Editing Spatial Layouts through Tactile Templates for
People with Visual Impairments
Jingyi Li
1
, Son Kim
2
, Joshua A. Miele
3
, Maneesh Agrawala
1
, Sean Follmer
1
1
Stanford University,
2
Vista Center for the Blind and Visually Impaired,
3
The Smith-Kettlewell Eye Research Institute
{jingyili, maneesh}@cs. stanford.edu, skim@vistacenter. org, jam@ski.org,sfollmer@stanford.edu
ABSTRACT
Spatial layout is a key component in graphic design. While
people who are blind or visually impaired (BVI) can use
screen readers or magniers to access digital content, these
tools fail to fully communicate the content’s graphic design
information. Through semi-structured interviews and con-
textual inquiries, we identify the lack of this information and
feedback as major challenges in understanding and editing
layouts. Guided by these insights and a co-design process
with a blind hobbyist web developer, we developed an in-
teractive, multimodal authoring tool that lets blind people
understand spatial relationships between elements and mod-
ify layout templates. Our tool automatically generates tactile
print-outs of a web page’s layout, which users overlay on
top of a tablet that runs our self-voicing digital design tool.
We conclude with design considerations grounded in user
feedback for improving the accessibility of spatially encoded
information and developing tools for BVI authors.
CCS CONCEPTS
• Human-centered computing → Accessibility systems
and tools; Accessibility technologies ;
KEYWORDS
accessible design tools; accessible web design; layout design;
templates; tactile overlays; blindness; visual impairments;
multimodal interfaces; accessibility
ACM Reference Format:
Jingyi Li, Son Kim, Joshua A. Miele, Maneesh Agrawala, Sean
Follmer. 2019. Editing Spatial Layouts through Tactile Templates
for People with Visual Impairments. In CHI Conference on Human
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies
are not made or distributed for prot or commercial advantage and that
copies bear this notice and the full citation on the rst page. Copyrights
for components of this work owned by others than the author(s) must
be honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior specic
CHI 2019, May 4–9, 2019, Glasgow, Scotland UK
©
2019 Copyright held by the owner/author(s). Publication rights licensed
to ACM.
ACM ISBN 978-1-4503-5970-2/19/05. . . $15.00
https://doi. org/10. 1145/3290605. 3300436
Figure 1: Our interface leverages interactive tactile sheets to
convey web page structure and content. Here, a user double
taps on a page element to hear more information.
Factors in Computing Systems Proceedings (CHI 2019), May 4–9,
2019, Glasgow, Scotland UK. ACM, New York, NY, USA, 11 pages.
https://doi. org/10.1145/3290605.3300436
1 INTRODUCTION
From books to screens, people around the world design and share
an abundance of visual media. A rich body of existing work has
been dedicated to nding the best way to visually encode, design
and present information [
28
,
30
,
34
,
47
]. Supporting access to this
visual language is especially important for people who are blind
or visually impaired (BVI) [
8
]. For example, people who are BVI
create slide decks to communicate and present their ideas in the
workplace, which is an important employability skill [
5
]. Many
BVI disability advocates create and maintain their own web blogs,
showcasing their work not only to BVI peers, but also to sighted
colleagues and curious strangers [6, 26].
Currently, people who are BVI use assistive technologies like
screen magniers, screen readers, and occasionally braille displays
to interact with digital content. Screen readers strip out graphical
information and present audio descriptions of the page content.
Although existing applications—such as PowerPoint or WordPress—
allow users to enter text in a pre-dened layout template, BVI au-
thors lack accessible ways of understanding and modifying the tem-
plate’s layout structure. Screen readers allow users to tab through
PowerPoint’s menu options and create widgets in WordPress with
their keyboard, but mouse-based interactions such as resizing or
repositioning elements are inaccessible. Creating layouts with a
screen reader oers no information about nor control over where
elements are placed besides their serialized order with regards to
other content on the page. While people with low vision who use
screen magniers can access their mouse, this technology does not
provide a sense of the page’s entire layout structure at once.
This lack of information about spatial layout makes it dicult
for BVI authors to understand how their actions impact the page’s
presentation and structure. From a series of contextual inquiries, we
observed that their current workows usually involve entering text
into a template and then using a screen reader to speak back that
text (whether by presenting their slides or visiting their webpage)
to verify their modications. Many ask sighted collaborators for
assistance on further tasks, like rearranging or adding non-textual
elements on a page, and “sanity checking” their edits. We see the po-
tential of a tool that supports BVI authors and leverages their tactile
and auditory perceptual abilities [48], to foster greater condence
and self-ecacy in not only understanding but also modifying their
own layouts.
We carried out an exploration with people who are blind and
visually impaired that identied common practices and their limita-
tions in working with layouts. Our ndings reveal that BVI people
are interested in layout design and want to make modications
to static templates beyond lling in text. Together with co-author
Son Kim, a blind Assistive Technology Specialist and hobbyist web
designer, we co-designed and prototyped a multimodal tool that
makes understanding and editing layouts—specically web pages—
more accessible. Our tool provides BVI authors access to layout
structures, augmenting their existing work ows of selecting and
lling out layouts. To address the lack of non-visual feedback cur-
rent technologies provide, our tool automatically generates tactile
sheets (which can be laser cut, embossed, or printed on capsule
paper [
2
]) from the bounding boxes of elements on a web page (Fig-
ure 1). Similar to prior methods that augment tactile graphics with
interactive audio [
3
,
14
,
25
], our system uses a tablet to verbalize
both the underlying content (e.g., Project 3 Title) and meta-content
(e.g., Heading, level 2; text alignment: left) of the layout. By break-
ing up what a screen reader would have spoken serially into two
modalities, we reduce the amount of information BVI authors have
to hold in short term memory.
In this paper, we contribute (1) the results of an initial inves-
tigation of how BVI people currently create and edit layouts and
the challenges they face, (2) a tool that aids blind people in un-
derstanding and editing spatial layouts through multimodal touch
interactions, and (3) design considerations for developing accessible
authoring tools, such as the trade-os of dierent types of spatial
representations, developing scaolding to support users learning
domain knowledge, and building systems that amplify the unique
perspectives assistive technology users bring to their designs.
2 RELATED WORK
Our approach is grounded in prior work in tactile graphics and how
to support layout design for sighted people. We draw inspiration
from existing cross- and multimodal design tools, and accessible
authoring tools.
Tactile Graphics
There is a wide range of literature [
1
,
16
,
36
] specifying best-practice
guidelines for creating tactile graphics for people with visual im-
pairments. Since these graphics traditionally have been static—
fabricated with a braille embosser or on capsule paper—they are
limited by the amount of information they can display [
9
]. Commer-
cial products like the Talking Tactile Tablet [
25
] or the ViewPlus
IVEO [
14
] add interactivity to printed tactile sheets by overlaying
them on a touchscreen that plays hand-made audio annotations
when touched. To explore tactile graphics, researchers have also
used personal fabrication machines. They have employed 3D print-
ing to let blind users access works of art [
39
], STEM education
[
11
,
20
], and traditional WIMP desktop interface elements [
4
]. Fur-
thermore, researchers have generated static laser cut tactile aids
to teach graphic design [
31
] and provide guidance for using touch
screens [
21
]. Recently, Avila et al. veried that laser cut tactile over-
lays of document layouts that—like our approach—speak underlying
text when touched were as comprehensible as screen readers for
text comprehension and navigation [
3
]. We extend this work with
a pipeline to automatically generate such sheets from web pages
and an authoring tool for spatial layouts.
Researchers continue to investigate and develop 2D matrices of
actuated pins: recent examples include the Holy Braille [
33
], the
HyperBraille [
38
], and the American Printing House for the Blind’s
Graphiti [
13
]. Although these tactile displays aord dynamic and
refreshable graphics, they are prohibitively expensive; for example,
the HyperBraille costs $50,000 USD for a half page size. Focusing
on low cost and wide reach, our tool only requires two pieces of
hardware: a tablet to run the tool and a machine, like a capsule
paper printer or embosser, to fabricate the tactile sheets.
Layout Design Support
For sighted users, researchers have extensively studied techniques
to support layout and graphic design. Design mining of the web has
produced explorable design galleries [
40
] and data driven analyses
of website layouts [
24
]. Both DesignScape [
35
] and Sketchplore
[
46
] aid novices by oering interactive layout suggestions during
the design process. R-ADoMC [
19
] and Kuhna et al. [
23
] suggest
designs for magazine covers: the former through high and low level
descriptors and the latter through image content. Focusing on lay-
out structure, researchers have developed systems to automatically
adapt grids to various viewport sizes [
18
] and make layout struc-
tures dynamic and reusable [29]. Our work also targets document
structures, but rather than oer design recommendations, we allow
BVI authors to understand and modify them.
Crossmodal and Multimodal Design Tools
Many researchers have looked at enhancing traditionally WIMP-
based design tools with multiple modalities—namely, voice and
gesture. Speak’n’Sketch [
41
] aimed to maximize the sketching
workspace by abstracting menu options to voice commands, and
Crossweaver [
43
] let designers prototype and test interactive sto-
ryboards through speaking and sketching. PixelTone [
27
], which
supported natural language and gestural inputs for image editing,
employed preview galleries of available edits to promote learning
the system vocabulary. Like PixelTone, our tool facilitates learning
through preview galleries; however, we focus not on augmenting
visual interfaces but translating them through tactile feedback and
audio annotations.
Authoring Tools for Accessibility
While 3D-printed or laser cut tactile interfaces allow for inexpen-
sive and customizable representations, they rely on pre-fabricated
models often made by researchers or experts. To address this issue,
researchers have created authoring tools that range from generating
tactile overlays of graphics [
17
] and data visualizations [
10
], to pic-
ture books [
22
] and annotated 3D models [
42
]. However, these tools
rely on a sighted user to design and produce the tactile artifacts for
a blind user, which could limit their independence and ability to
more deeply understand the material gained in the creation process.
Researchers have thus developed tools that enable blind users to
be the creators of their own media. In the domain of audio editing,
The Moose [
37
] enabled blind users to splice audio by acting as a
powered mouse with force feedback, while the Haptic Wave [
44
]
translated digital waveforms into the haptic channel. Facade [
15
]
employs crowdsourcing to let blind users generate overlays for
household appliances like microwaves, and Taylor et al. present a
system to create interactive 3D printed tactile maps [
45
]. Like these
systems, our tool allows blind users to create for themselves; we
focus specically on the domain of comprehending and iterating
upon layout designs.
3 FORMATIVE STUDIES
To better understand how BVI people engage in layout design, we
conducted semi-structured interviews with two low vision and ve
blind participants (Table 1). Additionally, we draw on the expe-
riences of the second and third authors, who are both blind. We
returned to three participants to conduct contextual inquiries where
we observed them creating and editing layouts for presentation
slides and websites, the domains our interviewees expressed were
the most challenging.
Procedure
We recruited participants by contacting local community organiza-
tions, email lists, and snowball sampling. We conducted an hour
long semi-structured interview with each participant that high-
lighted the challenges in accessing spatial information without
tactile aids. We revisited the oces of P2, 3, and 6 for contextual
inquiries where we observed them creating a website (P2) and/or
slide deck (P2, P3, P6). Data consisted of audio and video record-
ings, eld notes, and presentations and websites participants sent
us. We used a modied grounded theory [
12
] approach to code our
interview ndings. Our code categories t into two main themes:
participants’ current practices of layout design and the challenges
they encountered. We used this framing to help organize insights
from the contextual inquiries, and present our nal code categories
as the subsubsections below.
Findings: Current Practices
Workflows. Our participants created layouts mainly in the context
of their work—everyone had experience making slide decks, and
P2 and P7 maintained their own websites. Participants began with
content in mind and would either outline it in a word processor to
then paste into a layout template (P2), or write their content directly
into the template (P3, 6), and then speak or magnify the content
to verify their actions. All participants made their slide decks in
PowerPoint, except for P1 who used Google Slides. P6, who is
low vision, used memorized combinations of mouse movements to
access menu options in PowerPoint, while blind participants used
their keyboards to tab through and select the options. After using
their keyboard to select a text box identied by its placeholder
text, blind participants would type their own content and not make
further changes themselves. For websites, P2, who knew how to
program, edited HTML and CSS les directly. P7 used WordPress
for its interface that enabled selecting and creating widgets (like
images or post comments) from the keyboard. This is consistent
with the second author’s own experiences of editing websites, who
also used to directly edit HTML les, but prefers the higher level
interactions of WordPress.
Choosing Templates. Participants chose templates from higher level
descriptors, if they chose a template besides the default one at all.
For example, P7 searched for WordPress themes that used the “ac-
cessible” keyword, which mirrors the second author’s experiences.
Because screen readers do not convey graphical information like
the typography or color palette of templates, some participants re-
lied on sighted colleagues to choose a “pretty enough” or “socially
acceptable” theme. P2 mentioned, “I don’t feel condent in experi-
menting in choosing new layouts, because I would need someone
to explain to me what it looked like every time. I just go with the
default—that people have already described—because I know it.” P3
and P6 sent us PowerPoint presentations they had created without
external help, and both used the default template of black Calibri
text on a white background.
Modifications. During our contextual inquiries, blind participants
did not modify templates beyond entering text. They mentioned
when they wanted to add new non-textual content (such as im-
ages), they would ask someone sighted. Outside of content-based
modications, participants expressed a desire to control graphical
components of layouts that they believed would impact sighted
peoples’ judgments of their sites. This was aligned with the experi-
ences of the second author, who would ask sighted colleagues to
help choose new color palettes for his websites, because he wants
them to look appealing and professional, and not just “another
website which uses the same old template.” Low vision participants
modied themes to be even more accessible. P7, who runs a low
vision advocacy blog, dramatically increased the font size in her
theme, and P6 placed colorful widgets on her slides as place markers
so she would know where she was during her presentation, since
the text was too small for her to read.
Findings: Challenges
Lack of feedback. The biggest challenge blind participants faced
in understanding and editing layouts was that screen readers did
not provide information about the spatial relationships encoded in
layout elements, such as position and size. Instead, they experienced
layouts through serialized audio descriptions of textual content. P5
mentioned that his modifications to templates was “absolutely a
verification problem.” P3 echoed these sentiments, saying, “Even if
I have the ability to make a change, I don’t know how it actually
affected the layout. I just trust the computer did it, or ask someone.”
ID Gender Age Vision level and onset Desktop accessibility features Layouts made & frequency
1 Male 22
Born blind with light
perception
Screen reader
Word documents (weekly),
Slides (on occasion)
2 Male 23 Born blind Screen reader
Ocial documents (monthly),
Websites, Slides (on occasion)
3 Female 61
Born blind with light percep-
tion, total vision loss at 50
Screen reader Slides (on occasion)
4 Female 51 Born blind Screen reader
Slides (thrice a week),
Word documents (weekly)
5 Male 52 Blind at age 1 Screen reader Slides (on occasion)
6 Female 28 20/450 at age 17
Screen magnication &
text-to-speech
Slides, Graphs (weekly),
Websites, Fliers (on occasion)
7 Female 21 Low vision at age 3
Screen magnication & reader, high
contrast display, blue light lter
Websites (weekly),
Slides (on occasion)
Table 1: Participant demographics. Participants 1-5 are blind, and 6 & 7 are low vision.
Screen readers sometimes inadequately communicated the re-
sults of blind participants’ editing actions. For instance, we noticed
text overowing out of its container in a website P2 had designed
and edited, because he accidentally deleted a closing tag in the
HTML and his screen reader did not report the error. Similarly,
when the second author added a video inside WordPress’s content
editor, it was automatically resized to t the full page width. He
immediately asked, “How does it look, visually?” and assumed the
video was small and centered on the page. Without proper feedback,
the mental representations BVI authors held of the layout were
mismatched with the actual layout.
Lack of exposure to design standards. Because they have mainly
experienced web pages as a spoken ordered list of page content,
congenitally blind participants did not have opportunities to build
up their knowledge of standard layouts and design trends. P1 shared,
“Before I try to make my own PowerPoints pretty, I need to ask
what people think is pretty rst.” In the instance of the t to width
video, the second author was concerned visitors would be o put by
the large video player, asking if it “looked abnormal,” even though
t to width media have become standard in current design trends.
Furthermore, the kinds of information screen readers exposed, such
as font names, were unhelpful if participants didn’t know what the
font looked like in the rst place. P2 wished for a “design score” that
evaluated the page’s adherence to graphic design principles, and
P6 specically asked sighted colleagues if her slides were visually
appealing or not. These mirrored the experiences of the third author,
who mentioned, “What I care about is, does the font choice make
the page look childish, or cramped?”
High load on short-term memory. Screen readers translate the par-
allel experience of vision into a serial experience of speech. While
this serialization is adequate for accessing individual elements, it
makes exposing and understanding relationships between elements
challenging due to all the information BVI people must hold in their
short-term memory. The third author described his experiences
with interpreting dense modern web pages through a screen reader
as “playing chess in [his] head.” Participants with low vision who
used screen magniers also faced the issue of being unable to under-
stand the gestalt of their layouts, because big picture relationships
are lost when zooming in.
Sighted collaborators. Several participants had experience making
layouts in collaborative contexts (i.e., P6 was an editor of an aca-
demic magazine and P2 was a start up co-founder and had to make
slide decks to raise venture capital). While they trusted their sighted
collaborators to graphically design the content they provided, their
collaborators sometimes lacked the proper domain knowledge to
suciently visually communicate content only they understood,
such as making graphs (P6) or technical diagrams (P2). P7 also men-
tioned the social strife that could result from creative dierences
with collaborators. “When you ask someone else to do it, you’re
reliant on their judgment and understanding and tastes, which usu-
ally aren’t yours. And sometimes they take it personally, and then
you’re in a social mess.”
Confidence. None of our blind participants could move, align, or
resize content in their layout with a screen reader without external
help. We observed low vision participants using menu options and
buttons to align elements; P6 said she avoided direct manipulation
with the mouse for she feared her interactions inaccurate. This lack
of condence that they would not “mess up” the layout prevented
BVI users from editing content in the rst place. P2 said, “I lack the
ability to make last minute changes—I would never make a change
to my presentation without having a sighted person double check
it, not even if it’s as small as changing a word. I just don’t want to
screw it up.”
Design Imperatives
In summary, like sighted people, BVI people utilize templates in
creating layouts. However, their current assistive technologies re-
strict them to filling in content and making high level modifications
through PowerPoint or WordPress menus, like changing color or
type. Because they lack exposure to and feedback on layout elements,
BVI users express a desire to understand what “good” layout designs
are. They often juggle many elements in their short term memory
and have mental models differing from what they are designing.
BVI authors rely on sighted collaborators for assistance, which they
mentioned reduced their confidence in designing independently.
Based on these ndings, we developed several design guidelines
for our tool to help BVI authors in designing and understanding
layouts. We want to (1) leverage their existing workows of starting
from templates (i.e., by adapting our tool to work with existing tem-
plates); (2) provide feedback on and the ability to make spatial edits
(i.e., through interactive tactile sheets); (3) present content-layout
relationships in multiple modalities to avoid high cognitive loads
(i.e., by ooading some of the screen reader’s audio to the tactile
channel); and (4) support learning of unfamiliar layout designs and
concepts (i.e., by exposing options in the context of editing).
4 AN ACCESSIBLE LAYOUT DESIGN TOOL
Our tool enables users understand layout structures and make
modications in both page content and layout. It uses automatically
generated tactile sheets overlaid on a tablet that speaks and edits
HTML templates. We support two dierent layout representations
of these templates—pixel and inated—that respectively conform
to the original web page design and tactile graphics guidelines.
While our formative studies revealed insights both in domains
of slide and web design, we chose to build a web-editing tool for
a few reasons. First, the web is a platform with a rich tradition of
broadening access to information, and web layouts have complex
issues that are widely applicable to other layout tasks (which we
expand on in the discussion.) Second, web pages have an underlying
textual declarative representation (HTML/CSS) for visual layouts,
while this representation is hidden in other domains like closed-
source slide deck software. Our interface allows users to edit layouts
by manipulating this underlying representation, which generalizes
to a platform more widely used than creating our own specialized
software representations. Finally, we wanted to leverage building a
tool considering our co-designer’s interests and domain expertise.
Although our formative study participants had a variety of vi-
sual impairments, to address the issues brought forth by screen
readers, we frame our tool specically for people who are blind. We
imagine our tool to also be useful for people with low vision, but
acknowledge the wide spectrum of visual impairments and push
against a “one-size-ts-all” solution.
Co-Design Process
We closely designed our tool with co-author Son Kim, an Assistive
Technology specialist who is also a blind hobbyist web designer.
We iterated upon three design probes before arriving at our current
implementation. For the rst probe, we generated three sets of
tactile sheets by hand and encoded dierent page elements—like
paragraphs versus images—with dierent textures. We chose to use
tactile sheets because blind people use two hands with multiple
ngers while exploring tactile layouts [
32
], and the sheets aord
multiple nger and whole hand exploration. Our goals were to
verify that tactile sheets would lead to an adequate mental map of
spatial relationships as others have demonstrated [
3
], and to under-
stand how BVI designers might chose layouts with this additional
understanding over solely high-level tags. Son recreated layouts on
a felt board to demonstrate his understanding, and discussed with
the researchers where perceptual challenges occurred, which were
usually around closely-spaced elements.
In the second probe, we generated tactile sheets of two website
templates and mounted them on a tablet that provided audio an-
notations. We presented the tool without editing features to focus
on feedback on how it helped with layout comprehension. Son
took about 10 minutes to understand each layout, and suggested
creating a tactile sheet representation that exposed the web page’s
underlying grid structure and to help him more quickly identify
where elements were. We incorporated this functionality in reveal-
ing container relationships in the tool’s “inated" mode.
Finally, our goal in the last probe was to identify a set of edit-
ing actions to support. We presented ve web page templates to
modify. Son demonstrated what kinds of interactions and edits he
wanted to make, which we carried out with through Wizard of Oz
manipulations of the underlying website by changing the DOM in
real time and refabricating new tactile sheets to reect his edits.
Three themes emerged through the co-design sessions. First, legi-
bility diculties in the pixel-perfect layout representation reempha-
sized the importance of tactile graphics guidelines such as leaving
a minimum 1/8th inch between elements [
16
]. Second, with spatial
relationship knowledge, Son was able to assign semantic meaning
to the elements—for instance, he chose a layout in the rst probe
for a storefront because the large space allocated for an image made
it “showy and screamed, ‘Buy this now!”’ Third, breakdowns oc-
curred due to a lack of past exposure and contextual knowledge of
components (such as unfamiliarity with z-ordering or pagination),
mirroring our formative study ndings.
Understanding and Editing Layouts with Our Tool
We describe our tool with a hypothetical design scenario. Ben,
a blind student, wishes to make a personal portfolio website to
showcase his projects before applying to jobs. He starts by searching
for templates using tags, such as “accessible” and “portfolio,” and
downloads a prospective HTML le, akin to his current workow,
following from design imperative (1).
Providing access to spatial components and relationships. Ben loads
the rst HTML le (Figure 2A) into our tool on his tablet and selects
the Print button in the toolbar (Figure 3). This downloads an SVG
le that is the pixel representation of the layout—that is, bounding
box outlines of page elements (Figure 2B). He prints this SVG le on
capsule paper and runs the paper through a capsule paper printer to
generate raised lines. He axes this new tactile sheet to his tablet.
Ben now explores this web page by running his ngers over
the capsule paper. Ben double taps on a large element on the left
side of the layout, and he hears information about the element:
Element: image. Alternate text: Project 1 image placeholder. He then
taps below that, Element: Heading level 2. Content: Project 1. Text
align: left. This interaction follows from design imperative (2), and
splitting layout information across modalities—structure to touch
and meta information to audio—supports design imperative (3).
After exploring the other two templates in this fashion, Ben chooses
the rst layout because he likes that the layout isn’t too cluttered
and that it has partitioned sections for both his personal information
and portfolio.
Editing content and layout. First, Ben lls in his content text by
double tapping the Edit button in the tool bar to toggle on edit mode.
In edit mode, Ben rst selects the element he wants to modify—
here, the Heading level 1 to replace “Your Name” with “Ben.” He
double taps the Back and Next buttons to scroll through a menu
of operations, such as editing, deleting, swapping, or duplicating
Figure 2: A sample HTML template (A) and our tool’s automatically generated pixel (B) and inated (C) representations, fab-
ricated by laser cutting capsule paper. We also generate small multiples (D); here, we show several levels of bottom margin
between the project header and description. Note the outlines of the bounding b oxes are for gure purposes only.
Figure 3: A menu appears at the top of the design tool. Users
double tap the menu options to change the tool’s modality
and cycle through editing options.
the element, and traverses through the menu hierarchy: Selected
Heading level 1
→
Edit element
→
Edit text. After Ben enters text
using a keyboard connected to the tablet, he also lls in the images
he had planned for his website by specifying their URLs.
With the content lled in, Ben now wants to modify the lay-
out structure—he has three projects he wants to showcase, but the
template came with space for two. He toggles the print mode to
generate an inated representation of the layout that exposes ele-
ments’ containers (Figure 2C). This representation adds padding to
container elements to comply with tactile graphic guidelines [
16
],
including a minimum 1/8th distance between lines, and exposes the
underlying grid structure of the page. He selects an HTML column
(Figure 4.1), chooses the duplicate action, and then selects the new
container to duplicate the column in—in this case, its row (Figure
4.2). Our tool supports adaptive layouts and asks Ben if he would
like to keep the current column sizes, or shrink them to keep all
the elements on the same row; he selects the latter (Figure 4, after).
Now there is a mismatch in representation between the underlying
layout and tactile sheet, so Ben reprints and replaces the overlay,
which he keeps track of by a braille label at the top.
Ben now notices that his project heading is too closely spaced
to its paragraph description. He wants to add a bottom margin
to the element, but isn’t sure about what value would best match
the existing layout design. With the Edit margin action selected,
Ben double taps the Print button to generate a tactile sheet of a
preview gallery. This gallery contains multiple copies of his project
text description container (with the heading, paragraph text, and
“More Details” link) with dierent levels of bottom margin applied
to the heading (Figure 2D). Following design imperative (4), the
preview gallery aims to expose potentially unfamiliar concepts
situated in the context of the page. Ben applies a wider margin to
the heading, and then selects the heading and uses the Copy style
action to propagate the changes to the other two project headings.
Finally, Ben prints a pixel representation of his layout to verify how
his edits would appear on most screens.
Implementation
Our design tool relies on manipulating the web page’s Document
Object Model (DOM). To generate tactile sheets, our tool traverses
the DOM, gets an object’s bounding box coordinates, creates an SVG
rectangle. Generated tactile sheets have four dierent line patterns:
solid for elements (Figure 4D), dashed for general containers (4C),
dot-dashed for columns (4B), and dotted for rows (4A). Elements
are considered containers if their inner text matches the sum of
their children’s inner text. Notably, we cannot simply just consider
non-leaf nodes containers to avoid common false positives, such as
marking a paragraph a container if it has a link child. To generate
an inated layout, we add 1/8th inch of padding to any container
element whose padding is currently less than that. Because we
use responsive templates, increasing padding does not horizontally
overow elements at the tradeo of increased element heights. To
generate in-context preview galleries of small multiples, we extract
and modify the selected element in its parent container, applying
all available levels of padding, margin, alignment, or resizes.
Once users download the SVG, they can then fabricate a tactile
sheet. Users calibrate sheets to their devices by specifying their
screen width and a height oset. We used a laser cutter to fabricate
sheets, which takes an average of 10 to 120 seconds per sheet
depending on the complexity. However, sheets may be generated
on other 2-axis CNC machines, like an embosser, or with a capsule
paper printer, which requires a short xed time (about 15 seconds)
regardless of layout complexity. Once the sheets are fabricated,
users ax them to a tablet; we attach ours to an iPad Pro by securing
them under a beveled case.
Actions our tool currently supports include creating, deleting,
duplicating, and swapping content elements or their containers,
which we refer to collectively as layout components. Users can
edit layout components by editing their text content or tags (e.g.,
changing a paragraph to a heading), resizing them, aligning them
Figure 4: In the inated representation, we expose HTML rows (A), columns (B), other containers (C), and content elements
(D). Before: To duplicate a column, rst users select the element to be duplicated (the column) (1) and then in the container to
duplicate it in (the row) (2). After: Users have the option to adaptively resize the new columns in the resultant row.
Figure 5: Our system takes an HTML template as input and
allows users to iteratively edit templates and reprint and
replace them as tactile sheets. Once users are done editing,
they can export the edited HTML le.
within their container, text aligning them, editing their padding or
margin, or copying their style to another layout component. We
chose these actions based on the results of our co-design process.
Our tool is a Node.js web application. It takes an HTML le as
input, and outputs both pixel and inated SVG representations of
our tactile sheets and the edited HTML le (Figure 5). We employ
the ResponsiveVoice.js library for speech feedback, Hammer.js to
handle multitouch interactions, and Bootstrap as our templating
framework.
5 RESULTS
We tested our tool against popular website and web-based slide
templates and show some of the generated tactile sheets in Figure 6.
In addition to developing our tool with an experienced web designer,
we veried its usability and received positive informal feedback
from non-domain experts.
Informal Evaluation
Since our tool was co-designed with an experienced web designer,
we wanted to understand how the tool helps lower the threshold
for layout comprehension and design with novices. We recruited
two blind participants with no experience in creating layouts for
informal feedback. Neither participated in the formative studies
nor considered themselves familiar with tactile graphics. P1 lost
his vision nine years ago and has been using a screen reader on his
desktop computer since; P2 lost her vision at eight months and did
not regularly use a desktop computer, although she began using a
smart phone three months prior.
Participants underwent one hour of training with our tool and
one hour of a design task. During training, the proctor guided
them through all of the editing features provided. For the design
task, participants rst explored and chose from four Bootstrap
templates for personal portfolio websites, and then lled in content
and modied the layout.
Both participants were able to successfully identify and recall
layout elements, as well as populate them with content and make
structural layout modications. A template took from 15 to 20
minutes to fully comprehend. Participants initially had trouble
decoding the dierent dashed line representations for containers,
but learned after hearing the container type spoken a few times.
Both participants swapped tactile sheets when their edits resulted
in a large change in the page structure—like inserting an image
that was longer than its placeholder space—but were ne with
relying on audio feedback for actions like alignment and swapping
elements that were the same size. P2 preferred the inated layout
representations and said the wider spaces between elements were
more helpful to understanding “which boxes were the walls of the
room, and which ones were the furniture.” Both participants found
the design galleries helpful in understanding unfamiliar concepts
like padding. “Because I added another picture, I would want to
decrease the padding between them to loosen up the page,” said P1.
Despite their lack of layout experience, both participants found
the tool useful. “It’s a lot more detailed than I expected, but I like
that level of specic control. With more time to familiarize myself
prior work goes from being an impossibility to being possible,” said
P1. “I thought editing a website would be dicult because I didn’t
know anything about web pages and sites, but I actually enjoy
editing. I feel more control when I can place in images. I think I
understand a lot more about web pages now and am more condent
in my ability to make one,” said P2.
6 DISCUSSION AND FUTURE WORK
Limitations
People who are blind or visually impaired have diverse abilities
and experiences. It is important to note that all our formative study
interviewees had (or was working towards) a college degree and
were experienced users of accessible desktop technology.
Our informal evaluation also had a small sample size. Future
work should focus on a longitudinal “in the wild” deployment with
blind web designers.
Figure 6: Sample templates taken from popular themes and their corresponding pixel and inated tactile representations. A
and B are Wordpress themes, C and E are Bootstrap themes, and D is a sample slide from Remark.js, an HTML-based slide
library. Note that since template E did not have containers that needed to be inated, its representations are equivalent.
While our tool exposes layout structures, it does not commu-
nicate other design aspects like typography or color. Translating
these elements requires more research on choosing the appropri-
ate descriptors that facilitate their understanding for congenitally
blind people. Our tool aords iterative feedback and editing, but
the time it takes to fabricate and replace the tactile sheets prevents
an interactive workow. A refreshable tactile display [
13
,
38
] could
bypass this issue.
Our algorithm to generate tactile sheets sometimes falsely ex-
poses span elements that are used for stylistic markers (e.g., chang-
ing a font color) but irrelevant to spatial layout design (for example,
the last name in Figure 6E.) Since it relies on the bounding box
coordinates of elements, elements with bounding boxes that stretch
to ll to their parent but with sparse content appear to take up more
space than they actually do (for example, the footer text in Figure
6C.) These cases could be handled with a vision-based approach.
Design Considerations of Tools for BVI Designers
Through our co-design sessions and informal evaluations, sev-
eral themes emerged which we believe to be valuable to other
researchers developing authoring tools for BVI users: the trade-os
between pixel-perfect layout representations and tactile graphics,
how to support a wide range of abilities and backgrounds in de-
veloping domain-specic tools, and building tools that recognize
the valuable rst-hand perspectives users of assistive technologies
bring to their designs.
Pixel-Perfect Views and Tactile Graphic Guidelines. When gener-
ating tactile sheets, there is a tension between representing the
structure of a web page as a direct pixel translation of how it ap-
pears in the browser and modifying it to conform to tactile graphics
standards. Many child nodes share the same borders as their parent
containers, or are adjacent to other elements. These overlapping
borders are dicult to dierentiate tacitly—all users confused neg-
ative spaces (like the space between a navigation bar link and the
end of the bar) as small layout elements elements. Adding more
space between element borders addresses this issue; however, in-
creasing spacing leads to either overow (with increased margins)
or taller, “squished” content (with increased padding). Our tool
chooses the latter to preserve column structures central to modern
web templates as to avoid unnecessary layout re-comprehension.
One possible intervention is to only fabricate content elements and
present the container information in audio—for example, elements
could also speak the names of their parent containers. Another
option is to allow zooming in to the layout and add spacing at a
ner granularity for less disruption. Although more research needs
to be done to evaluate the trade-os of dierent tactile representa-
tions for layout design for BVI users, our ndings can inform the
design of tools built on refreshable tactile displays, which would be
particularly well suited for zooming and quickly toggling between
layout representations.
Scaolding for Novice Designers. As evidenced by our informal
evaluation, people have a wide range of visual impairments and
familiarity with domain-specic concepts. Our co-designer noticed
patterns of organizational structures throughout our sessions, such
as template elements being arranged side-by-side in the same row,
and not top-to-bottom as he assumed was the case since he used
the top and bottom arrow keys to access elements with his screen
reader. However, those without web design experience struggled
with understanding web-specic jargon, like “div.” In addition to
generating preview galleries that explain some terms in context, our
tool could come with a binder of simple examples that emphasize
dierent principles, or include training methods grounded in best
practices of teachers of the visually impaired.
We designed our tool to facilitate the existing practices of users,
allowing them to select and modify individual elements at the same
level as their screen readers. However, web design novices found
understanding our lower-level representations of a layout’s underly-
ing grid structure challenging. Due to concerns about the accuracy
of direct manipulation from the formative studies and past work
[
7
], we avoided pixel-based manipulations and computationally
abstract direct editing of CSS or HTML. High-level design tools
have supported sighted novice designers [
27
,
35
], and we believe
the same applies for users who are BVI. Our tool could oer global
heuristic-based design “sanity checks,” or execute edits following
high-level commands like “make this layout less cluttered.”
Supporting Edits for Accessibility. We observed users making edits
with accessibility in mind—for instance, Son swapped the image
and text columns so screen reader users would be able to have an
image description rst. While we have developed a tool for blind
people that represents how sighted individuals perceive websites,
we do not yet support how blind people could design websites for
other screen reader users. For example, a third layout representation
could expose a screen reader’s reading order of page elements, and
allow modications to this order for a better user experience.
7 CONCLUSION
We present the challenges people who are blind or visually impaired
face in understanding and editing spatial layouts, and address these
challenges by developing a multimodal tool to help blind people
understand and edit web pages by using tactile templates. Drawing
from our experiences co-designing the tool with a blind hobbyist
web developer and feedback from two blind web design novices, we
point out the tensions in representing spatial layouts with tactile
sheets and accommodating for a wide range of expertise. We believe
BVI users bring valuable accessibility-oriented perspectives to web
design and push for greater inclusion of their voices and insights.
ACKNOWLEDGMENTS
We wish to thank the formative study and informal evaluation
participants for their valuable time and insights, as well as Alexa
Siu, Elyse Chase, Eric Gonzalez, Evan Strasnick, Jennifer Jacobs,
and Niamh Murphy for their helpful conversations and assistance.
This work is supported by an NSF GRFP Fellowship #2017245378.
REFERENCES
[1]
Nacny Amick and Jane Corcoran. 1997. Guidelines for Design of
Tactile Graphics. http://www
.
aph
.
org/research/guides/. Accessed:
2018-09-06.
[2]
Sona K Andrews. 1985. The use of capsule paper in producing tactual
maps. Journal of Visual Impairment & Blindness (1985).
[3]
Mauro Avila, Francisco Kiss, Ismael Rodriguez, Albrecht Schmidt, and
Tonja Machulla. 2018. Tactile Sheets: Using Engraved Paper Over-
lays to Facilitate Access to a Digital Document’s Layout and Logical
Structure. In Proceedings of the 11th PErvasive Technologies Related to
Assistive Environments Conference (PETRA ’18). ACM, New York, NY,
USA, 165–169. https://doi. org/10. 1145/3197768. 3201530
[4]
Mark S. Baldwin, Gillian R. Hayes, Oliver L. Haimson, Jennifer Manko,
and Scott E. Hudson. 2017. The Tangible Desktop: A Multimodal
Approach to Nonvisual Computing. ACM Trans. Access. Comput. 10, 3,
Article 9 (Aug. 2017), 28 pages. https://doi.org/10. 1145/3075222
[5]
Edward C. Bell and Natalia M. Mino. 2015. Employment outcomes for
blind and visually impaired adults. (2015).
[6]
Edward C. Bell and Natalia M. Mino. 2016. Blogging when you’re
blind or visually impaired. https://yourdolphin
.
com/news?id
=
223.
Accessed: 2018-08-30.
[7]
Jens Bornschein, Denise Bornschein, and Gerhard Weber. 2018. Com-
paring Computer-Based Drawing Methods for Blind People with Real-
Time Tactile Feedback. In Proceedings of the 2018 CHI Conference on
Human Factors in Computing Systems (CHI ’18). ACM, New York, NY,
USA, Article 115, 13 pages. https://doi
.
org/10
.
1145/3173574
.
3173689
[8]
Erin Brady, Meredith Ringel Morris, Yu Zhong, Samuel White, and
Jerey P. Bigham. 2013. Visual Challenges in the Everyday Lives of
Blind People. In Proceedings of the SIGCHI Conference on Human Factors
in Computing Systems (CHI ’13). ACM, New York, NY, USA, 2117–2126.
https://doi. org/10. 1145/2470654. 2481291
[9]
Anke M Brock, Philippe Truillet, Bernard Oriola, Delphine Picard,
and Christophe Jourais. 2015. Interactivity improves usability of
geographic maps for visually impaired people. Human–Computer
Interaction 30, 2 (2015), 156–194.
[10]
Craig Brown and Amy Hurst. 2012. VizTouch: Automatically Gener-
ated Tactile Visualizations of Coordinate Spaces. In Proceedings of
the Sixth International Conference on Tangible, Embedded and Em-
bodied Interaction (TEI ’12). ACM, New York, NY, USA, 131–138.
https://doi. org/10. 1145/2148131. 2148160
[11]
Erin Buehler, Shaun K. Kane, and Amy Hurst. 2014. ABC and 3D:
Opportunities and Obstacles to 3D Printing in Special Education En-
vironments. In Proceedings of the 16th International ACM SIGACCESS
Conference on Computers & Accessibility (ASSETS ’14). ACM, New York,
NY, USA, 107–114. https://doi. org/10.1145/2661334. 2661365
[12]
Kathy Charmaz and Linda Liska Belgrave. 2007. Grounded theory. The
Blackwell encyclopedia of sociology (2007).
[13]
American Printing House for the Blind. 2016. American
Printing House for the Blind and Orbit Research Announce
the World’s First Aordable Refreshable Tactile Graphics Dis-
play. https://www
.
aph
.
org/pr/aph-and-orbit-research-announce-
the-worlds-rst-aordable-refreshable-tactile-graphics-display/
[14]
John A Gardner, Vladimir Bulatov, and Holly Stowell. 2005. The View-
Plus IVEO technology for universally usable graphical information. In
Proceedings of the 2005 CSUN International Conference on Technology
and People with Disabilities.
[15]
Anhong Guo, Jeeeun Kim, Xiang ’Anthony’ Chen, Tom Yeh, Scott E.
Hudson, Jennifer Manko, and Jerey P. Bigham. 2017. Facade:
Auto-generating Tactile Interfaces to Appliances. In Proceedings of
the 2017 CHI Conference on Human Factors in Computing Systems (CHI
’17). ACM, New York, NY, USA, 5826–5838. https://doi
.
org/10
.
1145/
3025453. 3025845
[16]
Lucia Hasty. 2018. Guidelines for Design of Tactile Graphics. http:
//www. tactilegraphics. org/index. html. Accessed: 2018-09-06.
[17]
Liang He, Zijian Wan, Leah Findlater, and Jon E. Froehlich. 2017.
TacTILE: A Preliminary Toolchain for Creating Accessible Graphics
with 3D-Printed Overlays and Auditory Annotations. In Proceedings
of the 19th International ACM SIGACCESS Conference on Computers
and Accessibility (ASSETS ’17). ACM, New York, NY, USA, 397–398.
https://doi. org/10. 1145/3132525. 3134818
[18]
Charles Jacobs, Wilmot Li, Evan Schrier, David Bargeron, and David
Salesin. 2003. Adaptive Grid-based Document Layout. ACM
Trans. Graph. 22, 3 (July 2003), 838–847. https://doi
.
org/10
.
1145/
882262. 882353
[19]
Ali Jahanian, Jerry Liu, Qian Lin, Daniel Tretter, Eamonn O’Brien-
Strain, Seungyon Claire Lee, Nic Lyons, and Jan Allebach. 2013.
Recommendation System for Automatic Design of Magazine Cov-
ers. In Proceedings of the 2013 International Conference on Intelli-
gent User Interfaces (IUI ’13). ACM, New York, NY, USA, 95–106.
https://doi. org/10. 1145/2449396. 2449411
[20]
Shaun K. Kane and Jerey P. Bigham. 2014. Tracking @Stemxcomet:
Teaching Programming to Blind Students via 3D Printing, Crisis Man-
agement, and Twitter. In Proceedings of the 45th ACM Technical Sym-
posium on Computer Science Education (SIGCSE ’14). ACM, New York,
NY, USA, 247–252. https://doi. org/10.1145/2538862. 2538975
[21]
Shaun K. Kane, Meredith Ringel Morris, and Jacob O. Wobbrock. 2013.
Touchplates: Low-cost Tactile Overlays for Visually Impaired Touch
Screen Users. In Proceedings of the 15th International ACM SIGAC-
CESS Conference on Computers and Accessibility (ASSETS ’13). ACM,
New York, NY, USA, Article 22, 8 pages. https://doi
.
org/10
.
1145/
2513383. 2513442
[22]
Jeeeun Kim and Tom Yeh. 2015. Toward 3D-Printed Movable Tactile
Pictures for Children with Visual Impairments. In Proceedings of the
33rd Annual ACM Conference on Human Factors in Computing Systems
(CHI ’15). ACM, New York, NY, USA, 2815–2824. https://doi
.
org/
10. 1145/2702123. 2702144
[23]
Mikko Kuhna, Ida-Maria Kivelä, and Pirkko Oittinen. 2012. Semi-
automated Magazine Layout Using Content-based Image Features. In
Proceedings of the 20th ACM International Conference on Multimedia
(MM ’12). ACM, New York, NY, USA, 379–388. https://doi
.
org/10
.
1145/
2393347. 2393403
[24]
Ranjitha Kumar, Arvind Satyanarayan, Cesar Torres, Maxine Lim,
Salman Ahmad, Scott R. Klemmer, and Jerry O. Talton. 2013. Webzeit-
geist: Design Mining the Web. In Proceedings of the SIGCHI Conference
on Human Factors in Computing Systems (CHI ’13). ACM, New York,
NY, USA, 3083–3092. https://doi. org/10.1145/2470654. 2466420
[25]
Steven Landau and Lesley Wells. 2003. Merging tactile sensory input
and audio data by means of the Talking Tactile Tablet. In Proceedings
of EuroHaptics, Vol. 3. 414–418.
[26]
Veronica Lewis. 2018. Veronica With Four Eyes. https://
veroniiiica. com/. Accessed: 2018-08-30.
[27]
Jason Linder, Gierad Laput, Mira Dontcheva, Gregg Wilensky, Walter
Chang, Aseem Agarwala, and Eytan Adar. 2013. PixelTone: A Multi-
modal Interface for Image Editing. In CHI ’13 Extended Abstracts on
Human Factors in Computing Systems (CHI EA ’13). ACM, New York,
NY, USA, 2829–2830. https://doi. org/10.1145/2468356. 2479533
[28]
Ellen Lupton. 2004. Thinking with type. Critical Guide for Designers,
Writers, Editors & Students (2004).
[29]
Nolwenn Maudet, Ghita Jalal, Philip Tchernavskij, Michel Beaudouin-
Lafon, and Wendy E. Mackay. 2017. Beyond Grids: Interactive Graphi-
cal Substrates to Structure Digital Layout. In Proceedings of the 2017
CHI Conference on Human Factors in Computing Systems (CHI ’17).
ACM, New York, NY, USA, 5053–5064. https://doi
.
org/10
.
1145/
3025453. 3025718
[30] David McCandless. 2012. Information is beautiful. Collins London.
[31]
Samantha McDonald, Joshua Dutterer, Ali Abdolrahmani, Shaun K.
Kane, and Amy Hurst. 2014. Tactile Aids for Visually Impaired
Graphical Design Education. In Proceedings of the 16th International
ACM SIGACCESS Conference on Computers & Accessibility (ASSETS
’14). ACM, New York, NY, USA, 275–276. https://doi
.
org/10
.
1145/
2661334. 2661392
[32]
Valerie S Morash, Allison E Connell Pensky, Steven TW Tseng, and
Joshua A Miele. 2014. Eects of using multiple hands and ngers
on haptic performance in individuals who are blind. Perception 43, 6
(2014), 569–588.
[33]
Valerie S Morash, Alexander Russomanno, R Brent Gillespie, and Sile
O’Modhrain. 2017. Evaluating Approaches to Rendering Braille Text
on a High-Density Pin Display. IEEE transactions on haptics (2017).
[34]
Josef Müller-Brockmann. 1996. Grid systems. Alemanha, Editora Braun
(1996).
[35]
Peter O’Donovan, Aseem Agarwala, and Aaron Hertzmann. 2015.
DesignScape: Design with Interactive Layout Suggestions. In Pro-
ceedings of the 33rd Annual ACM Conference on Human Factors in
Computing Systems (CHI ’15). ACM, New York, NY, USA, 1221–1224.
https://doi. org/10. 1145/2702123. 2702149
[36]
Braille Authority of North America. 2010. Guidelines and Standards
for Tactile Graphics. http://www. brailleauthority. org/tg/. Accessed:
2018-09-06.
[37]
M Sile OâĂŹModhrain and Brent Gillespie. 1997. The moose: A haptic
user interface for blind persons. In Proc. Third WWW6 Conference.
[38]
Denise Prescher, Jens Bornschein, Wiebke Köhlmann, and Gerhard
Weber. 2018. Touching graphical applications: bimanual tactile inter-
action on the HyperBraille pin-matrix display. Universal Access in the
Information Society 17, 2 (2018), 391–409.
[39]
Andreas Reichinger, Helena Garcia Carrizosa, Joanna Wood, Svenja
Schröder, Christian Löw, Laura Rosalia Luidolt, Maria Schimkowitsch,
Anton Fuhrmann, Stefan Maierhofer, and Werner Purgathofer. 2018.
Pictures in Your Mind: Using Interactive Gesture-Controlled Reliefs to
Explore Art. ACM Trans. Access. Comput. 11, 1, Article 2 (March 2018),
39 pages. https://doi. org/10. 1145/3155286
[40]
Daniel Ritchie, Ankita Arvind Kejriwal, and Scott R. Klemmer. 2011.
D.Tour: Style-based Exploration of Design Example Galleries. In Pro-
ceedings of the 24th Annual ACM Symposium on User Interface Soft-
ware and Technology (UIST ’11). ACM, New York, NY, USA, 165–174.
https://doi. org/10. 1145/2047196. 2047216
[41]
Jana Sedivy and Hilary Johnson. 1999. Supporting Creative Work Tasks:
The Potential of Multimodal Tools to Support Sketching. In Proceedings
of the 3rd Conference on Creativity & Cognition (C&C ’99). ACM,
New York, NY, USA, 42–49. https://doi. org/10. 1145/317561. 317571
[42]
Lei Shi, Yuhang Zhao, and Shiri Azenkot. 2017. Markit and Talkit:
A Low-Barrier Toolkit to Augment 3D Printed Models with Audio
Annotations. In Proceedings of the 30th Annual ACM Symposium on
User Interface Software and Technology (UIST ’17). ACM, New York, NY,
USA, 493–506. https://doi. org/10. 1145/3126594.3126650
[43]
Anoop K. Sinha and James A. Landay. 2003. Capturing User Tests in
a Multimodal, Multidevice Informal Prototyping Tool. In Proceedings
of the 5th International Conference on Multimodal Interfaces (ICMI
’03). ACM, New York, NY, USA, 117–124. https://doi
.
org/10
.
1145/
958432. 958457
[44]
Atau Tanaka and Adam Parkinson. 2016. Haptic Wave: A Cross-Modal
Interface for Visually Impaired Audio Producers. In Proceedings of the
2016 CHI Conference on Human Factors in Computing Systems (CHI
’16). ACM, New York, NY, USA, 2150–2161. https://doi
.
org/10
.
1145/
2858036. 2858304
[45]
Brandon Taylor, Anind Dey, Dan Siewiorek, and Asim Smailagic. 2016.
Customizable 3D Printed Tactile Maps As Interactive Overlays. In
Proceedings of the 18th International ACM SIGACCESS Conference on
Computers and Accessibility (ASSETS ’16). ACM, New York, NY, USA,
71–79. https://doi. org/10. 1145/2982142.2982167
[46]
Kashyap Todi, Daryl Weir, and Antti Oulasvirta. 2016. Sketchplore:
Sketch and Explore with a Layout Optimiser. In Proceedings of the 2016
ACM Conference on Designing Interactive Systems (DIS ’16). ACM, New
York, NY, USA, 543–555. https://doi. org/10. 1145/2901790. 2901817
[47]
Jan Tschichold. 1998. The new typography: A handbook for modern
designers. Vol. 8. Univ of California Press.
[48]
Jacob O. Wobbrock, Shaun K. Kane, Krzysztof Z. Gajos, Susumu Harada,
and Jon Froehlich. 2011. Ability-Based Design: Concept, Principles
and Examples. ACM Trans. Access. Comput. 3, 3, Article 9 (April 2011),
27 pages. https://doi. org/10. 1145/1952383.1952384
