(U) NIST500-317 Cloud Computing Accessibility Considerations_draft.pdf

NIST Special Publication 500-317 (DRAFT) 1

Cloud Computing and Accessibility 4

Considerations 5

Robert B. Bohn 9

James Tobias 10

This publication is available free of charge 13

UNCLASSIFIED

NIST Special Publication 500-317 (DRAFT) 29

Cloud Computing and Accessibility 31

Considerations 32

Robert B. Bohn 34

Advanced Networking Technologies Division 35

Information Technology Laboratory 36

James Tobias 38

Inclusive Technologies 39

Princeton, NJ 40

This publication is available free of charge 42

March 2016 45

U.S. Department of Commerce 56

Penny Pritzker, Secretary 57

National Institute of Standards and Technology 59 Willie May, Under Secretary of Commerce for Standards and Technology and Director 60 ii

This page is left intentionally blank. 63 iii

Reports on Computer Systems Technology 64

The Information Technology Laboratory (ITL) at the National Institute of Standards and 66 Technology (NIST) promotes the U.S. economy and public welfare by providing technical 67 leadership for the Nation’s measurement and standards infrastructure. ITL develops tests, test 68 methods, reference data, proof of concept implementations, and technical analyses to advance 69 the development and productive use of information technology. ITL’s responsibilities include the 70 development of management, administrative, technical, and physical standards and guidelines 71 for the cost-effective security and privacy of other than national security-related information in 72 Federal information systems. This Special Publication 500-series reports on ITL’s research, 73 guidance, and outreach efforts in Information Technology and its collaborative activities with 74 industry, government, and academic organizations. 75

National Institute of Standards and Technology Special Publication 500-317 77

Natl. Inst. Stand. Technol. Spec. Publ. 500-317, 49 pages (March 2016) 78

CODEN: NSPUE2 79

DISCLAIMER

This document has been prepared by the National Institute of Standards and Technology

(NIST) and describes technical research in support of the NIST Cloud Computing Program.

Certain commercial entities, equipment, or material may be identified in this document in order to describe a concept adequately. Such identification is not intended to imply recommendation or endorsement by the National Institute of Standards and Technology, nor is it intended to imply that these entities, materials, or equipment are necessarily the best available for the purpose.

This publication is available free of charge. 82 iv

Acknowledgements 84

This document emerged from a series of discussions with a number of stakeholders through 86 participation in the NIST Cloud Computing Accessibility Public Working Group, chaired by Robert 87 B. Bohn of the U.S. Department of Commerce, National Institute of Standards and Technology. 88 The document contains input from members of the NIST Cloud Computing Accessibility Public 89 Working Group who provided commendable contribution and guidance. 90 Contributors to this document are - 92 Steve Jacobs, The Ideal Group 94 Clayton Lewis, University of Colorado 95 Alex Li, Microsoft 96 Kathleen McCoy, University of Delaware 97 Jamal Mazrui, Federal Communications Commission 98 Alice A. Smith, Department of Homeland Security 99 Gregg Vanderheiden, U Wisconsin-Madison 100 Jay Wyant, State of Minnesota CIO 101 v

Table of Contents 102

Acknowledgements ......................................................................................................................... iv 103

List of Figures: .................................................................................................................................. vi 104

List of Tables: .................................................................................................................................. vii 105

1. Introduction and Background .................................................................................................. 1 106

2. User Experiences of Inaccessibility .......................................................................................... 2 107

2.1. Introduction to Use Cases ............................................................................................... 2 108

3. From Use Cases to Barrier Categories ..................................................................................... 5 109

3.1. Version control ................................................................................................................ 5 110

3.2. Reliance on browser ........................................................................................................ 5 111

3.3. Platform quandary ........................................................................................................... 5 112

3.4. Use of thin clients ............................................................................................................ 6 113

3.5. Rich data visualizations ................................................................................................... 6 114

4. Applications Categories and Accessibility Guidance ............................................................... 6 115

5. New Cloud-Based Accessibility Opportunities ........................................................................ 8 116

5.1. Global Public Inclusive Infrastructure (GPII) .................................................................... 8 117

5.2. Accessibility Application Programming Interfaces (AAPIs) ............................................. 9 118

6. The Role of Information Resources ....................................................................................... 11 119

6.1. Example: the FCC's Accessibility Clearinghouse ............................................................ 12 120

7. Cloud Computing Accessibility Taxonomy ............................................................................. 13 121

7.1. Audiences and Goals ..................................................................................................... 13 122

7.2. The Dimensions ............................................................................................................. 14 123

7.3. Draft Taxonomy ............................................................................................................. 15 124

7.4. Taxonomy Mind Map .................................................................................................... 16 125

8. Conclusions ............................................................................................................................ 19 126

Appendix A: Comments from Raising the Floor – International on "US Government Cloud 127

Computing Technology Roadmap, Release 1.0 (Draft)" ................................................................ 20 128

A1. Introduction ................................................................................................................... 20 129

A2. General Accessibility Comments ................................................................................... 20 130

A3. Global Public Inclusive Infrastructure............................................................................ 23 131

A4. Other Recommendations .............................................................................................. 24 132

Appendix B: Guidance for Cloud Computing Accessibility ............................................................ 25 133 vi

B1. General Guidelines ........................................................................................................ 25 134

B2. Enterprise collaboration Tools - email, calendaring, IM ............................................... 26 135

B3. Customer relationship management (CRM) ................................................................. 27 136

B4. Web server / Content Management ............................................................................. 28 137

B5. Identity Management .................................................................................................... 28 138

B6. Document Retrieval / Library Systems .......................................................................... 28 139

Appendix C: The Role of Standards ............................................................................................... 30 140

Appendix D: Current Taxonomies ................................................................................................. 33 141

D1. AbleData ........................................................................................................................ 33 142

D2. Access for All.................................................................................................................. 34 143

D3. Cloud4All Semantic Framework for Content and Solutions (SEMA) ............................. 36 144

D4. EASTIN ........................................................................................................................... 38 145

D5. Global Accessibility Resource Initiative (GARI) .............................................................. 39 146

D6. IndieUI User Context ..................................................................................................... 41 147

D7. ISO 9999:2011 ............................................................................................................... 42 148

D8. Job Accommodation Network (JAN) ............................................................................. 42 149

D9. M376 ............................................................................................................................. 44 150

D10. Raising the Floor ............................................................................................................ 45 151

D11. Section 508 .................................................................................................................... 46 152

D12. Tables of Taxonomies .................................................................................................... 47 153

D12.1. By Audience ............................................................................................................... 47 154

D12.2. By Dimension ............................................................................................................. 47 155

D13. Taxonomies found within technical standards; not included at this time .................... 48 156

D13.1. ICF .............................................................................................................................. 48 157

List of Figures: 159

Figure 1 - Cloud Accessibility MindMap ........................................................................................ 18 161

Figure 2 - Screenshot of AfA Voice Recognition attribute within the Control class. ..................... 34 162

Figure 3 - Screenshot of AfA Voice Recognition data model, showing “voice profile identity” as 163 one of several attributes. .............................................................................................................. 35 164

Figure 4 - Screenshot of AfA showing details of “voice profile identity”. ..................................... 35 165 vii

Figure 5 - Screenshot of SAT navigation panes. ............................................................................ 37 166

Figure 6 - Screenshot of SAT content pane. .................................................................................. 38 167

Figure 7 - Screenshot of EASTIN search results at the Cluster level, indicating the number of hits.168 ....................................................................................................................................................... 39 169

Figure 8 - EASTIN screenshot showing some of the results for “Assistive Products for 170 Communication and Information … Input devices for computers … Input software”. ................. 39 171

Figure 9 - Screenshot of GARI search portal front page. ............................................................... 40 172

Figure 10 - Screenshot of GARI search option page. ..................................................................... 41 173

Figure 11 - Screenshot of GARI list of accessibility features associated with cognitive disabilities, 174 with definitions. ............................................................................................................................. 41 175

Figure 12 - JAN’s main pathway: Information By Product or Service............................................ 43 176

Figure 13 - Product listing for Vision Impairments ........................................................................ 44 177

Figure 14 - Screenshot of RtF taxonomy, showing both the Feature and Product hierarchies and 178 first sub-categories. ....................................................................................................................... 46 179

Figure 15 - Screenshot of RtF taxonomy showing bottom-level feature details........................... 46 180

Figure 16 - Screenshot of ICF Browser showing Chapter structure within “Activities and 181 Participation”. ................................................................................................................................ 49 182

Figure 17 - Screenshot of ICF Browser showing functional detail. ................................................ 49 183

List of Tables: 185

Table 1: Comments on NIST SP 500-293 Volume I with respect to Accessibility .......................... 21 186

Table 2: Comments on NIST SP 500-293 Volume II with respect to Accessibility ......................... 22 187

Table 3: Comments on NIST SP 500-293 Volume I with respect to GPII ....................................... 23 188

Table 4: Excerpt from NIST SP 500-291 ......................................................................................... 32 189

Table 5: Table B.2 shows the relationship between functional performance statements and 190 specific requirements, indicating where the relationship is primary or secondary . .................... 45 191

NIST SP 500-317 Cloud Computing Accessibility Considerations

1. Introduction and Background 194

The NIST Cloud Computing Program (NCCP) released a draft two-volume US Government (USG) 196

Cloud Computing Standards and Technology Roadmap1 in November 2011 for public comments; 197 it was published in final form in October 2014. The USG Cloud Computing Technology Roadmap 198 lists ten requirements and several Priority Action Plans that should be followed to fulfill the 199 requirements. Requirement #7 is to “Define unique government regulatory requirements and 200 solutions.” 201

Section 508 of the Rehabilitation Act2, among other laws, requires that federal employees and 203 citizens have equal access to information and communication technologies (ICT) regardless of 204 their disabilities. This qualifies as a “unique government regulatory requirement”, and 205 accessibility is considered to be fundamental to this solution. 206

Some comments to the Roadmap focused on accessibility with regard to the above requirement 208

(see Appendix A). Although the Program’s original goal was to stress three major challenge areas 209 for USG adoption cloud computing in security, portability, and interoperability, it was evident 210 that accessibility is as valid a challenge for the USG. Cloud computing solutions that address and 211 highlight accessibility offer a path forward for an agency to fulfill its mission and requirements 212 by providing a larger number of potential solutions that USG ICT managers can use to be 213 creative in the development of new services and solve their unique accessibility requirements. 214

As work progresses in cloud computing, it is important to promote, incorporate and discuss 215 applicable standards in accessibility for cloud computing services as a discipline for investigation. 216

In response to the interest in cloud and accessibility, the NCCP formed a new Public Working 218

Group (PWG) on “Cloud Computing and Accessibility” (CCA-PWG) in June 2013. This PWG will 219 address the topics facing cloud computing with respect to accessibility, standards and usage. 220

This document details the activity of that PWG, and is in part the key deliverable to date of the 221

PWG. We hope that this document will be useful to its primary audience, ICT managers in 222 federal agencies who are seeking to build accessibility into their cloud computing procurement 223 and use. 224

1 US Government Cloud Computing Technology Roadmap Volume I: High-Priority Requirements to Further USG Agency Cloud Computing Adoption; and Volume II: Useful Information for Cloud Adopters, NIST SP 500-293, http://dx.doi.org/10.6028/NIST.SP.500-293 2 http://section508.gov/

UNCLASSIFIED

UNCLASSIFIED

http://section508.gov/ http://dx.doi.org/10.6028/NIST.SP.500-293 http://section508.gov/

2. User Experiences of Inaccessibility 225

2.1. Introduction to Use Cases 226

Accessibility, and thus inaccessibility, are almost entirely concerned with user experiences. 227

Background processes are relevant only to the extent that they affect how users receive, 228 perceive, and act upon the information and communication that is before them. 229

To fully explore the issues of accessibility to electronic information, we have developed a series 231 of use cases that explore how a user interacts in different circumstances or cases. This is 232 especially useful in considering accessibility in cloud computing, due to the strong interaction 233 effects among device, browser, assistive technology (AT), app, etc. In many cases it is clear that 234 there are policy and process barriers as well as purely technological ones. 235

The following personas do not always reflect specific personal experience, but were collected or 237 drafted to express known difficulties encountered by ICT users with disabilities, including federal 238 employees, some of whom contributed to this document. 239

Cora is a customer service in a specialized work group that covers income from foreign sources. 241

She is blind from birth and uses a Braille display connected to her computer. Most of the 242 information she handles shows up in specific blocks on the screen, and she has learned the 243 keyboard commands to give those blocks focus on the Braille display as needed. Unexpected 244 software updates to the internal cloud application sometimes change the layout and cause the 245

Braille display to lose focus. This requires some assistance from the IT support center to get her 246 back on track; in the meantime her productivity is compromised. The IT support people have 247 come to expect her calls whenever there is a software update. They serve other blind users, who 248 use a range of screen readers and Braille output devices. Cora and her blind peers have tried to 249 escalate this problem but have had limited success. 250

Garrett is a wounded veteran with a moderate cognitive disability as a maintenance technician 252 in a remote area. He drives between work sites and uses a mobile device that lets him navigate 253 by GPS and retrieve his work orders through the company’s app. Some work orders are 254 confusing and he needs help. He must place a call to his supervisor, and slowly read aloud the 255 text of the order. The supervisor then explains the work order and occasionally must text him a 256 complete, simplified order, in a regular text messaging app. Keeping track of the company’s app 257 and the separate text messages can be confusing as well, and makes Garrett’s recordkeeping 258 less accurate, but it does let him get his maintenance work done. He and his supervisor talk 259 about creating a simpler solution, but they do not have any resources to develop software or 260 even explore what their organization may already have that they could use. 261

Deena is a program administrator at a federal contractor working on many projects with several 263 agencies. She is an older worker and has been experiencing problems with her vision and 264 memory. Her company and the multiple agencies she works with all use different management 265 applications, some in the cloud and some not. She has trouble keeping track of her logins and 266 passwords; she keeps a ‘cheat sheet’ in her desk drawer, which is against policy. She also has 267 trouble copying and pasting information from an agency application into her company’s 268 management tool – the highlight color is yellow on a white background, which doesn’t work for 269 her. So she copies the whole page, pastes it into a word processor with higher contrast, selects 270 the text she wants to copy, and then pastes it into her company’s tool. At the end of the day, 271 she has many open word processor pages that need to be saved or discarded, and this is very 272 time consuming. She says “in the old days” most of her work was done by phone, informally 273 checking in with her colleagues in the federal agencies she was working with, but that nowadays 274 there is a need to document everything, which means more typing and reading than she can 275 easily do. She gets tired and has headaches many times during the week. She attended a 276 workshop for low vision solutions such as larger or clearer monitors and high contrast settings a 277 few years ago, but did not follow up with her supervisor. 278

Roberto works as a statistical analyst. He is deaf and uses video relay and video remote 280 interpreting to communicate with his workmates. His supervisor encourages him to use direct 281 text instead whenever possible, for budget reasons. Roberto finds sign language to be a more 282 effective form of communication. Some of the training videos he needs to use have poor quality 283 captions with errors, and the transcripts do not let him know what is being said at what part of 284 the video. He and his deaf peers informally exchange information about good and bad 285 captioning training videos; some admit that they choose what training they take based on the 286 quality of the captions. A recent problem has been an increase in the use of telecollaboration for 287 project meetings. Sign language is not available at all on these meeting calls. Real-time 288 captioning (CART) is not always available and the quality is mixed. When it is slow or incorrect, 289 he misses opportunities to ask questions or make comments. Moreover, when he is reading 290 captions he occasionally misses some content of the graphs in the main part of the screen, and 291 has to review the recording of the session to catch up. Lately, he has been using his tablet to 292 attend these sessions. This is good for his flexible schedule, but he has to balance the size of the 293 captioning and chat panes with the size of the text in the main content pane. Depending on the 294 content, this can be problematic. 295

Virginia uses multiple applications plus email on a daily basis. Because of a spinal cord injury, 297 she cannot use her hands for typing and has limited use of the mouse. She relies on Dragon 298

Naturally Speaking, a speech recognition system that lets her speak words to type and say 299 commands to control the computer. Virginia’s component is working with a cloud service 300 provider who plans to provide software as a service and to utilize a virtual desktop as a delivery 301 mechanism. Because the virtual desktop, Citrix Receiver, does not work with Dragon Naturally 302

Speaking, Virginia will not be able to work unless an exception is made to continue supporting 303 her current configuration. Virginia knows this because she works in the accessibility group. 304

Allen uses multiple applications plus email on a daily basis. Allen has been blind from birth. He 306 cannot use a mouse because he cannot see the screen to follow the cursor. He uses a screen 307 reader, assistive technology that speaks the screen information to him and lets him control the 308 computer with the keyboard. Allen is in Virginia’s component. His group is going to the virtual 309 desktop next week. Allen doesn’t know that his assistive technology will not work with the 310 virtual desktop and is facing a work stoppage next week. His situation and user needs in general 311 were not considered during the planning process. The virtual desktop vendors claim that 312 assistive technology will work if installed on the remote desktop. Allen’s agency has tested this 313 and found that it works poorly or not at all. However, there is no requirement in place for the 314 vendor to fix this compatibility problem in time for the transition; there is no timetable at all 315 built into the service contract. The current plan is to have Allen use a separate system until the 316 problem is resolved, but there is no clarity on how that will work or how long it will last. 317

In order to make compliance training more interesting and ‘game-like’ Jason’s agency has begun 319 to teach courses online so that units can track progress and compete for high scores and early 320 completion of training. These courses often use graphics and interface elements that make it 321 impossible for Jason, who is blind, to use a screen reader to independently complete the 322 required training. The software developers did not follow best practices regarding accessibility, 323 as it was not part of the contract. Jason often has difficulty completing courses, due to no fault 324 of his own. One typical problem is that as the training application calculates his work to show 325 progress graphically, his computer freezes and no record is retained. Jason’s supervisor has to 326 request system logs of Jason’s training instead, which is time-consuming and cumbersome. 327

Nancy is an agency financial analyst with Attention Deficit Disorder. Her agency regularly 329 requires her to teach using webinars. While the hosting software has the capacity to 330 demonstrate calculations and can show the impact of proposed changes on the agency’s 331 finances in near real time, the screen cannot be modified by individuals. Nancy is often 332 distracted by incoming questions when teaching and loses her train of thought. As a work 333 around, Nancy tried to call into the seminar and ask a colleague to change screens for her and 334 field questions but this has led to confusion when the slides were not in sync or when students 335 were asking for demonstrations of calculations needed in their jobs and Nancy needed to 336 comment on specific results. 337

The agency Tuan works at has approved apps on mobile devices to gather and support cloud-339 based data collection and analysis. Tuan, who has limited use of his hands, is knowledgeable 340 about the assistive technology he needs, a switch selection system on his tablet. However, the 341 apps his agency uses were not designed according to the accessibility recommendations for the 342 operating system, which often change. Since the last update, Tuan is unable to perform specific 343 key functions on the apps that let him import, compare, and aggregate data from multiple cloud 344 sources, limiting his ability to explore the data creatively. This reduces both his value to the 345 agency and the intrinsic reward he gets from his job. 346

3. From Use Cases to Barrier Categories 347

The next step in our analysis is to translate use cases like the ones listed above into a small set 349 of barrier categories experienced specifically in cloud computing environments. 350

Although cloud computing has changed the landscape in ICT, the current consensus is that its 351 unique accessibility implications are often minimal. That is, accessibility is more concerned with 352 the user interface than the infrastructure on which that interface runs, the processes to which 353 data is subjected, or the content the interface provides a port into. To an end user, whether the 354 computing environment is cloud-based or not may often be invisible or secondary to the 355 accessibility barriers encountered in the direct user experience. Taking that into account, the 356

CCA-PWG identified a few issues that are unique to managing accessibility in the cloud: 357

3.1. Version control 358

Changes to cloud computing software (e.g., fixes and updates) are generally controlled by the 359 provider, rather than an explicit part of customer enterprise operations management. This can 360 result in unexpected accessibility jeopardies to users who may not be able to roll back the 361 changes. 362

3.2. Reliance on browser 363

When cloud computing involves a browser rather than direct use of a cloud app, it inserts an 364 additional link in the ‘accessibility value chain’, the sequence of technologies that must 365 successfully interoperate for the appropriate accessibility features to reach the user. For 366 example, conventional computing may link an operating system, a screen reader, and a 367 spreadsheet program; a comparable cloud computing chain consists of an operating system, a 368 screen reader, a browser, and a cloud-based spreadsheet. Note that browser interposition is not 369 always a negative; browser settings such as enlargement and browser-based screen readers are 370 sometimes easier to find and use than other options. 371

3.3. Platform quandary 372

Cloud computing is intended to simplify ICT operations in part by reducing the need for 373 enterprises to provide dedicated support for too wide a range of platforms and operating 374 systems. In general, this is good for accessibility in that savvy users can select and become 375 expert in the platform that provides the best accessibility. However, it may unintentionally 376 jeopardize non-expert users with disabilities who may have difficulty navigating the complicated 377 decision space, while also not being able to rely on enterprise-supported ICT. For example, an 378 employee who is permitted to use a personal device (‘bring your own device’ or BYOD) for work 379 may have difficulty identifying which device (and version, and workplace applications, and 380 utilities, and assistive technologies…) will best fit both the workplace requirements and his/her 381 own accessibility needs. 382

3.4. Use of thin clients 383

The advent of cloud computing has brought the use of thin and very thin clients back into use. 384

These are characterized by terminals that collect input (keyboard, mouse) and display output 385

(screens), but leave all the computing to be performed at the other end. In fact, this is what 386 many people think of what is essentially different about cloud computing – that the actual 387 computing occurs in a network, not in a desktop or other user-facing device. If this includes 388 having the screen rendered at the other end (and just the image sent back to the terminal) then 389 many assistive technologies will not work. In particular screen readers (which would have to run 390 on the remote server to access the screen information) suffer from control and audio delays 391 that can significantly affect performance and use. Screen enlargers can be almost unusable if 392 run on the remote server – and have no access to screen information for tracking if run on the 393 local terminal. Other AT also may have to be installed and run on the remote servers – meaning 394 that a person may have to have their AT installed on many servers instead of just on their 395 personal terminal/computer. 396

3.5. Rich data visualizations 397

Dynamic data (e.g., weather) and large datasets are being used to create visualizations that 398 make it easier for a sighted user to understand what’s going on. Both the data and the 399 processing are in the cloud, often due to client-side limitations. Since these are dynamic, the 400 usual accessibility solutions such as static alt text, longdesc, or ARIA may not be possible, and no 401 resources are widely available for creating on-the-fly alternatives. Sometimes the visualizations 402 go beyond the complexity of the data. That is, if a screen reader user had access to the data 403 he/she might be able to make sense of it, but the visualization is the only access provided. 404

4. Applications Categories and Accessibility Guidance 405

The next step of the analysis is to examine cloud computing applications to understand where 407 and how they cause or prevent accessibility barriers. We begin by roughly categorizing them. 408

There are many kinds of cloud computing products, services, and features. To enumerate or 409 catalog them is beyond the scope of this document. However, it is possible to categorize the 410 most frequently used applications according to functionality as listed below. Note that, as is 411 characteristic for cloud computing, many of these functions can be and are performed on a 412 user’s device, via a browser or app, or completely in the cloud. From a user perspective it is not 413 necessarily important to understand the exact location of any computational function, since the 414 user’s experience is the result of the entire end-to-end interoperation. However, some of the 415 following functions are more purely ‘cloud’ than others. 416

• Enterprise collaboration tools – email, calendaring, messaging, document co-editing, 417 webinar platforms and related real-time collaboration tools 418

• Resource/budget planning/management and project management tools 419

• Customer relationship management (CRM), including customer support and 420 documentation 421

• Web server/Content Management Systems (CMSs) 422

• Identity management 423

• Document retrieval/Library systems/Online storage 424

Many of the technology products and services associated with these functions are designed to 426 be used in many situations, by many different types of users, for different purposes. That is, 427 even a specific product like a project management tool is intended for flexible use, and has a 428 wide range of capabilities, which are normally tailored down for a given environment of use. For 429 many of these situations, the pathway to solution begins with identifying the specific goals in 430 the environment of use. 431

To take one of the use cases we identified earlier, consider Garrett’s situation. Part of his 432 problem could be identified as pertaining to the mobile platform he uses. Selecting the right 433 platform for his needs, that also meets agency platform compatibility and operational realities, 434 is highly complex and rapidly evolving. One solution to part of his difficulty might be solved if he 435 could view his current task at the same time as seeing the exchange of text messages with his 436 supervisor, instead of having to shift back and forth between applications. An update to his 437 current platform’s operating system may permit such multi-tasking directly, or indirectly via a 438 simple gesture swapping between applications. 439

We will discuss later how information resources play an essential role in solving specific 440 accessibility problems. For now, we focus on the use case to develop the point that although 441 accessibility must be considered globally during procurement and development, it must be 442 managed locally during use. For Garrett, this means that identifying the specific function of 443

“seeing the task and the message thread at the same time” is a requirement for successful use. 444

Appendix B provides some specific guidance on some of these application categories in addition 445 to general guidance on cloud computing. We emphasize that for reasons of accuracy, 446 comprehensiveness, and especially harmonization, those looking for technological guidance 447 refer first to the Web Content Accessibility Guidelines (WCAG 2.03); not only the Guidelines 448 themselves, but the whole ecosystem of Success Criteria and other material provided there. 449

3 http://www.w3.org/TR/WCAG20/

UNCLASSIFIED

http://www.w3.org/TR/WCAG20/

5. New Cloud-Based Accessibility Opportunities 451

So far, this document has focused on how cloud computing either replicates or exacerbates 453 accessibility problems found in previous computing environment models. But cloud computing 454 offers new and exciting accessibility opportunities, not just new problems. This section describes 455 two of the most promising. 456

5.1. Global Public Inclusive Infrastructure (GPII) 457

The GPII4 is a global multi-year program to build an accessibility infrastructure in the cloud. The 458

GPII will allow people who have problems with standard interfaces and content – all of us, in 459 some situation or other – to be able to use information and communication technology (ICT) 460 products and services anywhere they encounter them. It will allow users to invoke the interface 461 and content adaptations they want or need, automatically, on any device, anywhere, anytime. 462

The purpose of the GPII is to ensure that everyone who faces ICT barriers due to disability or 463 usability issues, literacy, digital literacy, or aging, regardless of economic resources, can access 464 and use the Internet and all its information, communities, and services for education, 465 employment, daily living, civic participation, health, and safety. 466

The GPII will encompass a family of tools and services: 468

Tools for users to explore, select, store, and manage their preferences – the features 470 they want or need for interfaces and content 471

Ways to match those user needs and preferences with the capabilities of the devices 472 users encounter: public computers, websites, online videos, transaction machines, etc. 473

Ways to negotiate and implement those preferences on the user’s current device, in 474 real time – “automatic personalization” 475

Development tools to let assistive technology and mainstream companies, commercial 476 and non-commercial, build or adapt their accessibility software and services into the 477 cloud, including ways for them to reach and serve more potential users 478

Management tools so that all participating entities – consumers, software companies, 479

AT funding agencies, ATM manufacturers, educators, employers, retailers, etc. – can be 480 assured of highly reliable, effective, secure, private, efficient, and manageable 481 accessibility solutions. 482

The GPII will not create new access technologies or services itself. It is the infrastructure for 484 making their development, identification, delivery, and use easier, less expensive, and more 485

4 http://gpii.net http://gpii.net/ effective. This is similar to building a road system; roads do not provide transportation but 486 greatly enhance the ability of car companies and others to do so. They provide an infrastructure 487 that car companies themselves cannot make. The Internet is the infrastructure for general 488 information and commerce. The GPII enhancements to the Internet will enable the Internet to 489 be truly inclusive for the first time. 490

We have an unprecedented opportunity to fundamentally change and advance accessibility – to 492 create an infrastructure that supports both commercial and non-commercial efforts to make 493 accessibility more affordable, to reach more of the people who need it than the estimated 3 - 494

15% we reach now, to make it all simpler, to serve disabilities and aging groups we don't now 495 serve or serve well, and to build access that will work with the new technologies that are 496 coming. 497

The implications for US federal cloud computing are powerful. The GPII will offer federal entities 499 an integrated method to meet the needs of both employees and members of the public, as 500 required by Section 508 and other laws, with greater confidence and lower cost. In order to 501 redeem this promise, planning for federal cloud computing should factor in the development 502 and implementation of the GPII as one of the solution paths agencies can travel. Policies and 503 tools that NIST is building as part of its cloud computing roadmap responsibilities can help bring 504 accessibility into the mainstream of federal ICT where its impact is greater and costs are lower; 505 the GPII can be a significant part of that effort. 506

5.2. Accessibility Application Programming Interfaces (AAPIs) 507

An accessibility application programming interface (AAPI) is any capability built into an operating 508 system or platform that provides developers with the necessary software interface for 509 accessibility features such as screen reading and magnification. For example, the iOS platform 510 provides global screen reading and magnification capabilities as a default for all apps developed 511 on that platform. This vastly simplifies and standardizes accessibility solutions. AAPIs may not 512 have originally been intended for promoting accessibility features, but have that effect, which is 513 a natural outgrowth of ‘responsive design’ aimed at serving the wide range of mobile devices. 514

A persistent challenge for computer accessibility has been the limited adoption of accessibility 516

APIs by developers. Many developers perceived limited benefits to incorporating accessibility 517

APIs into their products. In the desktop era, software ran on a single device in a single context 518

(i.e., a desktop computer in an office) and the vast majority of users could use software without 519 the need for any adaptations. In the laptop era, software ran on two similar devices (desktops 520 and laptop computers) in a larger number of contexts, but the need for adaptations by most 521 people was still limited. The advent of cloud computing, in combination with mobile computing 522 devices (e.g., tablets, smartphones), has created an environment where the same software 523 can run on multiple devices with very different input and output capabilities and a wide variety 524 of use contexts (e.g., while driving; while walking; one-handed; in a noisy environment; in direct 525 sunlight). Developers adopt tools like HTML5 and CSS3 because they offer the promise of 526 the same software running across multiple devices. Developers should similarly adopt 527 accessibility APIs because they allow software to run in multiple contexts. Developers who want 528 their product to have a competitive advantage because users can access it in their office (on a 529 computer), in their living room (on a tablet) and out and about (on their smartphone) regardless 530 of where they are, what else they are doing and what is going on around them, but don't want 531 to implement all these capabilities from scratch, should use cloud computing, cloud storage, 532 non-native code and accessibility APIs. In this environment, accessibility should not be perceived 533 as an obligation or additional requirement, but rather as a fundamental aspect of software 534 development and a selling-point for a product. 535

Understanding the accessibility-related benefits of APIs allow system developers, system 537 owners, and other stakeholders to modernize their systems in ways that benefit all 538 stakeholders. 539

For example: 541

1. Accessible Ports-of-Entry: APIs can significantly leverage the ability to develop fully-543 accessible "ports-of-entry" to cloud-based resources. 544

2. Efficiency: API access allows content to be created one time in a manner that can 546 enable developers to make it accessible and available through many channels. 547

3. Wider Reach: By allowing anyone to create new presentation layers, like a 549 mobile/computer-based applications and websites, APIs can be used to create services 550 and information that is fully-accessible to people with disabilities. 551

4. Distribute Services and Information to New Audiences: APIs can be used to distribute 553 services and information to new audiences and in specific contexts that can be 554 customized to meet the access needs of individual consumers. 555

5. Leverage Government Assets: Data and information produced by the federal 557 government is a national asset, paid for by the American people. APIs can expose data 558 that was only available to a few and make it more available to everyone, including 559 people with disabilities. 560

6. Automation: APIs allow machines to handle workloads which might otherwise require 562 the manual work of a human being. This can be as simple as having one content update 563 propagate to multiple sections of an accessible website (or multiple accessible websites) 564 at once. 565

7. Application: Providing API access to information or a service sets up the use of that 567 information or service by accessibility experts developing mobile apps. This is especially 568 beneficial if an organization plans to build more than one app. APIs can serve as a 569 shortcut to developing an accessible second, third, and fourth app. 570

8. Partnerships: The mission of every agency is supplemented by like-minded non-profits 572 and businesses, who are interested in using agency information and services to provide 573 services. They do this by consuming and repurposing agency material into new, useful 574 products and putting agency content in front of their customers and clients. Enabling 575 businesses to easily create accessible interfaces to government information can fuel 576 innovation. 577

9. Increasing Efficiency and Reducing Cost: If developers plan to develop more than one 579 resource, APIs can serve as a shortcut to developing additional ones. 580

10. Integration: APIs allow content to be more easily embedded or interwoven throughout 582 websites and applications. It can help to ensure smooth, integrated, and accessible user 583 experiences. 584

11. Personalization: Users of government websites can benefit from the ability to 586 personalize and enhance the accessibility of sessions with the information and services 587 that are most useful to them. 588

12. Mashups: Mashups allow the public to better understand government information in 590 the context of other sources of information. An agency service or data stream can be a 591 small, important part of another service... especially if it is easily mashed-up and made 592 accessible. 593

13. Future-Ready: As needs change, APIs can help to support unanticipated future uses in 595 manners that make it easier to render deliverables in an accessible manner. 596

6. The Role of Information Resources 597

At the beginning of 2014, the CCA-PWG turned to a separate but related issue in cloud 599 computing accessibility. As can be seen in several of the use cases above, some accessibility 600 problems are caused by a lack of information in the right hands at the right moment, rather than 601 the total absence of a technological solution. In Garrett’s case, we suggested that his agency 602 keep an eye on the mobile environment to learn when that marketplace offers a mainstream 603 solution to a key requirement, Garrett’s need to see both his task and texts from his supervisor 604 at the same time. Can his organization survey the mobile environment itself to discover this 605 capability? If it cannot do this itself, can it learn from others, in the federal space or elsewhere, 606 about emerging solutions? 607

Finding accurate, timely, situationally-relevant information about accessibility is not trivial. 608

Although new features constantly pop up across the horizon of products and services, their 609 accessibility implications are not usually highlighted or reported broadly or consistently. The 610 speed and churn of modern mainstream ICT makes curating accessibility solutions challenging 611 even to full-time accessibility professionals, let alone the many more part-timers, including 612 people with disabilities themselves, who must identify and resolve specific problems with fewer 613 resources at hand. 614

However, this is a known problem. Efforts have been and are being made to address the need 615 for better information resources about accessible ICT. 616

6.1. Example: the FCC's Accessibility Clearinghouse 617

The FCC's Accessibility Clearinghouse5 is a web repository of information about accessible 618 communications products and services. It is mandated by the 21st Century Communications 619 and Video Accessibility Act of 2010 (CVAA), which also requires the FCC to promote this resource 620 to the public. The Accessibility Clearinghouse includes information on the following topics: 621

Accessibility features of mobile phones 622

Accessibility contacts at telecommunications and advanced communications services 623 companies 624

Free assistive apps for various computing platforms 625

Organizations implementing the National Deaf-Blind Equipment Distribution Program 626

(NDBEDP)6. 627

To enable more flexible and targeted use of Clearinghouse information, the FCC has released a 629 software development kit (SDK; blog post7; source code8; demo video9) that makes custom 630 programming of inquiries to the database relatively easy with a free, popular language called 631

Python. Python-based apps can query the Clearinghouse based on an application programming 632 interface (API). Clearinghouse Information may be selectively filtered and combined with data 633 from other sources as mashups. Views of information may be created that differ from the 634 presentation on the Clearinghouse website, thereby tailoring value for particular constituencies 635 or devices. 636

5 http://www.fcc.gov/accessibilityclearinghouse 6 https://www.fcc.gov/general/national-deaf-blind-equipment-distribution-program 7 http://www.fcc.gov/blog/fcc-releases-software-development-kit-accessibility-clearinghouse 8 http://www.github.com/FCC/clearinghouseSDK 9 http://www.youtube.com/watch?v=_Pk0igOFmqY http://www.fcc.gov/accessibilityclearinghouse http://www.fcc.gov/blog/fcc-releases-software-development-kit-accessibility-clearinghouse http://www.github.com/FCC/clearinghouseSDK http://www.youtube.com/watch?v=_Pk0igOFmqY http://www.fcc.gov/accessibilityclearinghouse https://www.fcc.gov/general/national-deaf-blind-equipment-distribution-program http://www.fcc.gov/blog/fcc-releases-software-development-kit-accessibility-clearinghouse http://www.github.com/FCC/clearinghouseSDK http://www.youtube.com/watch?v=_Pk0igOFmqY

7. Cloud Computing…