From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from w088.z208036108.was-dc.dsl.cnc.net ([208.36.108.88] helo=toccata.grg.afb.net) by speech.braille.uwo.ca with esmtp (Exim 3.35 #1 (Debian)) id 172zx4-0007zK-00 for ; Wed, 01 May 2002 15:39:47 -0400 Received: from localhost (janina@localhost) by toccata.grg.afb.net (8.11.6/8.11.6) with ESMTP id g41JYOj02059; Wed, 1 May 2002 15:34:42 -0400 X-Authentication-Warning: toccata.grg.afb.net: janina owned process doing -bs Date: Wed, 1 May 2002 15:34:24 -0400 (EDT) From: Janina Sajka X-X-Sender: janina@toccata.grg.afb.net To: uaccess-l@trace.wisc.edu, Subject: I swear to tell the truth, Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable X-MIME-Autoconverted: from 8bit to quoted-printable by toccata.grg.afb.net id g41JYOj02059 Sender: speakup-admin@braille.uwo.ca Errors-To: speakup-admin@braille.uwo.ca X-BeenThere: speakup@braille.uwo.ca X-Mailman-Version: 2.0.8 Precedence: bulk Reply-To: speakup@braille.uwo.ca List-Help: List-Post: List-Subscribe: , List-Id: Speakup is a screen review system for Linux. List-Unsubscribe: , List-Archive: The whole truth, and nothing but the truth, so help me God. >>From the web page http://www.microsoft.com/presspass/trial/mswitness/2002/hofstader.asp=20 IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF columbia =20 STATE OF NEW YORK ex rel. Attorney General ELIOT SPITZER, et al., Plaintiffs, v. MICROSOFT CORPORATION, Defendant.=20 =20 Civil Action No. 98-1233 (CKK) Written Direct Testimony of CHRIS HOFSTADER TABLE OF CONTENTS I. Background *=20 II. The Need for Assistive Technology * III. Microsoft’s Support for Accessibility * IV. Freedom Scientific’s Assistive Technology=20 Software Products * V. JAWS’ Reliance on "Middleware" * VI. JAWS Would Break if the "Middleware" on=20 Which It Relies Were Removed. * VII. The Removal or Alternation of Internet Explorer=20 Would Deprive Blind and Low Vision Users of the=20 Opportunity To Access the Internet. * VIII. Creating Different Versions of Windows Would=20 Greatly Increase Freedom Scientific’s Testing=20 Burden and Technical Support Calls. * IX. Anything that Alters the Consistent User Interface=20 of Windows Disadvantages Blind and Low Vision=20 Computer Users. * X. Microsoft’s Disclosures of APIs Are Timely and=20 More than Adequate. * =20 My name is Chris Hofstader. I am the Vice President of Software= =20 Engineering for Freedom Scientific Inc. Freedom Scientific is t= he=20 world leader in the development, manufacture and marketing of=20 assistive technology software and hardware products that people= use to=20 overcome the challenges of blindness, low vision and other=20 disabilities. Our software programs and hardware devices provid= e tools=20 for our customers to attend school, hold jobs, access personal=20 computers and simply live their everyday lives. Freedom Scienti= fic’s=20 company mission is: "To change the world for people with sensor= y=20 impairments and learning disabilities by creating innovative,=20 technology based solutions."=20 Not only do I manage the development of software products for b= lind=20 and low-vision people, but because I am blind, I also use our s= oftware=20 products both at work and at home.=20 Freedom Scientific is a privately-held company that has approxi= mately=20 183 employees today, the vast majority of whom work in the Blin= d and=20 Low Vision Group where they develop software and hardware produ= cts for=20 blind and low vision computer users. Over 30% of the employees = who=20 work in the Blind/Low Vision Group at Freedom Scientific are bl= ind or=20 vision impaired. Background I graduated from New York University in 1982 with a Bachelor of= =20 Science degree in Computer Science.=20 I have 23 years of professional experience working in software=20 development. From 1979 through the fall of 1998, I held various= =20 positions as a computer programmer and software developer for n= umerous=20 companies, and I have also worked as an independent software=20 consultant. A copy of my resume is attached hereto as Exhibit A= .=20 In October 1998, I joined Henter-Joyce, a company that develope= d=20 assistive technology software products for blind and low vision= =20 people. I was the Manager of Software Development at Henter-Joy= ce, and=20 I developed software products for blind and low vision people.=20 Henter-Joyce developed one of the first screen readers, which i= s=20 software that converts the contents of a computer screen to spe= ech and=20 Braille. Blind people use screen readers to access the informat= ion on=20 their computers.=20 Freedom Scientific was formed in April 2000 as a result of the=20 acquisition of Henter-Joyce and two other assistive technology=20 companies. After Henter-Joyce was acquired, I was promoted to t= he=20 position of Director of Software Development for Freedom Scient= ific.=20 Since August 2000, I have been Vice President of Software Engin= eering=20 at Freedom Scientific. I work in the Blind and Low Vision Group= and am=20 responsible for a group of approximately 25 software developers= who=20 create software products for blind and low vision people.=20 The Need for Assistive Technology There are between 1.5 and 2 million blind and approximately 10 = million=20 visually impaired people in the United States who are blind or = low=20 vision. They are likely to experience some degree of difficulty= in=20 working with computers. For example, they may be unable to see = text or=20 images on a computer screen and perform tasks that require eye-= hand=20 coordination such as moving a computer mouse.=20 Assistive technology software programs, such as those developed= by=20 Freedom Scientific and its competitors, make it possible for pe= ople=20 with disabilities to use computer technologies successfully at = work,=20 at school, and at home.=20 Accessible computers are more than just good business—the= y are a=20 matter of civil rights. I am not a lawyer, but based on my expe= riences=20 working at Freedom Scientific and Henter-Joyce, I have become f= amiliar=20 with the federal laws requiring accessibility for disabled peop= le. For=20 example, the Americans with Disabilities Act requires that empl= oyers=20 make their workplaces accessible to employees with disabilities= .=20 Similarly, the Individuals with Disabilities Education Act mand= ates=20 that school districts provide assistive technology products for= =20 disabled students. Freedom Scientific’s assistive technol= ogy products=20 enable employers and schools to comply with their obligations u= nder=20 federal law to provide accessible technology to blind and low v= ision=20 employees and students.=20 Moreover, Section 508 of the Rehabilitation Act requires that=20 electronic and information technology developed, procured, main= tained=20 or used by the federal government or federal agencies be access= ible to=20 people with disabilities. Thus, any computer or software purcha= se by=20 the federal government must comply with Section 508. States tha= t=20 receive federal funds under the Assistive Technology Act of 199= 8 are=20 also required to comply with Section 508 of the Rehabilitation = Act.=20 Microsoft’s Support for Accessibility=20 Microsoft has truly been a pioneer in empowering people with=20 disabilities by developing technologies in its products that en= able=20 them to overcome barriers to employment and participate fully i= n other=20 aspects of life. Microsoft was one of the very few software com= panies=20 that was committed to providing first-rate assistive technology= =20 support in its products before any federal law required that su= ch=20 things be available.=20 Microsoft has expended significant effort in making Windows ope= rating=20 systems accessible to people with disabilities, including blind= and=20 low vision people. Beginning in 1988 with Windows 2.0, Microsof= t began=20 including accessibility features in its operating system. Over = the=20 years, Microsoft has continually improved and expanded its effo= rts to=20 make its operating systems and applications accessible to peopl= e with=20 disabilities.=20 Most recently, Microsoft has built a wide range of accessibilit= y=20 options into its Windows 2000 and Windows XP operating system p= roducts=20 including an accessibility wizard to help people adapt various=20 operating system features to their needs. Windows XP contains d= ozens=20 of built-in features specifically for people with disabilities.= For=20 example, users can bypass the mouse altogether by taking advant= age of=20 a feature known as MouseKeys that allows them to perform mouse-= based=20 tasks using the numeric keypad instead. Microsoft has added clo= se=20 captioning and audio description technology to make multimedia = content=20 accessible to deaf and blind users. In addition, Windows 2000 a= nd=20 Windows XP include the Narrator utility, which is a very basic = text to=20 speech program for blind users, and the Magnifier utility, whic= h=20 provides a minimum level of screen magnification for low-vision= =20 individuals. Because these assistive technology utilities are a= lways=20 present on a Windows PC, a blind or low vision person can make = limited=20 use of their computer (or a friend’s or colleague’s= computer) without=20 installing a full-fledged assistive technology product like JAW= S=AE for=20 Windows from Freedom Scientific.=20 The experience I had installing Windows XP on my home computer = is just=20 one example of how much easier Microsoft has made it for blind = and low=20 vision people to use the most fundamental elements of the Windo= ws=20 operating system products, even without the addition of an assi= stive=20 technology aid such as those made by Freedom Scientific. On the= day=20 after Thanksgiving, my wife drove me to Circuit City so that I = could=20 purchase a copy of Windows XP. When I got home, I inserted the = Windows=20 XP CD-ROM into the CD-ROM drive of my computer. The Windows XP=20 Installation program, using the Microsoft Text to Speech Engine= via=20 the Microsoft Speech Application Programming Interface, then sa= id that=20 if I want to use the accessibility features, hit Windows key "U= ." I=20 did so and immediately the Narrator utility in Windows was laun= ched=20 and I could proceed with the installation process guided by the= spoken=20 prompts. Without any assistance from my wife, I was able to ins= tall=20 Windows XP "soup to nuts" by myself using the basic speech feat= ures=20 built into Windows XP. I then installed JAWS for Windows on top= of=20 Windows XP.=20 In addition, as discussed in more detail below, Microsoft has p= rovided=20 the best assistive technology application programming interface= s=20 ("APIs") in its operating systems. Independent software vendors= =20 ("ISVs"), like Freedom Scientific, can rely on these APIs when=20 developing assistive technology software programs that run on W= indows=20 operating system products. Mainstream software vendors (e.g., C= orel=20 and Lotus) can also use these APIs to communicate with assistiv= e=20 technology products like JAWS from Freedom Scientific.=20 No other operating system vendor comes close to matching the su= pport=20 Microsoft provides to assistive-technology ISVs. Microsoft devo= tes=20 considerable resources to helping ISVs who develop assistive=20 technology software products. Microsoft has a group called the=20 Accessible Technology Group that has more than 40 employees who= work=20 with assistive-technology ISVs, like Freedom Scientific, and wi= th=20 disability advocates to ensure that people with disabilities ca= n use=20 the software developed by Microsoft and other ISVs. The Accessi= ble=20 Technology Group works closely with Freedom Scientific and its=20 competitors to ensure that we receive the information and assis= tance=20 we need to create accessible applications to run on Windows.=20 For example, last summer Freedom Scientific attended a week-lon= g event=20 at Microsoft’s campus in Redmond which was designed to pr= ovide=20 assistance to assistive technology ISVs in porting their applic= ations=20 to Windows XP. Microsoft wanted to ensure that assistive techno= logy=20 ISVs had the assistance we needed to make sure that our product= s would=20 function properly on its latest operating system so that blind = and low=20 vision people would have Windows XP-compatible assistive aids=20 available when Windows XP was launched.=20 From time to time, Microsoft also discusses with Freedom Scient= ific=20 Microsoft’s plans for improving or incorporating addition= al assistive=20 technology features into future operating system products and s= olicits=20 our feedback on such improvements.=20 Not all operating systems are as supportive of assistive techno= logy as=20 Microsoft’s Windows operating system. Apple Computers, fo= r example,=20 has done little to make its Macintosh operating systems and=20 applications accessible to the blind. Only recently has Apple t= aken=20 the first steps to work with assistive-technology ISVs to discu= ss=20 developing assistive-technology products for the Apple Macintos= h=20 operating system. While there are a number of good accessibilit= y aids=20 available for the Linux operating system, none of those aids re= lies on=20 any functionality in the Linux operating system itself. Moreove= r,=20 Linux vendors such as Red Hat offer no assistive technology sup= port=20 for Linux users or ISVs. Instead, Red Hat refers blind users to= open=20 source volunteers who work hard to make Linux accessible to bli= nd and=20 otherwise disabled users.=20 Freedom Scientific’s Assistive Technology Software Products= =20 Freedom Scientific develops and offers for sale several softwar= e=20 products that blind and low vision people use to access compute= rs that=20 run on Windows desktop and handheld operating systems. All of o= ur=20 assistive technology software products are developed exclusivel= y for=20 Windows operating systems for several reasons. First, as discus= sed=20 below, Microsoft has outstanding support for assistive technolo= gy=20 products in its operating systems. Second, Windows operating sy= stems=20 are widely used and so our customers want assistive technology = that=20 runs on Windows. Third, because Linux applications are mostly f= ree, we=20 would not create an accessibility aid for the Linux operating s= ystem,=20 even if Linux were to provide operating system support for assi= stive=20 technology products. Freedom Scientific is in business to make = a=20 profit. Our ability to create and license our software for a pr= ofit=20 provides us with the incentive to continue developing products.= We do=20 not believe it would be in our interest to develop a product fo= r Linux=20 that we could not sell profitably.=20 Freedom Scientific’s flagship software product is called = JAWS for=20 Windows, which stands for Job Access With Speech. JAWS is the w= orld’s=20 best selling screen reader. A screen reader is software that co= nverts=20 data on a computer screen to speech and/or Braille so that a bl= ind=20 person can use a computer by having the text of a document read= aloud=20 by the computer or written in Braille on a refreshable Braille = display=20 like the Focus from Freedom Scientific. Among other things, JAW= S=20 allows a blind or low vision user to use many software applicat= ions,=20 such as Microsoft Excel, Microsoft Outlook or Corel WordPerfect= , to=20 create and edit documents, manage financial data, send and rece= ive=20 e-mail messages and to access the Internet. JAWS reads out loud= the=20 information displayed on the computer screen (such as the text = of an=20 e-mail message or a word processing document) when any of these= =20 programs are run. In fact, I used JAWS to read the States’= ; First=20 Amended Proposed Remedy and the draft text of my written direct= =20 testimony. There are approximately 80,000 registered users of J= AWS,=20 most of whom use the product to access the information on a com= puter=20 that they need to perform their jobs.=20 As part of my direct testimony, I intend to conduct a brief, li= ve=20 demonstration of how JAWS enables blind or low vision computer = users=20 to access information from their computers. Among other things,= I will=20 demonstrate how a blind user can create and spell check a word=20 processing document, view Web pages from the Internet, and send= and=20 receive e-mail messages on a computer installed with a Windows = desktop=20 operating system and JAWS.=20 In addition to JAWS, Freedom Scientific makes a product called = Connect=20 Outloud, an Internet access program designed for blind or low v= ision=20 people who are new to computing. Blind and low vision people us= e=20 Connect Outloud to send and receive e-mail, access the Internet= and do=20 simple word processing.=20 JAWS’ Reliance on "Middleware" As an ISV, Freedom Scientific takes advantage of many of the AP= Is=20 exposed by Windows operating systems in developing its software= =20 products. It is advantageous to Freedom Scientific to be able t= o call=20 on operating system functionality and not to have to spend the = time=20 and money developing that functionality in our products. Taking= =20 advantage of the APIs exposed by the operating system lowers th= e entry=20 costs for ISVs and allows them to focus their energy and money = on=20 creating truly innovative products, rather than replicating bas= ic=20 operating system functionality. In other words, because Microso= ft=20 provides the basic "plumbing" in its Windows operating system a= nd=20 makes it available for us to use in developing our products, Fr= eedom=20 Scientific can focus on adding value for its customers.=20 JAWS relies on five components of the Windows operating system = that=20 would be considered "Middleware" under the non-settling States&= #8217;=20 definition, because they expose APIs that ISVs, such as Freedom= =20 Scientific, call on in developing software applications. The fi= ve=20 "Middleware" components that expose APIs that JAWS calls on are= :=20 Microsoft Active Accessibility, Document Object Model, Speech=20 Application Programming Interface, Common Controls and Internet= =20 Explorer. I will discuss each of these below.=20 The Document Object Model ("DOM") is a standard published with = the=20 W3C, a recognized standard setting body. Microsoft’s impl= ementation of=20 the DOM in the Windows operating systems is first-rate. DOM is = a=20 platform- and language-neutral interface that permits programs = and=20 scripts to access and update the content, structure and style o= f a=20 document. For example, JAWS uses the DOM to communicate with Mi= crosoft=20 Excel files. Using the DOM, a JAWS user can query Excel to dete= rmine=20 what the active cell is (i.e., which cell the user’s curs= or is in).=20 Assume the user is told that the active cell is C3. Using the D= OM, the=20 user can then query Excel to find out what the contents of cell= C3=20 are. If the user is told that $250 is in cell C3, using DOM, th= e user=20 can then query Excel whether there is a mathematical formula=20 associated with cell C3. The user might then be told that C3 is= the=20 sum of cells C1 and C2. All of this is possible because DOM all= ows=20 JAWS to communicate with Microsoft Excel. In addition to Micros= oft=20 applications, many third party applications, including Corel= 217;s=20 WordPerfect and Quattro Pro, use Microsoft’s DOM to enabl= e their=20 applications to be accessible to blind users and to other appli= cations=20 that also communicate via the DOM.=20 Microsoft’s Active Accessibility ("MSAA") is a technology= that=20 simplifies the way accessibility aids, such as JAWS, work with=20 applications running on Microsoft Windows operating systems. Us= ing=20 MSAA, software developers can make applications more accessible= to=20 people with vision, hearing, motion and other disabilities. MSA= A is=20 used in magnifiers (software that increases the size of text or= images=20 on the screen for low vision users, like MAGic from Freedom=20 Scientific), screen readers (software that reads aloud or provi= des in=20 Braille the text on the screen), tactile mice (devices that, us= ing=20 force feedback, provide the user with the feel of information o= n the=20 screen), specialized keyboards, speech recognition packages and= many=20 other products.=20 MSAA provides a standard way for accessibility aids, like JAWS,= to=20 access information about user interface elements and to manipul= ate=20 these elements. For example, MSAA enables JAWS to provide a bli= nd user=20 with the choices from a drop down menu in Windows. It also prov= ides=20 application developers who use MSAA greater flexibility in desi= gning=20 the user interfaces of their programs. Third party applications= such=20 as Adobe Acrobat and Lotus Notes have implemented MSAA to enabl= e JAWS=20 to support their applications. As another example, if there wer= e an=20 image of the U.S. Constitution saved in Adobe Acrobat format (i= .e., as=20 a PDF file), a blind user of JAWS could have the text of the=20 Constitution read to him using synthesized speech. This is poss= ible=20 because Windows, JAWS and Adobe Acrobat all support MSAA.=20 SAPI is Microsoft’s speech API. SAPI is a speech-to-text = and speech=20 recognition interface. SAPI enables ISVs, such as Freedom Scien= tific,=20 to communicate with a wide variety of speech synthesizers from=20 different manufacturers via a single, generic interface. Thus, = JAWS is=20 able to read to a user with a much larger collection of voices = in a=20 wide array of different languages without any custom programmin= g done=20 by Freedom Scientific engineers to support the different softwa= re=20 synthesizers. In some countries, SAPI synthesizers are the only= way a=20 blind person can hear the information on their screen.=20 Common Controls are used to create standard dialogue boxes, but= tons,=20 and other on-screen displays in Windows operating systems and i= n=20 applications that run on Windows. Standardized boxes and button= s=20 facilitate the implementation of assistive technology solutions= . For=20 example, JAWS uses Common Controls to make buttons on the Windo= ws=20 desktop talk to a blind user.=20 Internet Explorer is the only web browser that is accessible to= blind=20 and low-vision PC users. Nearly one hundred percent of blind pe= ople=20 who browse the Internet do so using Internet Explorer. JAWS uti= lizes=20 Internet Explorer to allow blind and low vision users to access= the=20 Internet.=20 JAWS Would Break if the "Middleware" on Which It Relies Were Remo= ved. As discussed above, JAWS and Freedom Scientific’s other a= ssistive=20 technology products call on five kinds of "Middleware" in the W= indows=20 operating system, as that term is defined in the States’ = Remedy=20 Proposal, to perform certain functions. (Although Internet Expl= orer is=20 one of the five kinds of Middleware called upon by JAWS, I will= =20 discuss the implications of removing Internet Explorer in a sep= arate=20 section of my testimony because it raises additional issues.) I= f a=20 version of Windows were sold from which an OEM or other Windows= =20 redistributor had removed one or more of MSAA, DOM, SAPI or Com= mon=20 Controls, JAWS would not function properly and a blind or low v= ision=20 user would not be able to use their computer. Moreover, because= the=20 absence of that "Middleware" would be unexpected, the blind or = low=20 vision user may not be able to determine why their computer was= not=20 working. Not only would the absence of any one of these pieces = of=20 "Middleware" break JAWS, but it would also impede the creation = of new=20 assistive technology products to the detriment of blind and low= -vision=20 PC users.=20 If multiple, non-standard versions of Windows containing differ= ent=20 configurations of Microsoft and non-Microsoft Middleware were t= o=20 proliferate, ISVs could no longer rely on the inclusion of Micr= osoft’s=20 Middleware in the operating system. As a result, Windows would = no=20 longer be the stable, consistent platform that it is today.=20 If Freedom Scientific could not be sure that any particular=20 "Middleware" component would be present in Windows, we would be= much=20 less likely to rely upon that Middleware, even if it offered su= perior=20 assistive technology functionality. Depriving ISVs of the oppor= tunity=20 to use the assistive technology Middleware built into the opera= ting=20 system would have harmful consequences for blind and low vision= users.=20 The Removal or Alternation of Internet Explorer Would Deprive Bli= nd and=20 Low Vision Users of the Opportunity To Access the Internet. Internet Explorer is also "Middleware" under the States’ = Proposed=20 Remedy.=20 Internet Explorer is the only commercially available browser th= at is=20 accessible to blind people, because Microsoft has included assi= stive=20 technologies (such as DOM and MSAA) in Internet Explorer. Micro= soft=20 designed Internet Explorer so that Freedom Scientific and other= ISVs=20 that make accessibility aids can use Internet Explorer to provi= de=20 assisted browsing. If Internet Explorer were removed from Windo= ws,=20 blind people would not be able to access the Internet at all. I= f a=20 blind person received a computer with a version of Windows that= did=20 not include Internet Explorer, he or she could not use another = browser=20 to get to the Internet to download Internet Explorer, because t= here is=20 no other browser that is accessible to blind people. A computer= =20 without Internet Explorer deprives a blind person of the opport= unity=20 to access the Internet.=20 If Internet Explorer were open-sourced and different versions o= f=20 Internet Explorer were created that did not implement assistive= =20 technology solutions (such as MSAA and DOM), blind and low visi= on=20 computer users would not be able to access the Internet with th= ese=20 non-standard versions of Internet Explorer. Similarly, if devel= opers=20 who were not aware of MSAA and DOM added new features to open-s= ourced=20 versions of Internet Explorer, those new features would most li= kely be=20 inaccessible to blind and low vision people.=20 Internet Explorer generally points users to the default home pa= ge of=20 MSN (the Microsoft Network). The MSN website is mostly accessib= le to=20 blind people and largely complies with industry standards for c= reating=20 accessible websites. If an OEM or third party were even innocen= tly to=20 change Internet Explorer’s default home page to a website= that did not=20 comply with industry standards for creating accessible websites= , a=20 blind or low vision user might launch Internet Explorer and fin= d their=20 computer talking gibberish because the default website is not=20 accessible. Such a result would, at a minimum, be annoying and=20 unpleasant to blind and low vision users.=20 It is not an acceptable answer that a blind person who received= a=20 version of Windows without Internet Explorer or with an inacces= sible=20 version of Internet Explorer could call Microsoft and request a= copy=20 of Internet Explorer on a CD-ROM. While the blind person waits = for the=20 copy of Internet Explorer to arrive in the mail, he or she cann= ot=20 perform job-related or personal tasks that involve accessing th= e=20 Internet.=20 The Netscape Navigator browser is not accessible to blind and l= ow=20 vision users. About a year ago, Netscape asked Freedom Scientif= ic how=20 it could make its browser accessible to blind and low vision pe= ople.=20 Freedom Scientific advised Netscape that it should implement MS= AA and=20 DOM in its browser to make it accessible to blind and low visio= n=20 users. Netscape is currently working on an accessible browser, = and I=20 have tested an alpha version of a Netscape browser that support= s MSAA=20 and DOM. However, Netscape, in my opinion, is still a long way = from=20 commercially releasing an accessible browser.=20 Because the alpha version of the Netscape browser also relies o= n MSAA=20 and DOM to make it accessible, the Netscape browser would not f= unction=20 if it were installed on a version of Windows that did not inclu= de the=20 MSAA and DOM Middleware. Many other ISVs would also be adversel= y=20 affected by the removal of MSAA and DOM from Windows.=20 Creating Different Versions of Windows Would Greatly Increase Fre= edom=20 Scientific’s Testing Burden and Technical Support Calls. If there were multiple versions of Windows with differing combi= nations=20 of Microsoft and non-Microsoft Middleware, it would dramaticall= y=20 increase the testing burdens on Freedom Scientific and other IS= Vs. The=20 proliferation of different versions of Windows, containing vari= ous=20 combinations of Microsoft and non-Microsoft Middleware, would i= mpose=20 tremendous testing burdens on Freedom Scientific.=20 Freedom Scientific currently employs approximately 8 full-time = testers=20 who test Freedom Scientific’s software products. JAWS con= tains=20 approximately 1,200 innate features, which are tested on each v= ersion=20 of Windows. Freedom Scientific currently tests JAWS on Windows = 95,=20 Windows 98, Windows 98 SE, Windows ME, Windows NT Service Pack = 4, 5=20 and 6, Windows 2000, Windows XP Home and Windows XP Professiona= l.=20 Freedom Scientific also tests JAWS with a wide variety of softw= are=20 applications that our customers use, and with a wide variety of= =20 possible hardware configurations (e.g., sound cards) that might= be=20 installed on our customers’ computers.=20 Multiplying this test matrix by a large number of additional ve= rsions=20 of Windows, each containing different combinations of Microsoft= and=20 non-Microsoft Middleware, would make it prohibitively expensive= for=20 Freedom Scientific to test our products as thoroughly as we do = now. We=20 could not afford to employ the significantly increased number o= f=20 testers we would need to test JAWS against the many versions of= =20 Windows with differing combinations of Microsoft and non-Micros= oft=20 Middleware.=20 Because we could not test each of the different, non-standard v= ersions=20 of Windows, in reality we would likely test our product on a co= uple of=20 major versions of Windows (the Compaq and Dell versions, for ex= ample)=20 and just ship our product without testing it on the other versi= ons.=20 This "solution" is far from ideal, however. For example, if a c= ustomer=20 used JAWS on the Gateway version of Windows that contained a=20 combination of Middleware that we did not test, the customer mi= ght not=20 be able to make JAWS work properly on that version of Windows. = The=20 customer would then have to call Freedom Scientific to ask why = JAWS=20 will not work with their computer. Before we could begin to det= ermine=20 the cause of the problem, we would have to find out what hardwa= re the=20 customer had, what version of Windows the customer had, which=20 Middleware was and was not included in the customer’s ver= sion of=20 Windows, and which applications the customer was running. Only = after=20 Freedom Scientific learned that information from the customer c= ould we=20 begin to determine what caused the problem and how to fix it. E= ven=20 then, because the combination would likely be one that Freedom=20 Scientific had not tested, our ability to determine the cause o= f the=20 problem and to advise the customer how to fix it would be limit= ed.=20 During all of this time, the blind or low vision customer would= not be=20 able to use their computer. If the customer were using the comp= uter to=20 perform her job, they could not work productively (if at all) u= ntil we=20 could resolve the problem.=20 Let me illustrate this problem with an example that did not inv= olve=20 any modifications to Windows. In the past year, some OEMs added= an=20 "Internet Keyboard" to their computers to try to distinguish th= eir=20 products from other OEMs. An "Internet Keyboard" is a keyboard = that=20 contains special new keys, in addition to the standard keys on = most=20 keyboards, that automatically launch a browser to access the In= ternet,=20 can be used to access a Favorites list, or to move Back or Forw= ard=20 among visited web pages. Exactly what these keys do, and how th= ey are=20 configured, varies from OEM to OEM. After Christmas last year, = Freedom=20 Scientific received several dozen technical support calls from = JAWS=20 users who had purchased computers with Internet Keyboards and w= ho=20 called to ask why the Internet keys on their Internet Keyboard = did not=20 perform any functions. Freedom Scientific did not even know tha= t OEMs=20 had added these Internet Keyboards to their computers and thus = was not=20 in a position to make those keys work with JAWS. Anything on a=20 computer that is non-standard makes it more difficult for blind= and=20 low vision people to use computers and for ISVs to make their p= roducts=20 work with every variation or "doo-dad" that an OEM might add to= try to=20 give them a competitive edge. This problem, however, is much si= mpler=20 to address than the problems that would be caused by the States= ’=20 Remedy Proposal. In this case, our users knew there were new ke= ys, and=20 once alerted to the existence of the keys, we explained that JA= WS had=20 not been designed to work with those keys. Under the StatesR= 17; Remedy,=20 however, the actual software code of Windows will be different,= but=20 this difference will not be readily apparent to users or to Fre= edom=20 Scientific. Therefore, the difficulties in designing our softwa= re,=20 testing our software and troubleshooting our products will be=20 increased many times over.=20 Freedom Scientific’s inability to test the various versio= ns of Windows=20 with various Middleware configurations will lead to a dramatic=20 increase in customer support calls as our customers experience=20 problems running JAWS on their particular version of Windows.=20 Responding to customer support calls is expensive for any ISV, = and the=20 additional burden of responding to the increased customer suppo= rt=20 calls caused by non-standard versions of Windows would increase= =20 Freedom Scientific’s costs. We would have to pass those i= ncreased=20 costs along to our customers in one of several ways: we could i= ncrease=20 the price of our software products, we could begin charging our= =20 customers for technical support, or we could do some combinatio= n of=20 those two options.=20 If Freedom Scientific had to expend additional resources to tes= t=20 different, non-standard versions of Windows and to respond to=20 increased customer support calls, we would have fewer financial= and=20 developer resources to devote to adding innovative new features= to our=20 existing products and to developing entirely new products. Blin= d and=20 low vision consumers would suffer from the lack of innovation.=20 Anything that Alters the Consistent User Interface of Windows=20 Disadvantages Blind and Low Vision Computer Users. The consistent layout and operation of the Windows user interfa= ce is=20 very important to people with visual impairments and to assisti= ve=20 technology ISVs. By providing a stable and consistent user inte= rface,=20 Windows operating systems have provided benefits to large numbe= rs of=20 blind and low vision people.=20 Assistive technology products rely upon standardization and=20 predictability. For example, JAWS expects a "button" on the Win= dows=20 desktop to look and act a certain way. This expectation is, in = part,=20 how our products identify buttons, and enables us to relay info= rmation=20 regarding buttons to our blind or visually-impaired users. A St= art=20 Menu or desktop that is non-standard would pose significant=20 difficulties for assistive technology vendors, because not all = items=20 on the desktop (or in the Start Menu) would share the character= istics=20 normally associated with those items. This would increase devel= opment,=20 testing, and technical support costs.=20 For example, the Start Menu in Windows 2000 and Windows XP look= =20 different, because Microsoft added new user interface elements = as part=20 of its improvements to Windows XP. As a result, Freedom Scienti= fic=20 software developers had to do development work to make JAWS wor= k with=20 the Start Menu in Windows XP. We only had to do the development= work=20 once, however, because there is only one new version of the Win= dows XP=20 interface.=20 If there were multiple versions of Windows with different user=20 interfaces, Start Menus or desktops, blind people would be adve= rsely=20 affected. For example, JAWS and our users expect to be able to = find a=20 "My Documents" folder on the Windows desktop or in the new Star= t Menu.=20 Even a minor change by an OEM or a third party to the Windows d= esktop,=20 like removing the "My Documents" folder or renaming the folder = "My=20 Favorite Things," could create problems for the JAWS user. Such= =20 changes result in additional testing burdens and increased supp= ort=20 calls for Freedom Scientific.=20 Moreover, different non-standard versions of Windows with diffe= rent=20 user interfaces make it less likely that blind and low vision u= sers=20 can transfer their knowledge about how personal computers work = from=20 one computer to another. For example, if a blind user was emplo= yed by=20 a company that used the Dell version of Windows, and then switc= hed=20 jobs to a company that used the Gateway version of Windows that= had a=20 different user interface and Start Menu, the blind user may fin= d=20 himself unable to adapt to the Gateway version without addition= al=20 training. For this reason, a standard, consistent platform is=20 especially beneficial to blind and low vision users and their=20 employers.=20 Freedom Scientific does not have the resources to create differ= ent=20 versions of JAWS for different versions of Windows with differe= nt user=20 interfaces, Start Menus or desktops. It is likely that Freedom=20 Scientific would decide only to support certain versions of Win= dows=20 (such as the Compaq or Dell versions) and devote its limited de= veloper=20 resources to making versions of JAWS that worked with the Compa= q or=20 Dell versions of Windows. However, if a blind user worked for a= =20 corporation that used the Gateway version of Windows, that blin= d user=20 may find her computer inaccessible. This will, of course, also = limit=20 choices for blind people by forcing them to use only the most p= opular=20 brands of personal computers.=20 Many of Freedom Scientific’s customers are first-time com= puter users.=20 For that reason, Freedom Scientific distributes detailed traini= ng=20 materials with its products to assist its customers in familiar= izing=20 themselves with their PCs. To the extent that items as basic as= the=20 desktop and Start Menu are reconfigured, Freedom Scientific wou= ld have=20 to create different versions of its training materials that cor= respond=20 with the non-standard versions of Windows. Creating training ma= nuals=20 is already costly for Freedom Scientific. It could be prohibiti= vely=20 expensive to create multiple versions of training materials for= =20 non-standard versions of Windows.=20 Microsoft’s Disclosures of APIs Are Timely and More than Ad= equate. Because Microsoft is continually innovating and adding new feat= ures to=20 improve its operating systems, there is always a risk that a so= ftware=20 product created for an older version of Windows may not work as= well=20 on a newer version of Windows. As part of its efforts to minimi= ze such=20 compatibility problems, Microsoft makes available to ISVs a bet= a test=20 release of Windows well in advance of that operating systemR= 17;s=20 commercial release. This gives Freedom Scientific and other ISV= s=20 sufficient time to test their products to ensure that they will= =20 continue to run well on the Windows operating system, and to de= velop=20 new features that take advantage of new technologies Microsoft = is=20 including in the upcoming version of Windows.=20 Freedom Scientific also subscribes to the Microsoft Developer N= etwork=20 ("MSDN") to get access to disclosures of and technical informat= ion=20 about Microsoft’s APIs. I have found Microsoft’s di= sclosures of APIs=20 on MSDN to be the most thorough collection of documentation abo= ut any=20 desktop operating system available anywhere. As a software deve= loper,=20 I would always like to have more information, of course. Howeve= r, in=20 my experience, not only are Microsoft’s disclosures and t= echnical=20 information more than adequate to enable ISVs to make their pro= ducts=20 take advantage of the Windows APIs, they are the best in the in= dustry.=20 Although it is not good practice, Freedom Scientific does utili= ze some=20 undocumented interfaces in Windows. There are hundreds or thous= ands of=20 undocumented interfaces in Windows, some of which have been pub= lished=20 in books like Undocumented Windows. Some Microsoft software eng= ineers=20 are not even aware of the existence of many of the undocumented= =20 techniques that can be used to gather information from Windows.= I know=20 this because when Freedom Scientific disclosed the techniques w= e used=20 to access various undocumented interfaces in Windows to some Mi= crosoft=20 employees, I recall Microsoft being pretty impressed at our dis= covery.=20 Although we utilize these undocumented interfaces, it is a dang= erous=20 hack. It is dangerous for several reasons. First, the interface= s have=20 not been tested by Microsoft for all the possible consequences = of=20 using them in the way that we do. It is often the case that usi= ng an=20 interface in some untested way can cause unintended or unexpect= ed=20 effects somewhere else in the program. This is especially true = for a=20 program as large and complex as Windows. Second, when Microsoft= =20 documents an interface, it is in effect telling ISVs that we ca= n=20 expect that interface to behave in a predictable fashion for fu= ture=20 versions of Windows (at least until Microsoft tells us it is ma= king a=20 change). Microsoft does all the work needed to ensure, to a gre= at=20 extent, that the thousands of documented Windows APIs work from= =20 version to version for the vast majority of ISVs and our progra= ms.=20 However, when Microsoft modifies Windows to improve it, Microso= ft=20 often changes undocumented interfaces. If Freedom Scientific is= using=20 an undocumented interface, and Microsoft changes it, our progra= m may=20 not operate properly. For those reasons, I do not believe that=20 Microsoft should be require to document these interfaces.=20 I declare under penalty of perjury that the foregoing is true and c= orrect. Executed this ___ day of April 2002 at St. Petersburg, Florida. =20 _____________________ Chris Hofstader Christian D. Hofstader 3019 8th St. North St. Petersburg, FL 33704 (727) 896-6393 (727) 803-8000 x1061 (o) chrish@freedomscientific.com =20 Experience October 1998 – Present Vice President/Software Engineering Director of Software Development Development Manager Freedom Scientific, Inc. http://www.freedomscientific.com Freedom Scientific (formerly Henter-Joyce) is the world’s lar= gest=20 developer of software for blind and visually impaired computer user= s. Its=20 products include: JAWS, a screen reader; MAGic, a screen magnificat= ion=20 program with speech; Open Book, an OCR product which reads books al= oud to=20 its users; and a variety of hardware devices designed specifically = for=20 blind users. All of our products deliver information to the user th= rough=20 text to speech synthesizers and some drive Braille displays. My primary role at Freedom Scientific/Henter-Joyce is to manage the= =20 software development department. My responsibilities include buildi= ng a=20 development team, implementing standard engineering procedures, tra= cking=20 schedules, ensuring timely deliveries of products, budgeting the te= am’s=20 resources, managing third party development and improving communica= tions.=20 Prior to the merger, I reported directly to the General Manager; I = now=20 report to the President of the organization. When I joined Henter-Joyce, there were virtually no formal procedur= es for=20 designing, implementing or testing software. We had no way of track= ing=20 schedules, and products were released in an ad hoc manner. Typicall= y,=20 product releases occurred once every 14 months. During my tenure, w= e have=20 built a formal test department, grew the team from 6 to 25 programm= ers,=20 have improved communications and reporting procedures and have had = 16=20 successful software releases. July 1996 – August 1998 Consultant My services during this period included consulting to both software= and=20 other companies with software engineering needs in a variety of are= as=20 involving a number of engineering related issues. Specifically, I h= ave=20 performed tasks varying from recruiting and team building to=20 internationalizing Windows '95, NT and Macintosh programs to develo= ping a=20 sophisticated education product in HTML, Java and CGI. June 1994 - July 1996 President Ignis Technology Corporation Coordinated the startup phase of the corporation, assembling the st= artup=20 team, leading the incorporation process and developing the organiza= tion=20 throughout its first year. In its first year, Ignis developed Dashboard for Windows '95 for St= arfish=20 Software as well as a number of other products for a variety of oth= er=20 software companies. My role was both project manager and individual= =20 contributor. In its second year, Ignis added management consulting. My role incl= uded=20 project management and development, and I also assumed the lead rol= e in=20 Ignis' management consulting areas. November 1989 - May 1994 Technical Leader=20 Turning Point Software Coordinated the creation of TPS' Technical Leader Committee, which = became=20 responsible for the technical design of all projects developed at T= PS.=20 Other tasks that the Technical Leader group performed included main= tenance=20 of the TPS Coding Standards, code reviews for the engineering staff= and=20 development of TPS cross-platform libraries. In addition to defining and organizing the Technical Leader Committ= ee, I=20 continued my daily activities as a software engineer working on pro= jects=20 for MS Windows in C++. I was also responsible for designing and developing software for MS= -DOS,=20 Microsoft Windows and UNIX based systems. Responsibilities included= =20 scheduling projects, client interaction and design and implementati= on of a=20 wide variety of software products.=20 November 1988 - October 1989=20 Principal Engineer=20 Experra Corporation Responsibilities included writing and maintaining Experra's Extract= Plus=20 financial database product. I designed and implemented a set of cod= ing=20 standards and implemented the usage of software engineering tools. = I was=20 also responsible for the management of the rest of the development = staff. July 1987 - November 1988=20 Software Engineer=20 Number Nine Computer Corporation Responsible for writing software graphics device interfaces between= =20 commercially available CAD/CAM packages and the graphics hardware d= esigned=20 at Number Nine. All the work was on IBM PC/AT/386 machines and was = written=20 in C and Assembly Language. February 1983- July 1987=20 Independent Programming Consultant Clients included Exxon Research and Engineering, Arrow Paper, Ginsb= erg=20 Bros., a major New York Commodities firm and Congressman Mel Reynol= ds. May 1979 - February 1983=20 Programmer=20 Lincoln Savings and Loan Computer programmer. Organizational Experience January 2002 – Present Access Forum This group, funded by GSA helps the Federal Government interpret Se= ction=20 508. Its goals are to help set standards for accessibility and prov= ide the=20 myriad Federal agencies with the information they need to comply wi= th this=20 important legislation. May 2000 – May 2001 Committee on Accessibility for People with Impairments City of St. Petersburg, Florida As a member of this committee, I participated in advising the Mayor= and=20 City Council on issues regarding accessibility. April 1989 - February 1997 Co-founder, President, Director League for Programming Freedom Along with Richard Stallman, I founded this influential organizatio= n and=20 helped develop its growth until it merged with another IP law relat= ed=20 organization. LPF was credited with organizing the friend of the co= urt=20 brief which helped Borland win the lawsuit with Lotus in the Federa= l=20 Circuit Court and the U.S. Supreme Court. =A92002 Microsoft Corporation. All rights reserved.