WEBVTT

1
00:00:00.732 --> 00:00:04.202 line:90% position:50% align:center size:90%
Identifying accessibility
and usability needs.

2
00:00:04.798 --> 00:00:07.743 line:90% position:50% align:center size:90%
Using a text editor
through speech recognition.

3
00:00:08.965 --> 00:00:11.184 line:90% position:50% align:center size:90%
Alejandro Rodríguez Ascaso.

4
00:00:11.384 --> 00:00:14.475 line:90% position:50% align:center size:90%
School of Computer Engineering. UNED.

5
00:00:18.146 --> 00:00:23.136 line:90% position:50% align:center size:90%
To analyse a system’s accessibility and
usability needs, we start with a person,

6
00:00:23.136 --> 00:00:28.470 line:90% position:50% align:center size:90%
a goal and a context of use. In our
example, Claudia has limited manual

7
00:00:28.470 --> 00:00:33.892 line:90% position:50% align:center size:90%
dexterity and wants to write and edit
a text. She uses her voice to enter it.

8
00:00:33.892 --> 00:00:37.725 line:90% position:50% align:center size:90%
We will examine which
needs this system meets.

9
00:00:39.514 --> 00:00:43.292 line:90% position:50% align:center size:90%
Let’s review the accessibility
components of the interaction sequence

10
00:00:44.668 --> 00:00:48.695 line:90% position:50% align:center size:90%
Perception. Perception has
an interaction component in which several

11
00:00:48.695 --> 00:00:52.744 line:90% position:50% align:center size:90%
aspects play a role, namely:
the existence and position of the system’s

12
00:00:52.744 --> 00:00:56.687 line:90% position:50% align:center size:90%
controls, the state the system is in,
and the feedback produced after

13
00:00:56.687 --> 00:00:58.942 line:90% position:50% align:center size:90%
any control is activated by the user.

14
00:01:00.155 --> 00:01:04.835 line:90% position:50% align:center size:90%
Operation. Operation occurs when
the user makes use of any of the system’s

15
00:01:04.835 --> 00:01:09.598 line:90% position:50% align:center size:90%
functions. The system shall allow
users to use all the system’s functions

16
00:01:09.598 --> 00:01:10.395 line:90% position:50% align:center size:90%
or actions.

17
00:01:12.374 --> 00:01:16.245 line:90% position:50% align:center size:90%
Comprehension. Comprehension involves,
on the one hand, understanding

18
00:01:16.245 --> 00:01:20.206 line:90% position:50% align:center size:90%
the output information produced
by the system, and on the other,

19
00:01:20.206 --> 00:01:23.036 line:90% position:50% align:center size:90%
understanding in general
how the product is used.

20
00:01:25.092 --> 00:01:29.359 line:90% position:50% align:center size:90%
Let us review that sequence in Claudia’s
case. The speech recognition

21
00:01:29.359 --> 00:01:33.585 line:90% position:50% align:center size:90%
software indicates that it is
listening. Claudia says a sentence

22
00:01:33.585 --> 00:01:38.519 line:90% position:50% align:center size:90%
and the words appear in the document.
She then gives the command to insert

23
00:01:38.519 --> 00:01:42.857 line:90% position:50% align:center size:90%
a new line, and the cursor changes
position. The feedback allows her

24
00:01:42.857 --> 00:01:45.472 line:90% position:50% align:center size:90%
to check that she has
achieved what she intended:

25
00:01:45.472 --> 00:01:48.985 line:90% position:50% align:center size:90%
entering text and giving
a command to the text editor.

26
00:01:48.985 --> 00:01:53.832 line:90% position:50% align:center size:90%
Claudia understands the recogniser’s
waiting state, as well as how it

27
00:01:53.832 --> 00:01:55.946 line:90% position:50% align:center size:90%
works together with the text editor.

28
00:01:58.819 --> 00:02:03.770 line:90% position:50% align:center size:90%
Next, we analyse each step of the
interaction sequence, looking mainly

29
00:02:03.770 --> 00:02:07.574 line:90% position:50% align:center size:90%
for the accessibility and usability
needs that specifically apply

30
00:02:07.574 --> 00:02:09.464 line:90% position:50% align:center size:90%
to Claudia’s profile.

31
00:02:11.856 --> 00:02:17.188 line:90% position:50% align:center size:90%
We begin by discussing perception.
Claudia perceives the system’s state:

32
00:02:17.188 --> 00:02:21.690 line:90% position:50% align:center size:90%
it indicates that it is listening
to her, ready to recognise her voice.

33
00:02:21.690 --> 00:02:26.575 line:90% position:50% align:center size:90%
When she speaks, the system confirms
that it has carried out the instruction,

34
00:02:26.575 --> 00:02:30.087 line:90% position:50% align:center size:90%
either by displaying the recognised
word within the text itself

35
00:02:30.087 --> 00:02:33.583 line:90% position:50% align:center size:90%
or by displaying the recognised
command, for example “new line”.

36
00:02:35.851 --> 00:02:39.593 line:90% position:50% align:center size:90%
Claudia has no specific accessibility
need regarding perception,

37
00:02:39.593 --> 00:02:44.662 line:90% position:50% align:center size:90%
so we record the general category
“Needs related to perceivability”,

38
00:02:44.662 --> 00:02:49.358 line:90% position:50% align:center size:90%
together with the heuristic
“Visibility of system status”.

39
00:02:52.482 --> 00:02:57.117 line:90% position:50% align:center size:90%
Regarding operation, we begin by noting
that Claudia trains the speech

40
00:02:57.117 --> 00:03:02.266 line:90% position:50% align:center size:90%
recognition system. She reads aloud
a sentence displayed on the screen.

41
00:03:03.306 --> 00:03:08.216 line:90% position:50% align:center size:90%
The system allows Claudia to train it
personally so that speech recognition

42
00:03:08.216 --> 00:03:11.729 line:90% position:50% align:center size:90%
works better (7-2-2-b).

43
00:03:13.382 --> 00:03:17.303 line:90% position:50% align:center size:90%
Claudia dictates a word and it
appears in the document.

44
00:03:17.303 --> 00:03:21.123 line:90% position:50% align:center size:90%
She then says “new line”
and the cursor moves to the next line.

45
00:03:23.897 --> 00:03:27.918 line:90% position:50% align:center size:90%
Claudia uses speech recognition
to write in the document and control

46
00:03:27.918 --> 00:03:35.392 line:90% position:50% align:center size:90%
the text editor (need 7-4-5-a, “Operate
a program using speech recognition”).

47
00:03:35.392 --> 00:03:37.818 line:90% position:50% align:center size:90%
This need breaks down into two others.

48
00:03:40.396 --> 00:03:44.516 line:90% position:50% align:center size:90%
Claudia dictates a sentence
and the words appear in the document.

49
00:03:44.516 --> 00:03:48.463 line:90% position:50% align:center size:90%
She then spells out an acronym
and the editor inserts it into the text.

50
00:03:49.335 --> 00:03:53.763 line:90% position:50% align:center size:90%
She writes by dictation and uses
spelling to enter particular words

51
00:03:53.763 --> 00:04:00.351 line:90% position:50% align:center size:90%
or acronyms (need 7-1-1-e, “Have
alternative modalities for text input”).

52
00:04:03.813 --> 00:04:07.697 line:90% position:50% align:center size:90%
Claudia says “new line” and the
cursor moves to the next line.

53
00:04:07.697 --> 00:04:13.130 line:90% position:50% align:center size:90%
After entering a word, she says “undo”,
and the word she has just entered

54
00:04:13.130 --> 00:04:13.890 line:90% position:50% align:center size:90%
disappears.

55
00:04:14.897 --> 00:04:18.347 line:90% position:50% align:center size:90%
When she says “Go to sleep”,
speech recognition pauses.

56
00:04:18.347 --> 00:04:24.687 line:90% position:50% align:center size:90%
All of this corresponds to need 7-3-e,
“Have a method of controlling

57
00:04:24.687 --> 00:04:26.551 line:90% position:50% align:center size:90%
the system using speech”.

58
00:04:28.356 --> 00:04:32.391 line:90% position:50% align:center size:90%
When speech recognition is paused,
Claudia can talk to another person

59
00:04:32.391 --> 00:04:37.066 line:90% position:50% align:center size:90%
and the document stays unchanged.
She then resumes dictation by saying

60
00:04:37.066 --> 00:04:38.971 line:90% position:50% align:center size:90%
the command “Get to work”.

61
00:04:39.287 --> 00:04:45.510 line:90% position:50% align:center size:90%
We therefore record need 7-4-4-i,
“Take a break from interaction and return

62
00:04:45.510 --> 00:04:48.261 line:90% position:50% align:center size:90%
later with a minimum of additional work”.

63
00:04:50.481 --> 00:04:55.119 line:90% position:50% align:center size:90%
While dictating and giving voice commands,
Claudia keeps her hands at rest.

64
00:04:55.119 --> 00:04:58.929 line:90% position:50% align:center size:90%
The text and cursor change
in response to her voice.

65
00:05:00.225 --> 00:05:05.780 line:90% position:50% align:center size:90%
For that reason, another need we can
record is “7-4-2-b — Operate without

66
00:05:05.780 --> 00:05:11.274 line:90% position:50% align:center size:90%
using the hands”. In fact, it is less
precise for Claudia’s case because

67
00:05:11.274 --> 00:05:15.500 line:90% position:50% align:center size:90%
she specifically prefers using
her voice to work on the document.

68
00:05:18.398 --> 00:05:22.948 line:90% position:50% align:center size:90%
In general, during operation, we observe
that Claudia uses her voice to write,

69
00:05:22.948 --> 00:05:26.106 line:90% position:50% align:center size:90%
give commands and pause
interaction when she needs to.

70
00:05:27.085 --> 00:05:31.216 line:90% position:50% align:center size:90%
We therefore add the corresponding
Nielsen heuristic, “Flexibility

71
00:05:31.216 --> 00:05:32.829 line:90% position:50% align:center size:90%
and efficiency of use”.

72
00:05:37.156 --> 00:05:41.352 line:90% position:50% align:center size:90%
We now address understanding,
which in practice is present throughout

73
00:05:41.352 --> 00:05:46.668 line:90% position:50% align:center size:90%
the interaction sequence. Seeing
an indicator that the system is listening

74
00:05:46.668 --> 00:05:50.443 line:90% position:50% align:center size:90%
is not enough to know whether
the system is listening.

75
00:05:50.443 --> 00:05:54.619 line:90% position:50% align:center size:90%
We also need to interpret its meaning.
Similarly, perceiving that a new

76
00:05:54.619 --> 00:05:58.746 line:90% position:50% align:center size:90%
word has been written or that the cursor
has moved to a new line is not

77
00:05:58.746 --> 00:06:01.984 line:90% position:50% align:center size:90%
enough if we do not understand
what these changes mean.

78
00:06:03.566 --> 00:06:07.416 line:90% position:50% align:center size:90%
Claudia has no specific accessibility
need regarding understanding.

79
00:06:07.416 --> 00:06:12.904 line:90% position:50% align:center size:90%
Therefore, for her profile, we record
the general category “Needs related

80
00:06:12.904 --> 00:06:16.915 line:90% position:50% align:center size:90%
to understandability” to make clear
that we have not forgotten them.

81
00:06:19.941 --> 00:06:23.580 line:90% position:50% align:center size:90%
More specifically, we focus on
the microphone icon and its label,

82
00:06:23.580 --> 00:06:28.468 line:90% position:50% align:center size:90%
which indicate that speech
recognition is listening. In addition,

83
00:06:28.468 --> 00:06:33.136 line:90% position:50% align:center size:90%
the way the editor is used does not
essentially change because the text

84
00:06:33.136 --> 00:06:34.071 line:90% position:50% align:center size:90%
is dictated.

85
00:06:34.573 --> 00:06:39.848 line:90% position:50% align:center size:90%
The preceding example is related
to compliance with two Nielsen heuristics:

86
00:06:39.848 --> 00:06:43.750 line:90% position:50% align:center size:90%
“Speak the users’ language”
and “Consistency and standards”.

87
00:06:46.248 --> 00:06:50.502 line:90% position:50% align:center size:90%
The help provides instructions
and example commands for using speech

88
00:06:50.502 --> 00:06:51.259 line:90% position:50% align:center size:90%
recognition.

89
00:06:52.018 --> 00:06:56.716 line:90% position:50% align:center size:90%
Claudia also mentions the help
provided by the system, so the “Help

90
00:06:56.716 --> 00:06:58.920 line:90% position:50% align:center size:90%
and documentation” heuristic is met.

91
00:07:00.899 --> 00:07:04.979 line:90% position:50% align:center size:90%
In addition to the needs arising
directly from applying the interaction

92
00:07:04.979 --> 00:07:07.035 line:90% position:50% align:center size:90%
sequence, other needs also apply.

93
00:07:08.444 --> 00:07:11.983 line:90% position:50% align:center size:90%
For example, the “undo” command
removes the word that has just been

94
00:07:11.983 --> 00:07:13.244 line:90% position:50% align:center size:90%
entered by mistake.

95
00:07:13.919 --> 00:07:20.061 line:90% position:50% align:center size:90%
This connects with need 9-6-a,
“Undo the last actions”, which in turn

96
00:07:20.061 --> 00:07:23.563 line:90% position:50% align:center size:90%
corresponds to the heuristic
“User control and freedom”.

97
00:07:27.853 --> 00:07:32.256 line:90% position:50% align:center size:90%
The speech recogniser enters the words
Claudia dictates into the word processor.

98
00:07:34.231 --> 00:07:39.234 line:90% position:50% align:center size:90%
We also record need 11-1, “Use
assistive products with the system’s

99
00:07:39.234 --> 00:07:43.733 line:90% position:50% align:center size:90%
functionalities”. The speech recogniser
works with the word processor

100
00:07:43.733 --> 00:07:45.481 line:90% position:50% align:center size:90%
for the actions demonstrated.
