think
that,
basically
like
it
doesn't
definitely
doesn't
favor.
Discoverability
I
mean
like
if
you,
if
you
don't
see
the
security
tab,
although
you
don't
have
security
jobs.
A
Let
me
see
if
I
can
show
this
quickly.
I
was
looking
around
each
lab.
I,
don't
think
we're
very
consistent
with
that
like
if
you
go
to
something
like,
for
instance,
I'm
gonna,
go
to
this
repo
and
you
go
to
security
and
I
go
to
security.
Were
here
like
we,
we
don't
hide
security,
but
right
like
we,
we
have
like
an
empty
state
right
like
where
we
say
like
Oh,
learn
more
about
this
thing.
A
If
you
want
to
use
it
right,
so
I
was
wondering
why
we
hide
these
stops
there,
even
if
that
functionality
exists
there
right
so
I,
don't
know
that.
Basically,
it's
kind
of
like
that.
The
last
be
that
I
want
to
say,
because
I
think
that's
kinda,
like
something
for
discussion.
I
want
to
hear
your
thoughts
on
that,
because
I'm
inclining
my
thoughts
towards
idea
that
we
should
have
all
the
available
cops
always
showing
up,
and
if
you
don't
have
those
things
you
should
click
on
them
and
there's
an
empty
state
that
says
like.
A
B
Think
that's
a
pretty
valid
point
about
I
mean
it's
one
of
our
larger
underlying
UX
problems
across
kilobits
discoverability
of
features,
so
I
would
guess
the
reason
why
we've
gone
to
this
hide
tabs
scenario
is
probably
born
out
of
I'll.
Try
not
to
get
on
a
rant
here,
but
there's
one
of
my
least
favorite
patterns
in
gitlab
right.
C
C
B
The
filters,
and
now
we
use
tabs
as
filters
and
even
at
the
point,
I
understand
why
you're
breaking
out
the
screenshots,
in
my
opinion,
long
term
ideal
situation
would
be
like
this
is
a
test
and
it's
a
filterable
set
of
that.
But
we
don't
have
these
robust
filters
on
a
lot
of
these
views.
So
we
end
up
I
feel
like
going
to
this
tab
pattern
run
and
then
I
think
that's
why
it's
hidden
right,
because
my
sense
of
failed
jobs
isn't
there
to
just
hide
that
top
right.
I!
Really
don't
like
this
pattern.
No.
A
I
didn't
I
agree
with
you,
I,
like
basically
counting
what
I
was
thinking
at
the
beginning
is
like
I.
Don't
want
to
create
a
new
table
right
I
want
to
keep
everything
routine
tests
because
text,
it's
like
what
like
these
horrible
tests
right,
but
a
little
bit
of
context,
I,
don't
think
this
is
gonna,
make
sense
until
I
actually
show
it.
A
A
D
A
A
So
what
I
was
trying
to
get
to
is
like,
while
these
loads,
like
my
initial
plan,
was
like,
let's
add
some
type
of
filter
into
these
right,
like
maybe
there's
like
something
like
like
a
control
at
the
top
of
this
table.
That
he's
like
show
me
all
the
suites
that
have
screenshots.
So
maybe
you
have
screenshots
here
and
everything
right,
but
yeah,
that's
the
problem
that
I
see
is
the
like.
First
of
all,
we're
not
ready
for
that
like
to
create
like
that.
Verbose
then
be
like
the
whole
thing
about.
A
And
there's
like
a
lot
of
things
going
on
here
right
like
there's
all
these
tests,
you
know
so
like
you
need,
if
you
had
filters
outside
of
that
context,
you
need
to
have
also
filters
inside
these
contexts,
so
the
design
becomes
way
more
complex
right
like
well.
The
other
one
is
like
isolated
right.
It's
just
kinda
like
that.
Yeah,
like
all
these
pipeline,
has
the
screenshots,
which
matter
to
me,
because
I
want
to
see
how
I
think
we're
behaving
rows.
B
B
I
think
would
the
way
we
should
go.
I
think
this
is
a
perfectly
valid
NBC
I
think
I'm,
just
also
suggesting
that
whenever
I
see
15,000
of
anything
yeah,
we
need
to
filter
like
eventually,
I
would
like
to
see
this.
Maybe
you
know
acts
I.
Think
this
view
is
illustrating
that
it's
it's
kind
of
lost
its
value.
A
At
this,
oh,
it's
totally
on
you
I
mean
like
so
we're
not
totally
usable
but
like
what
what
value
do
I
get
from
these
really
like
out-of-the-box,
not
not
much
right
like
it's.
Unless
I
do
come
on
fine
and
I
tried
to
find
like
something
like
a
pool.
Well,
I,
don't
know
all
right!
That's
you
know,
like
that's
kind
of
how
I,
like
that's
kinda,
like
the
hockey
way
that
I
will
use
this
view
somehow,
but
I.
Don't
even
think
that
that's
like
really
that
good!
It's
just
this.
B
Use
case
yeah,
but
I
mean,
if
it's
possible,
it's
possible
right
and
right
and
and
when
you
run
tests,
it's
not
like
a
human
being.
Actually
like
did
this
on
purpose.
This
is
what
happens
when
you
write
code
with
tests
and
to
get
coverage,
and
things
like
this
yeah.
That's
the
nature
of
writing.
Tests
is
right
there.
Obviously
it's
going
to
be
a
lot
of
them.
I
mean
you're
right.
Gila
organs
is
enormous
right.
A
C
A
D
A
B
D
D
This
is
one
of
those
cases
where
it's
clear
that
we're
extending
beyond
any
original
scope
for
this
cordis
page
right,
I,
agree
with
Mike
that
this
is
a
very
valid
MVC
and
that
we
should
move
forward
in
order
to
move
forward.
But
I
do
think
that
you
know
we
have
the
need
to
rethink
our
concept
of
extending
the
pipeline
page
it
definitely
if
that
is
part
of
purely
CI
I,
don't
think
so
at
the
same
time,
because
it
touches
so
many
other,
you
know
workflows
and
stage
groups.
D
A
Yeah
right
right,
yeah
I
mean
like
definitely
we're
not
gonna
solve
this
problem
here
on
this
school,
but
I
think
we
should
start
thinking
about
how
we're
gonna
feel
with
these
I
also
feel
that
we
can
like
abuse,
taps
to
matching
good
love.
You
know
like
when
we
don't
know
how,
like
a
way
when
we
don't
have
fine,
like
a
good
way
to
play.
A
Something
in
context
of
a
screen
is
like
Oh
put
it
in
a
top
like
the
top
is
like
the
catch-all
pattern
you
know
and
tops
are
good,
but
I
think
like
we
should
have
like
a
healthy
balance
it.
It
cannot
be
that,
like
every
time
that
we
have
something,
it's
like
a
new
top
right
like
it's
not
good
for
users,
he
does
some
favorites
cover
ability
and
our
tops
are
not
great
I
mean
I'm.
Gonna,
say
it.
A
I
feel
that
our
tops
are
like
sometimes
I,
look
at
them
and
like
it's
like,
because
I
work,
a
deal
up,
I
know
that
how
they
work
but
I
think
like
if
I
was
a
user.
When
you
go
to
an
issue
like
it
took
me
a
while
I
mean
let
me
open
one.
He
took
me
a
while
to
discover
like
the
design
stuff
you
know
like
I
did
like
it's
kind
of
like
until
someone
said.
Oh
there's,
a
design
stub,
you
know,
so
why.
D
D
Issue
on
especially
like
there
are
two
parts
here
right:
it's
legacy,
UI,
that's
being
extended
beyond
its
original
idea
and
then
there's.
The
second
thing,
which
is
the
design
of
our
tabs
component,
is
not
fulfilling
its
made
of
being.
You
know,
accessible
and
cleared
once
a
user.
This
is
a
concept.
It
was
very
clear,
especially
for
the
pipeline's
tab
within
the
merge
quest
page
where
users
were
scrolling
over
it.
Looking.
D
For
a
sign
where
you
know
some
hint
of
pipeline-
and
they
just
didn't
see
it,
there's
a
there's,
dedicated
issue
to
improving
it.
Okay,
but
you
know
it's
it's
it's
the
same
thing
with
image
cosplay,
you
know
the
original
intent
didn't
include
this
much
expansion
right.
This
needs
to
be
a
dedicated
effort,
it's
not
something
that
is
easily
so.