CMSC 20370/30370 Winter 2020 Dealing with constraints Case Study: - - PowerPoint PPT Presentation

CMSC 20370/30370 Winter 2020 Dealing with constraints Case Study: Developing Countries

Feb 10, 2020

Quiz Time (5-7 minutes).

Quiz on SmartSpeakers

Principles of Good Design

Administrivia

• Thank you for the constructive feedback

– Will help improve next iteration of this class – You are the first to take the class and this is the first class

• n making technology inclusive

– 15 quizzes in total, we will drop your lowest 5 scores, quizzes are to jog memory for class/check on readings

• GP1 grades are back
• GP2 is on Wed for feedback

– No need to print poster – Can paste onto poster board – Main thing is to get design alternatives sketched out based

• n any user research to date

– Moved GP2 report due date to Monday 17 Feb

Today’s Agenda

• Finishing off human abilities
• Recap on inclusion
• Dealing with constraints
• Case Study: Developing systems for

developing countries

Case Study: SeeingVR

• Low vision users – not fully blind, blurred vision, light

sensitivity, can’t correct with contact lenses or glasses

• Formative work with low vision users
• SeeingVR prototype of 14 augmentations to VR e.g.

Bifocal lens

• Unity toolkit for integrating SeeingVR into applications
• Evaluation with low vision users and developers
• Limitations

– Small user sample – Biases from apps tested – Within subjects

Learning

• How to learn to use a computer-

based application

• Using a computer-based application

to understand a given topic

• People find it hard to learn by

following instructions in a manual

• prefer to learn by doing

Toddlers and iPads

• Why do you think it works?

Design implications

• Design interfaces that encourage

exploration

• Design interfaces that constrain and

guide learners

• Dynamically linking concepts and

representations can facilitate the learning of complex material

• Case study SeeingVR: Augmentations

are not as useful if there is a steep learning curve

Reading, speaking, and listening

• The ease with which people can read,

listen, or speak differs

– Many prefer listening to reading – Reading can be quicker than speaking/ listening – Listening requires less cognitive effort than reading or speaking

The vision

• Then:

Design implications

• Use natural interactions
• Distinguish from other interactions
• Decide when its appropriate to use

voice vs text vs other forms of interaction

• Privacy and security implications?
• SeeingVR: text to speech augmentation
• Other cases: using sonification

Problem-solving, planning, reasoning and decision-making

• All involves reflective cognition

– e.g. thinking about what to do, what the

• ptions are, and the consequences
• Often involves conscious processes, discussion

with others (or oneself), and the use of artifacts – e.g. maps, books, pen and paper

• May involve working through different

scenarios and deciding which is best option

Design implications

• Provide additional information/functions for users

who wish to understand more about how to carry

• ut an activity more effectively
• Use simple computational aids to support rapid

decision-making and planning for users on the move

• SeeingVR: highlight what’s salient, could help

find clues in a game

Summary of human abilities

• Cognition involves several processes including attention,

memory, perception and learning

• Interface design can greatly affect how well users can

perceive, attend, learn and remember how to do their tasks

• We have only talked about cognition but humans differ on

many axes and we need to accommodate accordingly

• Accessibility is about accommodating differing human

abilities (e.g. cognition, motor, sensory, physical, etc)

• Inclusion is about accommodating humans of different

abilities, ages, cultures, demographics, in different contexts

Recap: why are we here?

• This class is to help you learn about

examples of making technology inclusive

• Most systems are designed for the

“typical” user

– Assumes Internet connection – Assumes literate, educated, informed

• This is not often the case….

What is being marginalized/underserved?

• Here we are looking at cases where the typical user

assumptions break

– E.g. accessibility – -> humans have different abilities – According to the world bank, 15% of all humans have at least

• ne disability

– In the USA, according to the Center for Disease control (CDC) 1 in 4 adults have a disability – You are learning about how to place a user in the center of a design process to make your designs inclusive – The readings are examples of inclusive systems made using this user-centered design process – The group project is to help you try out the process for itself – The individual assignment was to help you think critically about a system design and whether it is inclusive

What examples have we covered…

• Systems for accessibility
• Systems for low income users
• Systems for users in difficult circumstances

such as abuse

Where are we heading…

• Systems for developing countries
• Systems for LGBQT
• Systems for gender
• Systems for workers in the gig economy
• Systems for children
• Systems for addiction and chronic illness
• Asking ourselves the question is

technology the answer?

Today: Developing countries

Source:https://commons.wikimedia.org/wiki/File:Developed_and_developing_countries.PNG and CIA World Fact book

Source: World Bank https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups

Why are these countries any different?

Why are these countries any different?

• Users may not be literate
• Usually more mobile broadband than

home broadband users

• Users usually accessing internet from

mobile device – may be first access

• Bandwidth can be expensive
• Poor infrastructure in parts

Case Study: SmartSpeakers

• Why Q&A?
• Why voice?
• Text and technology illiterate in the slum
• Resource-constrained -> this is not like

asking Alexa or Google home for something

• Public space interactive speech system

• Prior work -> smart speakers

– Trouble with accents – Trouble with conversations and multi-person interaction – Neighbourly with Google

• What if you WoZ it? Replace backend with

a person?

Focus on emergent users of tech

• Use products with static buttons e.g.

feature phone

• Able to navigate websites/smart phones
• Able to type/input text for searching

Longitudinal deployment

• Also looked into systems ability to answer

questions

• Types of questions asked for each system
• Frequency of questions

Discussion

• Propose hybrid Artificial Intelligence –

Human-powered delay system

– Machine takes first pass and also passes to human if requires more complicated answer – Public speech system seems useful

What about….

• Privacy issues?

– Asking questions in public?

• Guidelines for what types of questions are

appropriate?

• Fairness in use of the system

– E.g. one girl used the system for 1 hour in 135 interactions!

• Everyone wants to be entertained…

Is this specific to India?

• Yes and no
• Some characteristics are common to low-

middle income country settings

• Have to be highly contextualized and

locally relevant

• Same system may not work in South Africa

for instance…

The Internet Is Mobile

5million fixed line

broadband subscriptions

354 million active

mobile broadband subscriptions

vs

Source: https://www.itu.int/en/ITU-D/Statistics/Documents/statistics/2019/ITU_Key_2005-2019_ICT_data_with%20LDCs_28Oct2019_Final.xls

41

Gupta, A., Calder, M., Feamster, N., Chetty, M., Calandro, E., and Katz-Bassett, E. “Peering at the Internet’s Frontier: A First Look at ISP Connectivity in Africa” PAM 2014.

42

Most recent study in South Africa

• Qualitative interviews with 52 low income

mobile users

• RQ: How do these users manage privacy
• n Facebook, WhatsApp, Instagram etc

Crime and physical safety

• Non-real time posting
• Concern for personally identifiable

information contained in photos

• Hard to overcome hacking owing to data

costs

Constraints are everywhere

• Cost
• Literacy level
• Education level
• Infrastructure
• Access to devices
• Access to Internet
• Any others?

Summary: So what does this mean for inclusive technology?

• We have to consider people’s

environments in addition to their abilities

• Have to consider culture
• Humans don’t exist in a vacuum
• Being inclusive means considering human

abilities, context, and constraints

Coming up…

• GP2 poster session
• GP2 due date is moved
• Considering different communities

Get in touch:

Office hours: Fridays 2-4pm (Sign up in advance) or by appointment JCL 355 Email: marshini@uchicago.edu