1
00:00:07,170 --> 00:00:09,956
- [Instructor] Automation
helps us keep up.

2
00:00:09,956 --> 00:00:11,624
Let's look at some good ways

3
00:00:11,624 --> 00:00:14,721
to approach automation
in agile development.

4
00:00:14,721 --> 00:00:16,505
In traditional projects, we know

5
00:00:16,505 --> 00:00:18,360
when automation is going to happen.

6
00:00:18,360 --> 00:00:20,767
It's going to happen after a
whole bunch of other things,

7
00:00:20,767 --> 00:00:22,756
after everything is stable.

8
00:00:22,756 --> 00:00:24,641
And the sole purpose of automation

9
00:00:24,641 --> 00:00:27,000
is usually for regression testing.

10
00:00:27,000 --> 00:00:28,802
In agile development,

11
00:00:28,802 --> 00:00:30,973
automation is not only
for regression testing.

12
00:00:30,973 --> 00:00:34,561
It helps us get fast
feedback on new features,

13
00:00:34,561 --> 00:00:38,085
on changes, and it's not
only functional testing,

14
00:00:38,085 --> 00:00:40,040
we may be testing our infrastructure,

15
00:00:40,040 --> 00:00:42,280
we may be using automation to help

16
00:00:42,280 --> 00:00:44,773
with continuous delivery and appointments.

17
00:00:44,773 --> 00:00:46,863
We need to automate when it makes sense,

18
00:00:46,863 --> 00:00:48,227
and we're going to talk later

19
00:00:48,227 --> 00:00:50,280
about what not to automate.

20
00:00:50,280 --> 00:00:52,133
Automation is hard, right?

21
00:00:52,133 --> 00:00:54,474
We need a diversity of
skill sets on the team

22
00:00:54,474 --> 00:00:56,360
to be able to solve problems.

23
00:00:56,360 --> 00:00:59,240
So, having people with
different backgrounds

24
00:00:59,240 --> 00:01:01,412
collaborating on automation problems

25
00:01:01,412 --> 00:01:03,114
help us tackle them.

26
00:01:03,114 --> 00:01:06,178
For example, a tester and a developer

27
00:01:06,178 --> 00:01:08,758
can pair to fix so-called flakey tests

28
00:01:08,758 --> 00:01:11,271
that fail some of the time.

29
00:01:11,271 --> 00:01:13,981
Maybe there is a real bug
there that the tester can spot.

30
00:01:13,981 --> 00:01:16,152
Or maybe there's just something
wrong with the code design

31
00:01:16,152 --> 00:01:18,596
that the developer can spot.

32
00:01:18,596 --> 00:01:21,123
This has to be the priority
for the whole team.

33
00:01:21,123 --> 00:01:23,192
Also reach out to other teams,

34
00:01:23,192 --> 00:01:25,027
if you're working at a larger company,

35
00:01:25,027 --> 00:01:27,655
maybe you have a community
practice or guild

36
00:01:27,655 --> 00:01:32,577
or even just reach out to the
community at large for help.

37
00:01:32,577 --> 00:01:36,299
Share ideas on tools
on how different people

38
00:01:36,299 --> 00:01:39,768
have approached different
automation obstacles.

39
00:01:39,768 --> 00:01:42,613
Some things are definitely
harder to automate than others.

40
00:01:42,613 --> 00:01:45,878
For example, if your user interface

41
00:01:45,878 --> 00:01:47,933
has a lot of business logic built in,

42
00:01:47,933 --> 00:01:50,272
like a Javascript application,

43
00:01:50,272 --> 00:01:52,481
that definitely poses
automation challenges

44
00:01:52,481 --> 00:01:56,690
that you don't have if you're
testing a server side API,

45
00:01:56,690 --> 00:01:58,895
so discuss up front,

46
00:01:58,895 --> 00:02:01,757
as you're starting to talk
about developing a new feature,

47
00:02:01,757 --> 00:02:06,374
how can you design the code
to facilitate the automation.

48
00:02:06,374 --> 00:02:09,102
We need reliable, quick feedback loops

49
00:02:09,102 --> 00:02:11,272
via our continuous integration,

50
00:02:11,272 --> 00:02:13,953
and we need to automate not only tests,

51
00:02:13,953 --> 00:02:16,465
but things like automating our deploys

52
00:02:16,465 --> 00:02:18,215
to our test environments,

53
00:02:18,215 --> 00:02:19,797
to our production environments.

54
00:02:19,797 --> 00:02:22,255
There are so many aspects
of our infrastructure

55
00:02:22,255 --> 00:02:24,448
that we need to automate.

56
00:02:24,448 --> 00:02:26,086
Let's talk about some things

57
00:02:26,086 --> 00:02:28,463
that testers are particularly good at.

58
00:02:28,463 --> 00:02:30,349
Our whole team works together,

59
00:02:30,349 --> 00:02:32,960
but we each add value with our deep skills

60
00:02:32,960 --> 00:02:34,696
out of our T-shaped skills.

61
00:02:34,696 --> 00:02:37,645
And testers are really
good at specifying tests

62
00:02:37,645 --> 00:02:39,480
and asking questions.

63
00:02:39,480 --> 00:02:42,291
When you think about hiring
testers for your agile team,

64
00:02:42,291 --> 00:02:44,249
do you really want to hire more coders,

65
00:02:44,249 --> 00:02:47,211
or do you want to hire people
with great testing skills.

66
00:02:47,211 --> 00:02:49,719
Testers who use their
critical thinking skills

67
00:02:49,719 --> 00:02:52,953
to verify quality attributes
beyond the functional,

68
00:02:52,953 --> 00:02:56,495
such as a quadrant four test
we discussed in lesson two,

69
00:02:56,495 --> 00:02:58,689
that may be who you want on your team.

70
00:02:58,689 --> 00:03:00,541
To testers and developers,

71
00:03:00,541 --> 00:03:03,523
collaborating together is the
ideal scenario on your team.

72
00:03:03,523 --> 00:03:05,815
Leverage those different
skill sets together

73
00:03:05,815 --> 00:03:07,653
to solve problems.

74
00:03:07,653 --> 00:03:11,191
Test code is equally or more
valuable than production code,

75
00:03:11,191 --> 00:03:12,859
in my opinion.

76
00:03:12,859 --> 00:03:15,270
As Janet mentioned, it's
living documentation.

77
00:03:15,270 --> 00:03:17,829
It tell us how our code works now.

78
00:03:17,829 --> 00:03:20,575
Developers who write production code

79
00:03:20,575 --> 00:03:23,813
are also the ideal people
to write test fixture code,

80
00:03:23,813 --> 00:03:26,372
because they're really
good at code design,

81
00:03:26,372 --> 00:03:28,969
and testers are ideal
for a specifying test,

82
00:03:28,969 --> 00:03:31,312
because they're using their
critical thinking skills

83
00:03:31,312 --> 00:03:32,878
all the time,

84
00:03:32,878 --> 00:03:34,965
analyzing risk, and knowing what to cover

85
00:03:34,965 --> 00:03:37,353
with automated tests.

86
00:03:37,353 --> 00:03:39,863
DevOps means a combination of developers,

87
00:03:39,863 --> 00:03:44,057
testers, system administrators,
database experts, and others

88
00:03:44,057 --> 00:03:46,804
collaborating on our test environments,

89
00:03:46,804 --> 00:03:48,707
our continuous integration,

90
00:03:48,707 --> 00:03:50,850
and our deployment infrastructure.

91
00:03:50,850 --> 00:03:54,878
So we need collaboration
across all these skill sets.

92
00:03:54,878 --> 00:03:57,015
But the key is in the collaboration.

93
00:03:57,015 --> 00:03:59,109
For example, the programmers may think

94
00:03:59,109 --> 00:04:02,425
to add code to the test script to tell it

95
00:04:02,425 --> 00:04:04,233
when the page is loaded,

96
00:04:04,233 --> 00:04:05,868
so that the test can proceed,

97
00:04:05,868 --> 00:04:07,673
that way we avoid having any timeouts,

98
00:04:07,673 --> 00:04:09,173
or false failures.

99
00:04:10,348 --> 00:04:13,364
We need data experts, we need operations

100
00:04:13,364 --> 00:04:15,473
and system administration experts,

101
00:04:15,473 --> 00:04:17,849
as well as coding and testing experts.

102
00:04:17,849 --> 00:04:19,849
We like generalizing specialists,

103
00:04:19,849 --> 00:04:23,832
but we also like bringing
in the skill set we need

104
00:04:23,832 --> 00:04:26,426
and transferring those
skills across the team.

105
00:04:26,426 --> 00:04:28,517
Look for tools that promote collaboration,

106
00:04:28,517 --> 00:04:31,111
and that allow people
who do not write code

107
00:04:31,111 --> 00:04:34,461
to participate in
specifying automated tests.

108
00:04:34,461 --> 00:04:37,797
Not only testers, but business analysts,

109
00:04:37,797 --> 00:04:40,423
maybe product owners,
maybe even designers.