          COMPUTER SCIENCE UPDATE - SUMMER-WINTER, 1992


                          Published By
                                
                National Federation of the Blind
                       in Computer Science
                                
                    3530 Dupont Avenue North
                  Minneapolis, Minnesota  55412

                     Phone:  (612) 521-3202



Correspondence with Artic Technologies
How is Artic Perceived by the Blind?
by Curtis Chong

     Editor's note:  Artic Technologies International is a
     company best known for its screen access software and
     hardware for the IBM PC and compatible computers.  Anyone
     who has even a passing acquaintance with screen access
     technology for the blind has undoubtedly heard about
     Artic Vision and Artic Business Vision.  In 1986, when
     Artic Vision first emerged, everyone regarded it as a
     highly superior product and Artic Technologies as an
     innovative and forward-looking company.  As the following
     correspondence suggests, this may no longer be the case. 
     In short, Artic Technologies seems to have lost its
     innovative edge in the marketplace of screen access
     technology for the blind.  Does Artic Technologies
     acknowledge this fact, and, perhaps more important, is it
     in a position to do anything about this situation?  Will
     the recent release of Artic's new external speech
     synthesizer, the TransPort, help to improve the company's
     image and, more importantly, restore its leadership
     position?  Read the following correspondence.  It points
     out both good and bad things about Artic Technologies
     International.  It brings to light some facts which
     consumers would do well to ponder.

     One fact that may escape the reader's attention has to do
     with the availability of the SONIX2 module for Artic's
     micro-channel speech card.  Without this module,
     competing screen access systems cannot generate speech
     using Artic's micro-channel speech card.  As of this
     writing, no such module yet exists even though a lot of
     people recall promises made by Artic officials that such
     a module would be released.  Interestingly enough, Mr.
     McDaniel, in his reply, says nothing about this.  Perhaps
     after all is said and done, the TransPort will provide a
     satisfactory solution to this problem.  Only time will
     tell.

      NATIONAL FEDERATION OF THE BLIND IN COMPUTER SCIENCE

3530 Dupont Avenue North
Minneapolis, Minnesota  55412
April 21, 1992

Dale McDaniel
Artic Technologies International
55 Park Street
Suite 2
Troy, Michigan  48083

Dear Dale:

I am writing this letter to bring to your attention some of the
current thinking that has taken place among blind people in this
country concerning Artic Technologies International.  I regret the
circumstances that have motivated me to write this letter. 
However, if Artic Technologies is to maintain any favorable
standing within the blind community and if it is to continue to be
a leading supplier of screen reading technology for the IBM PC and
compatible machines, it would do well to consider the points raised
here.  I trust that you will communicate what I have said to the
other principals of the company.

When Artic Technologies and its program, Artic Vision, came upon
the scene in 1986, it was clear to everyone that here was a company
that had developed a screen reading system with some unique,
innovative, and exciting features.  In 1986, no other screen
reading program for the blind could respond as quickly and as
smoothly as Artic Vision, no other screen reading program could
silence unwanted speech as quickly as Artic Vision, and no other
screen reading system could offer synthetic speech for as
economical a price with as high quality as Artic Technologies could
with its SynPhonix speech card.  Overnight, Artic Technologies had
established itself as a company of prominence in the field.

Alas, events that have taken place over the last few years have
done much to erode Artic's favorable standing.  In my conversations
with blind people throughout the country, these key factors are
cited:

 1.  Artic's perceived attempt to "lock up" the SSI 263 chip market
     by making itself the sole distributor of the chip.

 2.  Artic's tendency to pre-announce software and hardware as
     "soon to be available" when, in reality, many months or even
     years go by before anything is brought to market.  In this
     context I refer to Artic's micro-channel speech card, Vision
     3, the not-yet-released keyboard macro feature for Vision 3,
     and Artic's external speech synthesizer.

 3.  Artic Technologies' decision to release its newer SynPhonix
     cards only with Artic Vision instead of immediately providing
     an open interface (SONIX2) to permit the card to be used with
     competing screen reading programs.

 4.  Packaging the Artic Vision system such that the program can be
     used only with one SynPhonix speech card.

Consider the question of exclusive distribution rights for the SSI
263 speech chip.  While we can engage in endless debate and
discussion about whether or not Artic Technologies had any designs
to monopolize the market for this chip, the fact remains that Artic
Technologies was perceived as attempting to monopolize the chip. 
Proponents of this view argued that Artic Technologies, as a
marketer of screen reading software and speech synthesis technology
for the blind, should not be vying for the exclusive distribution
rights for a speech chip (the SSI 263)--especially, since that chip
was an indispensable component in speech systems marketed by its
competitors.  It was felt that as an exclusive distributor of the
SSI 263 chip, Artic Technologies would be able to exercise an
unfair advantage, thereby stifling competition.  Artic's repeated
denials of any intent to monopolize the chip and its assertion that
it would be the best distributor of the chip because of its ability
to provide technical support were either not heard or not believed
by the vast majority of blind consumers.

A lot of people have told me that they simply do not give any
credence to new announcements made or prototypes shown by Artic
Technologies.  Given the delays encountered getting Vision 3 to
market, the commitments made to put out an as yet unreleased key
macro capability for Vision 3, unfulfilled promises made to develop
a SONIX2 open interface for the SynPhonix micro-channel speech
card, and the not-yet-released but widely talked about external
speech synthesizer, this can hardly be considered surprising. 
There are those who have said that Artic Technologies never
releases a product on time, and some times not at all.  Recent
experiences such as those cited above give substance to this
assertion.

When Artic Technologies developed its new line of speech cards (the
successors to the old SynPhonix 200), it initially sold them only
as part of an Artic Vision or Business Vision package.  At the
initial release date, Artic Technologies chose not to market an
open interface module that vendors of other screen reading software
could use to drive the newer speech cards.  As you know, the old
SynPhonix 200 cards could not run well on computers running at more
than 10 megahertz.  People who wanted the SynPhonix speech cards
for their newer and faster computers found that it was somehow
easier to purchase Artic Vision instead of sticking to their older
screen reading system.  A lot of people thought that the open
interface module should have been released as soon as the new
SynPhonix cards came to market and resented what was perceived to
be a clear attempt by Artic Technologies to increase the sales of
its Artic Vision product.

And what about software protection.  Everyone is well aware that
the Artic Vision product is packaged so as to only work with one
SynPhonix speech card.  This has been and continues to be a bone of
contention with many blind people.  Although no one denies that
Artic Technologies has a right to protect its software, it is
generally felt that Artic Technologies could have come up with a
better protection scheme that would have permitted licensed users
of the program to run the software on more than one machine.  As it
is today, a licensed user who wants to run Artic Vision on two
computers has to buy two copies of the program plus two speech
cards.  This is not a major issue in and of itself.  But when
combined with everything else, it reinforces the perception that
Artic Technologies does not hold the interests of blind consumers
as a high priority.

I would not go so far as to say that Artic Technologies has
deliberately sought to earn the ill will of blind consumers.  It
only seems that way.  The company has taken actions and adopted
policies which have done tremendous damage to its favorable
standing.  There are those who have said that Artic Technologies is
no longer a leader in the field and that Artic Vision, as a screen
reader, has been surpassed by competing software.  There are those
who have said that Artic Technologies promises much and delivers
too little too late.  My belief is that Artic Technologies has made
some bad decisions and that it has made too many premature
announcements.  These have resulted in the loss of the company's
leadership position in the field.

A lot of blind people will tell you that Artic Technologies is
viewed today as a stable if not spectacular company.  Vision 3 is
still viewed as one of the leading screen reading systems on the
market.  However, no one holds out any real hope that the program
will make any significant strides in the near future.

It seems to me that Artic Technologies needs to take deliberate
steps to regain its favorable standing among blind people.  For one
thing, the company must do a better job of announcing new products
and then releasing them on time.  For another, the company must do
a better job of listening to and communicating with blind
consumers.

Consider what might be done to facilitate more open access to your
speech cards on the new 33 megahertz 486 processors.  I understand
that your current speech cards do not work well on some 486
machines.  Rumor has it that you intend to remedy this problem by
making a software patch to Artic Vision.  Why not include the same
fix in your SONIX2 product so that other screen reading software
will be able to use your speech card on the 486 processors.  And
speaking of SONIX2, it is high time that a SONIX2 module was made
available for the micro-channel version of your speech card.

I hope that you and your colleagues will give careful consideration
to the issues raised in this letter.  I do not wish to destroy or
discredit Artic Technologies.  Rather, I am trying to help the
company perform in a manner that is more in keeping with the wishes
of the consumers who purchase its products.

Yours sincerely,

Curtis Chong
President
National Federation of the Blind in Computer Science



                                                      May 4, 1992

Mr. Curtis Chong
President
National Federation of the Blind in Computer Science
3530 Dupont Avenue North
Minneapolis, Minnesota  55412

Dear Curtis,

I would like to thank you for your letter of April twenty-one. 
Constructive criticism is necessary and welcome for any company
that desires to grow and improve.  All too often, individuals or
agencies who take exception with a company's actions are willing to
express their dissatisfaction with everyone but the company itself. 
This can create unfavorable and distorted attitudes among consumers
who have had this dissatisfaction repeated to them second or third
hand.  Not to mention the fact that it makes it difficult if not
impossible for a company to take corrective action if they are not
informed.  We appreciate the fact that you have taken the time to
inform us of your concerns directly and for your expressed
confidence that we can rectify these matters.

Curtis, we are pleased that "A lot of blind people will tell you
that Artic Technologies is viewed today as a stable" company and
that "Vision 3 is still viewed as one of the leading screen reading
systems on the market."  We have watched a number of companies and
many products come and go in the adaptive technology field, and are
proud of the fact that our customers can rely on our continued
presence, enhancements to our existing products, and the
introduction of new products to broaden the horizons of access for
them.

At the same time, I can understand in part the reasons behind your
contention that Artic is currently not a "spectacular company" and
that Vision 3 might not "make any significant strides in the near
future".  While I can understand the frustrations that have led to
such statements, neither I nor anyone at Artic can accept them as
unchangeable fact.  We realize and accept that as part of our
evaluation as a company we have made mistakes.  Be assured that we
have learned and will continue to learn from these mistakes and do
our utmost to justify the confidence that you and so many others
have placed in us.  Artic gained a reputation as an innovative,
technologically advanced company based on our numerous
accomplishments.  Anyone who thinks that this reputation is a thing
of the past is vastly underrating us.

I realize that you ordered the comments in your letter to reflect
a historical perspective and that several of the items have already
been rectified.  However, for ease of presentation, I would like to
respond to the points that you raised in the order in which you
raised them rather than by the order of importance.

1)   Regarding Artic's "perceived attempt to `lock up` the SSI 263
     speech synthesis chip market by making itself the sole
     distributor of the chip", we are neither the sole distributor
     of this chip nor have we ever refused to sell the chip to
     anyone wishing to purchase it.  When this issue was first
     raised at the NFB convention in Chicago several years ago, I
     met with Dr. Cranmer, yourself, other members of the NFB
     staff, and numerous manufacturers of speech products.  At that
     time, I explained the salient points that led to Artic
     becoming a distributor for the SSI 263 chip.  The principals
     of Artic designed the 263 chip while we were with the Research
     & Development group of Votrax, Inc.  SSI simply manufactures
     the chip under license from what was then Votrax's parent
     company (Votrax has since gone out of business).  Since SSI
     did not design the chip, they have never had anyone on their
     staff who understands the operation of the chip or can provide
     technical support.  SSI had contracted Votrax, Inc. to provide
     technical support but, since everyone involved with the 263
     design project had left the company, Votrax had no one on
     staff that could support the chip either.  This situation,
     coupled with the fact that SSI sells relatively few 263 chips
     in comparison to their other products, led SSI to announce
     that they were discontinuing production of the 263.  To ensure
     a continued supply of this chip for ourselves and others,
     Artic entered into an agreement with SSI under which we would
     provide technical support and distribution for the 263 chip. 
     We made it abundantly clear at that time that we would make
     the 263 chip available to any and all individuals, companies,
     or agencies who wished to purchase it.  Since that time, we
     have sold chips to numerous companies and individuals,
     regardless of their application or potential market. 
     Interestingly, those companies who manufacture speech products
     for low vision and blind individuals have not been among our
     customers.  These companies obviously obtain the 263 chip
     through other sources.  Consequently, not only does Artic
     provide the 263 chip to any interested party, we also are not
     the "sole distributor" of this device.  The reality of the
     situation is that, if not for the intervention of Artic, the
     263 chip would have ceased to exist several years ago.

2)   Artic is aware of its "tendency to pre-announce software and
     hardware".  In the past, we have released products late and
     presently have some still awaiting release.  We realize and
     regret the frustration that these situations have created, and
     can assure you that no one feels this frustration more deeply
     than we do.  It obviously does not benefit us in any manner to
     announce a product and then not deliver it by the promised
     date (your letter is simply proof of that).  Unfortunately,
     sophisticated software design and scheduling are not exact
     sciences for any company and the temptation to announce
     exciting new concepts can sometimes be irresistible (MicroSoft
     being two years late with Windows is a good example of these
     points).  At this point, all that we can do is offer our
     apologies for our past mistakes and assure you that they will
     not be repeated in future.  We will no longer discuss any new
     products until we are in a position to quote a guaranteed
     delivery date.  Additionally, products that were previously
     announced will be completed and released as soon as possible. 
     I will be happy to provide you with firm release dates for our
     external speech synthesizer (TransPort) and the keyboard macro
     feature for Vision 3 as they become available.  As you are
     aware, the other products that you mentioned (Vision 3 and the
     micro-channel version of our synthesizer (SynPhonix-315)) have
     been available for well over a year.

3)   Our decision to release our "new SynPhonix cards only with
     Artic Vision instead of immediately providing an open
     interface (SONIX2) to permit the card to be used with
     competing screen reading programs," is in reality an
     indication of our commitment to our customers.  That some
     people mistakenly view this as a negative action on our part
     is obvious from your statement that "A lot of people thought
     that the open interface module should have been released as
     soon as the new SynPhonix cards came to market and resented
     what was perceived to be a clear attempt by Artic Technologies
     to increase the sales of Artic Vision products".  Such
     perceptions illustrate my earlier comment that distorted
     attitudes can develop when people do not discuss their
     dissatisfactions with the manufacturer.  The SynPhonix
     synthesizers marketed with SONIX2 are a completely separate
     product line from the SynPhonix synthesizers with VEST and
     Visions.  Consequently, when a synthesizer upgrade is planned
     we must decide which product will be upgraded first.  If we
     opted to upgrade the SynPhonix with SONIX2 and release it
     first, those customers who rely on our Vision software for
     computer access would find themselves in the position of
     waiting for a new release while those who use other screen
     access programs were already benefiting from an upgraded
     synthesizer.  What message would this deliver regarding
     Artic's loyalty to its customers?  The only other alternative
     would be to upgrade one product line and then delay its
     release until the upgrade of the other product line was
     completed so they could be released simultaneously.  This
     approach might allow us to avoid some criticism, but it would
     cause one group or the other to needlessly wait for an upgrade
     that could be of great assistance to them.  While we are
     committed to considering the needs of the entire low vision
     and blind community, our first loyalty must go to our
     customers who rely upon us for their access needs.

     A related issue that you raised deals with a modification to
     ensure proper operation of SynPhonix synthesizers with the new
     33 megahertz 486 processors.  You state that "Rumor has it
     that you intend to remedy this problem by making a software
     patch to Artic Vision.".  The reality is that any such
     modifications would have to be made to the software that
     controls the synthesizer, not the screen access program which
     uses the synthesizer as an output device.  I can assure you
     that any such modifications to the synthesizer drivers will be
     released simultaneously for all versions of our synthesizers.

4)   Regarding the serialization of Visions software to match the
     SynPhonix synthesizer, you state that such software protection
     "has and continues to be a bone of contention with many blind
     people".  This is an area of great concern to us, as the
     introduction of serialization was an attempt to address an
     unfortunate situation in a manner that did not penalize
     legitimate customers.  When we first released Visions, the
     software was completely open and could operate with any
     SynPhonix synthesizer.  Unfortunately, over the course of
     several years we encountered numerous instances of individuals
     running copies of Vision that they had not purchased.  Many
     times these instances became apparent when the individual
     called us for technical support.  Such instances damage
     everyone in a variety of ways.  They reflect a loss of revenue
     which could have been utilized to hire additional engineering
     staff to enhance existing products and research new products,
     they utilize technical support time that should be going to
     legitimate users, and they divert personnel and resources from
     productive activities that benefit the Artic consumers. 
     Eventually, such occurrences became too frequent and
     unmanageable to be overlooked.  At that point, we had several
     options to resolve the situation.  We could have taken the
     approach used by many large software manufacturers and
     increased our prices to offset these losses.  However, we did
     not then and do not now feel that it is appropriate to
     penalize licensed users for the actions of others.  In fact,
     while our products have increased in sophistication and
     functionality over the years, our prices have not increased. 
     A second option would have been to discontinue selling
     SynPhonix synthesizers for general usage and offering them for
     sale only with a Visions package.  This option was not
     rejected due to any profits that might be realized from such
     synthesizer sales.  In fact, of all the products that we
     manufacture, synthesizers have the lowest profit margin.  As
     difficult as it may be for some people to accept, Artic is not
     simply a faceless company driven by a profit motive.  Artic is
     composed of people who take pride in the fact that their work
     enhances the lives of its consumers rather than simply turning
     out meaningless products for the mass consumer market. 
     Consequently, we did not feel that, in good conscience, we
     could turn our backs on customers who depended on our
     synthesizers.  The third and final alternative was some form
     of software protection.  The most common form of software
     protection is a coded disk that forces the user to perform
     complicated installation procedures and limits the number of
     installations and back-up copies.  Further complicating
     matters was the fact that such procedures would have to be
     performed without benefit of speech.  We felt that these
     complications and limitations were an unacceptable burden to
     our customers.  However, by serializing the software to the
     synthesizer, we were able to provide a system under which
     installation is trivial and the user can maintain back-up
     copies without limitation.  The only difficulty arises when an
     individual needs access to more than one computer.  While you
     state that the user "has to buy two copies of the program plus
     two speech cards.", this is not the case.  From the day that
     we implemented this system, we have offered a complete second
     system to licensed individuals for the price of the
     synthesizer alone.  Consequently, any individual who needs to
     run on more than one system pays for the software once and
     then purchases the additional synthesizer just as they would
     with any other system on the market today.  We also offer
     licenses to accommodate the needs any company or organization
     that requires access for multiple computers.  Even this
     necessity will soon be removed since the introduction of our
     TransPort external synthesizer will enable the user to access
     any computer they happen to be near.

     As I said, our intent was to address an unfortunate situation
     by implementing a solution that did not penalize legitimate
     customers.  As always, we would welcome any alternative
     recommendations that you or any other individual might have.

Curtis, I hope that the responses that I have provided answer the
questions and concerns that you expressed.  I would be happy to
answer any additional questions or provide further clarification
for you or anyone else who desires it.

Again, I deeply appreciate your efforts to assist us in being a
company that meets the needs of the low vision and blind community
to the best of our ability.  I hope that you and others will
continue to provide us with the feedback we require to accomplish
this goal.

                              Sincerely,

                              Dale McDaniel
                              Vice-President of Marketing


Prodigy: Not Meant for the BlinD
by Jim Barbour

     Editor's note:  Jim Barbour served for two years on the
     Board of the National Federation of the Blind in Computer
     Science, relinquishing that position at the 1992 NFBCS
     meeting.  Currently, Jim is employed as a Systems
     Administrator/Programmer by the National Oceanographic
     and Atmospheric Administration (NOAA).

Prodigy is a computer service that is becoming increasingly
available across the country.  When used on a home computer
equipped with a modem, it allows access to a variety of services
including ordering merchandise from several major shopping outlets,
making your own plane reservations, ordering and having delivered
groceries from a local store, playing "online" computer games and
much more.  Unfortunately, because of the way that prodigy
interacts with the IBM Personal Computer and compatible machines,
it is not useable with most screen reading software designed for
the blind.

Prodigy is not the only service of its kind available.  CompuServ,
Genie, and others offer similar services.  Using Prodigy, you pay
a fixed fee per month, rather than a fee based on how long you are
connected (online) to the system.  Prodigy is able to keep its user
fees down because of two sources of revenue:  First, in order to
use the service, you have to purchase special software (available
through local computer software outlets) from Prodigy.  Secondly,
similar to commercial television, Prodigy exposes its users to
numerous paid advertisements, which are interspersed with the
services and information provided by the system.

One thing that distinguishes Prodigy from other online services is
the way in which data is presented to its users.  Most online
services have limited control over the computer screen.  They
display only text.  Prodigy is different.  It communicates with
users entirely through graphics.  It can display text in different
sizes, shapes (fonts) and colors.  It can also display pictures. 
Because of the graphics nature of the system, it has been
traditionally impossible for the text-based screen access systems
used by the blind to work with Prodigy.

Why Won't Screen Access Software Work with Prodigy? There are two
reasons why standard screen reading software won't work with
Prodigy.  First, when text is displayed on an IBM PC screen, it is
also stored as text in the computer's memory.  Screen access
programs rely on this information when they want to find out what
is on the screen.  However, when graphics are used, as with
Prodigy, the text is not kept in memory.  Instead, information is
stored as individual dots (pixels) in the computer's memory. 
Second, the codes used to communicate information from the prodigy
computers to the local PC have not been made public.  This prevents
third party programmers from writing software that could intercept
and interpret the Prodigy data stream.

Both of the above reasons account for the difficulty in making
Prodigy accessible to screen access systems for the blind. 
Imbedded in the data stream sent from the Prodigy computers to your
PC is the actual text that needs to be displayed on the screen.  It
is theoretically possible to capture this information and make use
of it.  However, because the text is buried beneath a sea of
graphic and other computer codes that are difficult to translate,
and because of Prodigy's reluctance to make its coding scheme
public, most developers of screen reading systems have not even
attempted the task.

The NFB in Computer Science and Prodigy  For some time, the
National Federation of the Blind in Computer Science (NFBCS) has
been communicating with Prodigy Services in an effort to achieve
independent access to the system by persons who are blind.  In a
letter to Mike Pepper, a technical assistant to Prodigy, Curtis
Chong, NFBCS President, described the problems blind people were
having with Prodigy and invited Mr. Pepper to the 1990 NFBCS
meeting in Dallas, Texas.  Part of President Chong's letter reads
as follows:

     The success of screen reading programs developed for use
     by the blind rests upon the underlying assumption that
     the information displayed on the computer's video monitor
     is stored in a buffer as ASCII text and not as a pattern
     of bit-mapped graphic elements.  The textual nature of a
     wide variety of computer applications has made it
     possible for blind people to use computers independently
     at home and on the job.  More often than not, we can
     operate the same word processors, database programs,
     spreadsheets, and terminal emulation systems as our
     sighted peers.  This means that we can access the same
     systems and data as our sighted colleagues, enabling us
     to compete on terms of absolute equality.

     As the Prodigy system operates today, everything that is
     displayed on the screen is done so using bit-mapped
     graphics.  In other words, the Prodigy system cannot be
     used independently by the blind; and if the system
     achieves an increasing share of the market, the very real
     possibility exists that blind people on the job will lose
     independent access to information they can readily obtain
     today from other networks.

     There is, in the blind community, more than a little
     annoyance and frustration with Prodigy Services.  There
     is a widely held belief that Prodigy Services doesn't
     really care whether or not blind people can use its
     system.  Some people have reported that their letters and
     phone calls to the company on this subject go unanswered,
     and this only lends credibility to the belief that, to
     Prodigy Services, the blind are of little or no
     consequence.

     ...We, in the National Federation of the Blind in
     Computer Science, are willing and able to work with
     Prodigy Services to have a system that blind people can
     use and benefit from.  This assumes a willingness on the
     part of Prodigy Services to work with us.

In a letter dated June 5, 1990, Mr. Pepper declined President
Chong's invitation and responded in part:

     Prodigy is pursuing ways to provide access to screen
     readers and other adaptive technology.  I have been
     appointed project leader to investigate ways in which the
     PRODIGY service can be adapted for a blind audience.  I
     hope you will be pleased to hear that Prodigy is now
     actively involved with the American Foundation for the
     Blind's National Technology Center.  Over the next
     several months, I will be working closely with Elliot
     Schreier, director of the center.

     ...Since my work is just beginning, I have nothing of
     substance to bring to your Dallas convention at this
     time.  When that changes, I will let you know.  In the
     meantime, I ask your patience and support for my work.

Needless to say, President Chong was not satisfied with Mr.
Pepper's response.  In a letter dated June 24, 1990, President
Chong wrote in part:

     I am pleased that at long last, Prodigy has appointed a
     project leader (you) to investigate ways in which its
     system can be adapted for a blind audience.  As you and
     I have discussed over the telephone, given the current
     technology used by Prodigy, the task promises to be a
     complex one.  It is unfortunate that the original
     technology incorporated into the system was not adapted
     at the outset.

     You indicated in your letter that Prodigy is now actively
     involved with the American Foundation for the Blind's
     National Technology Center and that you will be working
     closely with Elliot Schreier, the Center's director.  I
     have spoken with Mr. Schreier about what he perceives his
     role to be in this project, and he has indicated that his
     interest lies in providing Prodigy with a "wish list" of
     services which it should make accessible to the blind. 
     In other words, Prodigy is not likely to obtain any help
     from the American Foundation for the Blind in the way of
     specific programming or systems design effort.

     If a "wish list" was all that you wanted, you could
     certainly have obtained it by coming to the National
     Federation of the Blind convention in Dallas.  In a very
     real sense, this is where you would have been able to put
     your finger on the pulse of the blind community.  If
     Prodigy is to learn anything of substance about the whys
     and hows of blind people using its services, it must
     learn to deal with blind people and their organizations
     directly.  You would be making a tragic mistake if you
     viewed the American Foundation for the Blind as your
     primary source of information about blind people and
     their use of computer systems.

In the summer of 1991, Interface Systems International, well known
for its FREEDOM1 and ISOS screen reading software products,
announced Prospeak, its program to provide access to Prodigy.  As
of this writing--March, 1992--we have heard nothing to indicate
that prodigy is working toward any kind of a broader solution to
provide access to its systems with all known screen reading
technology for the blind.  It would appear that as soon as
Interface Systems International released Prospeak, Prodigy stopped
all work to make its system accessible to blind persons.  Since at
the very best, Prospeak can only be viewed as a stop gap measure
while Prodigy figures out a way to provide its information in a
textual form, Prodigy is not off the hook.  As far as this writer
is concerned, Prodigy is still responsible for making its system
available in an ascii text format rather than using a graphical
user interface.

About ProSpeak and ISOS  I have tried to use ProSpeak on several
different occasions to gain access to prodigy.  I have yet to get
it to work.  To be fair, I have not spent as much time as I would
have liked with Interface Systems International attempting to fix
the problem.  Initial indications are that it is not a simple
process to use Prospeak with Prodigy.

Prospeak requires the use of Interface Systems International's ISOS
(Interactive Speech Operating System) screen reading program. 
While I found ISOS to be a very powerful product brimming with
online user help, I also found it to be very awkward to use. 
Others have reported that ISOS makes use of key sequences that are
difficult to remember.  They have said that although the online
help is nice, there is no total manual for the program.

This is by no means meant to be an evaluation of either Prospeak or
ISOS, merely my first impressions after having limited experience
with these products.  I am sure there are blind individuals using
both Prospeak and ISOS who are extremely happy with both programs.

I am told that Interface Systems International now offers a product
called Protege, which sells for about $149.  Basically this is a
stand-alone Prodigy interface system that contains just enough ISOS
functions to enable you to use Prodigy.  However, you still cannot
access Prodigy with any other screen reading system.

An Addendum  Fortunately, Prodigy is not the only system to provide
services such as shopping, airline reservations, and online games. 
CompuServ now offers a $7.95 per month flat fee for limited service
on its system, and another good online system to check out is
GENIE.  Since these are text based systems, useable with any ASCII
telecommunications program, I would recommend that blind persons
interested in online services check with them before looking into
Prodigy.


Two Reviews from the
International Braille and Technology Center

     From the Editor:  David Andrews is the director of the
     Federation's International Braille and Technology Center. 
     Readers of the Braille Monitor know that the Center
     contains just about every braille embosser made in the
     world today plus a variety of speech and braille screen
     access systems.  The Center also has a few optical
     character recognition (OCR) systems in-house.  In other
     words, for us "techies," it is a wonder to behold.

     David Andrews has an opportunity to examine and evaluate
     a great variety of software and hardware designed for use
     by the blind.  He has graciously contributed two reviews
     for this issue of Computer Science Update: one on ASAP
     (Automatic Screen Access Program) and the other on the
     new DoubleTalk speech synthesizer.  Here is what Mr.
     Andrews has to say:

Early in 1991, two new products were introduced, the DoubleTalk PC
from RC Systems, and ASAP from MicroTalk.  These products are a
speech synthesizer and a screen review program, respectively. 
Initially, the Double Talk only worked with ASAP, and, initially,
ASAP only supported the DoubleTalk.  Both of these situations have,
over time, changed for the better.  The DoubleTalk PC is today
supported by most screen review programs and ASAP supports a wide
variety of other speech synthesizers.

Here are short reviews of both these products.  You can of course
see them and many other things at the International Braille and
Technology Center for the Blind, located at the National Center for
the Blind, 1800 Johnson Street, Baltimore, MD  21230.  You can also
give us a call at (410) 659-9314 with your questions.

DoubleTalk PC

The perfect speech synthesizer would offer human-like speech at an
affordable price.  While such a device does not yet exist, we may
be one step closer with the introduction of the DoubleTalk PC from
RC Systems, Inc. of Bothell, Washington.

The company, which sells the DoubleTalk PC for $279, is best known
for its Apple products including the SlotBuster and DoubleTalk. 
The unit is a half-slot internal card for IBM PC, XT, AT, 386, 486,
and compatible machines.  The synthesizer is somewhat unusual in
that it does not require the use of one of your computer's
interrupts, as most other internal cards do.  The DoubleTalk PC
only uses an I/O memory address.  There is a jumper on the board to
change this address should any conflicts occur with other software. 
MicroTalk and other companies are selling an external, battery-
powered model of the DoubleTalk PC called the LiteTalk for $395.

One subjective measure of speech synthesizer quality is how easily
it can be understood by a person who has not heard synthesized
speech before.  The DoubleTalk PC scores high using this measure. 
In fact, of all the synthesizers in the International Braille and
Technology Center for the Blind, I would rate it second or third
behind the DEC-Talk for immediate understandability by novices.  

Does this mean that the DoubleTalk PC has DEC-Talk quality speech? 
No, not really.  However, for the most part, the speech is very
clear and understandable.  The majority of synthesized speech users
are familiar with the SSI 263-based speech synthesizers including
the Artic SynPhonix, Votrax Votalker, Accent, Sounding Board and
Braille 'n Speak.  I would describe those synthesizers as sounding,
in general, somewhat mechanical or robot-like.  By comparison, I
would describe the DoubleTalk PC as sounding somewhat electronic. 
There is some high pitched distortion and sibilance in the speech
generated by the synthesizer.  This sound is sometimes called
ailiasing.  Some people hear it as an echo.  Also in my office
computer (a CompuAdd 325), the DoubleTalk PC picks up a little
extraneous noise from the hard disk and other components.  This
does not seem to be a problem in most computers, however.  RC
Systems is aware of this problem and says that it is quite rare. 
Randy Carlstrom, the designer of the board, called me on a
Saturday, with a possible cure.  The fix was not successful, but it
is nice to know that he is working on the problem.  Finally, the
top speed on the original DoubleTalk PC card was not as fast as I
might have liked.  RC Systems has since come out with a new ROM
that increases the top speed somewhat.  The new chip is available
for twenty-five dollars.  While I personally would still prefer a
little more speed from the card, I believe it will be fast enough
for most people.

The DoubleTalk PC synthesizer is not based on a traditional speech
chip such as the SSI 263.  Speech generation is software based and
performed right on the board.  The DoubleTalk PC also has an on-
board microprocessor so that it does not require any resources from
your PC to produce speech.  There is a very small (approximately
1K) driver program, similar to PORTTALK, which can either be run
separately or built into a screen review program.  To my knowledge,
all the screen review programs that currently support the
DoubleTalk PC incorporate this driver code into their innards so
all you will have to do is drop the board into an empty slot and
run your screen review program.  Most screen review programs,
including ASAP, JAWS, Vocal-Eyes, ISOS, Soft Vert, PROVOX and Tiny
Talk, now offer support for the DoubleTalk.

When I first heard the DoubleTalk PC, I liked it very much.  In
fact, I went out and bought one for my personal use.  I have used
the unit a good deal over the past year.  I must admit that now, I
do not like it as much as I once did.  In taking thousands of
persons on tour through the International Braille and Technology
Center for the Blind, I have learned that how a person hears a
particular synthesizer and what he or she likes about it is a very
personal matter.  Thus, what I may like, you may hate.  With this
in mind, it is difficult for me to put my finger on my present
dissatisfaction with the DoubleTalk.  Here are some things I do not
like about the synthesizer.  It is somewhat difficult to
distinguish between some letters and/or words, and it puts an odd
emphasis on some words.  This in combination with the ailiasing
seems to have gotten on my nerves of late.  However, in its
defense, I must also say that some people like it very much,
particularly those who are unfamiliar with synthesized speech.  At
times, the DoubleTalk has a very human-like quality, more than most
other synthesizers.

The DoubleTalk PC from RC systems offers a good combination of
performance, speech quality and value.  It should be considered by
anyone who is looking for a responsive, relatively inexpensive and
understandable speech device.

Automatic Screen Access Program (ASAP)

If you were to ask me what blind people need in terms of computer
access technology, another entry into the already crowded screen
review market would not be high on my list.  Nevertheless, the
latest entry into this field, ASAP, offers some interesting
alternatives and is likely to effect the way we all compute over
the next couple of years.

ASAP stands for Automatic Screen Access Program.  The program can
be purchased for $525 from MicroTalk of Louisville, Kentucky. 
MicroTalk is best known for its apple II products, ProWords,
ProTerm and ProBraille.  ASAP is written by Larry Skutchan, a
skilled blind programmer who understands what tools a blind person
needs to interact with a computer.  Skutchan also works for the
American Printing House for the Blind and is responsible for all
recent developments on the TEXTALKER program used to produce speech
on, among other things, Apple II computers.

As its name implies, ASAP's goal is to provide the PC user with
automatic access to his or her computer and its application
programs without the need to set up configurations, macros, speech
windows etc.  It does this by watching the screen, analyzing
changes, and deciding what you need to know to use the program you
are running.  Skutchan says that the program employs "artificial
intelligence techniques."  The January 1991 Byte Magazine defined
artificial intelligence as the process of making a computer think
like a human.  ASAP, or any other program for that matter,  does
not really do this.  In a later article in the same issue, Byte
further defined artificial intelligence as a technique for making
your computer smarter.  ASAP does accomplish this.  I would choose
to call the techniques Skutchan uses to be "intelligent screen
processing" or some such thing.  However, what it is called is
probably less important than what it accomplishes.

In addition to its automatic features, ASAP is a full-featured
screen review program.  When the program was first introduced, this
was not quite true.  In early versions of ASAP, you could not save
your settings in a configuration file.  You could adjust the speed
of the synthesizer, vary the pitch, set up temporary windows,
determine video attributes to be monitored, etc. via the control
panel, from the command line or batch files; but the program did
not save these settings in configuration files as do conventional
screen review programs.  Presumably, Skutchan's reasoning was that
configurations were not necessary because the program did
everything automatically.  However, besides a wide variety of
refinements, this is the major thing that has changed with ASAP
over the past year and a half.  It is now possible to save all your
settings in configuration files called ".SET" files.  In fact,
".SET" files are saved under the name of the application currently
running and are automatically loaded when that application is run
again.  This system generally works well and is quite automatic. 
All one has to do is issue a Control-S from within the Control
Panel to save all current settings.  While working with a program
called Off-Line Express, I was unable to properly save settings. 
For some reason, ASAP saved my settings under the "ASAP.SET" file
name.  This was the only time I had problems with the feature when
testing ASAP.  

ASAP is basically a command-driven program.  There are no long
menus or hand-holding help-type files.  You will need to read the
manual to learn how to fine tune the program and alter its
settings.  The manual is available on disk, and purchasers get one
cassette and one print copy as well.

ASAP has both a normal operating mode and a review mode.  The
review mode, which is called the "control panel,"  is full
featured.  You control most settings, speed, pitch, punctuation
pronunciation level, etc. from the control panel.  In addition,
there is a command that can be issued from within an application to
execute one review mode command without having to enter the control
panel.  This is called a "one-shot" command and is invoked by first
pressing the zero key on the numeric keypad.  The program also uses
the numeric keypad on the keyboard to manipulate a pointer cursor,
which is called the reading cursor.  This is a separate audio
cursor that can be moved around the screen to read text.  It is
independent of the system cursor and can be moved without going
into a review mode or freezing the running application.  I found
the reading cursor to be very handy for quickly checking things. 
While ASAP will work with the old-style 84 key keyboard, it is much
more convenient to use with an extended 101 key keyboard because of
the presence of the separate numeric keypad and cursor keys.  With
the 101 key keyboard, you can use the numeric keypad for moving the
pointer and the cursor cross for moving the application cursor.

There are two types of commands available from within the Control
Panel.  The first is an Alt-letter or Control-letter combination. 
These commands are used to specify different attributes to monitor,
control level of punctuation pronunciation, and so on.  In general,
you press the proper key combination and are told the default
setting and given a yes/no choice.  You can always cancel a choice
with the Escape key, making it safe to explore from within the
Control Panel.  The second type of Control Panel command involves
typing in a number, then a letter to change that characteristic. 
For example, "9s" would change the speed of your synthesizer to
rate 9, and "2p" would change the pitch to setting two.  Finally,
you can explore the screen character by character, word by word,
line by line, etc. from within the Control Panel.  It is also
important to remember that with the One-Shot Control Panel command
it is possible to issue any Control Panel command from an
application without actually having to enter the Control Panel. 
One minor quibble I have about the Control Panel is that when you
enter it, the program reads you the line your system cursor is on. 
Thus, it might be blank, and say nothing, or will in most cases
just be a repeat of the DOS Prompt.  I personally would prefer that
the program say "Entering Control Panel" or something similar so I
definitely know I am there.

One important feature of any synthesizer/screen review combination
is the ability to quickly cut speech.  ASAP excels in this
department, over all other screen review programs, because of the
number of alternatives it provides.  As with most programs, any key
will interrupt ongoing speech.  The Alt key will cut speech until
another key is pressed.  The Control key will cut speech until the
speech buffer fills up again.  Finally, the Shift key cuts speech
for the current line and starts reading the next line immediately. 
This feature, which is unique to ASAP, allows you to read a screen
of text, skimming down until you find what you are looking for. 
This feature seems to work best with the DoubleTalk PC and not as
well or at all with other supported synthesizers.

I tested ASAP with the DoubleTalk, SynPhonix 215 using "SONIX/TTS"
and an Audapter.  The program works best with the DoubleTalk.  If
you are at all seriously considering the purchase of ASAP, you
should consider purchasing the DoubleTalk as well.  While ASAP did
work with other synthesizers, the line by line speech shutup did
not work.  When using a modem with both the SynPhonix and Audapter,
ASAP had some problems cutting speech cleanly.  Further, it was
difficult to quickly silence the Audapter.  The Audapter is
normally a very responsive speech synthesizer, but ASAP was unable
to utilize the synthesizer to its full potential.

Other unique features of ASAP include the ability to "remember" a
piece of information found on your computer's screen.  This snippet
of information can be recalled later or written directly into an
application.  I found that this feature, while potentially useful,
was somewhat involved to set up.  ASAP can also dial your modem
using a telephone number found on the screen and specified by you. 
The program can also type a word directly into an application or
onto the DOS command line.  The word is indicated by the reading
cursor.

ASAP originally only supported the DoubleTalk PC and LapTalk. 
Thanks to user demand, the program now supports additional
synthesizers including the SynPhonix and Votalker boards, the DEC-
Talk, the Audapter, the Braille 'n Speak, the Echo PC, the Accent
line and the Sounding Board.  A demo version is available from the
author or from various bulletin board systems, including NFB NET
(410) 752-5011.  MicroTalk also offers a bulletin board, the
MicroTalk Support BBS at (502) 893-2269.

The question of course is, does this automatic stuff work?  In
general, the answer is yes.  I have used the program with a variety
of applications, with varying success.  It worked beautifully with
Lotus 1-2-3 and Dbase, fairly well with WordPerfect, Managing Your
Money, and Microsoft Bookshelf and not well at all with Paradox or
Fastback Plus Version 2.1.  It must be remembered that these tests
were conducted without any special configuring or setup.

I suspect that most people will be interested in the WordPerfect
results.  Before the ability to save and retrieve "SET" files was
added to ASAP, it worked fairly well overall.  It automatically
read me the status line at the bottom of the screen, read the full-
screen menus automatically and worked well with the thesaurus.  In
the files menu it read the first part of file names but did not
read the file extensions.  It did work with the spell checker, but
was a little chatty.  It would in some instances read the menu of
choices twice.  You also had to pay attention to what it was saying
to hear the misspelled word.  Overall, the spell checker was
usable, but it took some practice and required the user to pay
attention.  Using the supplied .set files for WordPerfect, speech
is much less sloppy now.  There are configurations for the main
editing screen, the spell checker, help screens and Files Menu. 
ASAP is able to change to the proper configuration automatically. 
The configurations do improve the use of WordPerfect, most
dramatically for the Spell Checker.

I noticed that with some complex screen layouts, such as those used
with Norton Utilities Version 5, or Silver Express (an off-line
reader used by some bulletin boards), ASAP does not necessarily
read the screen in the order you would expect.  It might for
instance read text from the middle, then choices from the top or
bottom.  I presume that this has to do with the way the application
is updating the screen.  Skutchan asks that you bring problem
applications to his attention.  Skutchan states, "it is my goal to
produce the finest screen reader around."  He added, "you haven't
seen nothing yet."  He does update the program on a regular basis,
at least once per month, and the best way to use it is to have a
modem so you can download updates from the MicroTalk or other
BBS's.  When you buy ASAP, you get a "Brand" program that changes
the downloaded demo into a permanent unrestricted copy.

When it works well, ASAP is a valuable aid in running complex
applications.  I use the program to explore complex new CD-ROM
applications and other software.  One possible danger with relying
too heavily on an automatic program is that when it doesn't provide
you with complete or accurate information, you aren't going to know
it.  However, these times seem to be the exception rather than the
rule.  ASAP is constantly being updated by the author.  The manual
originally said that users are entitled to six months of free
upgrades.  My BRAND program has worked for almost a year and a half
and I don't recall any time restrictions being mentioned in the
current manual.  The upgrades can be downloaded from MicroTalk's
bulletin board.  When I found a problem with ASAP working with the
speech-oriented directory lister SDIR, Skutchan had a fix the very
next day.  Another problem I encountered took longer to solve, but
Skutchan was willing to stick with it.  It seemed that when
"Extended Activity Checking" was on I was unable to load the
Arkenstone Reader software on a CompuAdd 286.  Extended Activity
Checking is ASAP's mode that helps it analyze the screen for
changes.  I am now able to do so, although I have heard of others
who have also had problems with ASAP and the Arkenstone.  The
interactions between memory resident software, other applications,
high memory addresses and timing factors can get very complex. 
ASAP, because of its constant screen analysis, may be somewhat more
demanding of your computer resources than other screen review
programs.  You may notice that the program takes a little longer to
react than some.  However, there are much more sluggish programs
out there that do much less.

Overall, I am very impressed with ASAP.  It accomplishes a lot in
a very small amount of memory, about 40K.  It allows a blind person
to use most applications fairly easily and works exceptionally well
with many.  Finally, I think that ASAP would be a good choice for
a beginner or for a more advanced user who is more interested in
running a wide variety of complex applications than in fiddling
with a screen review program.  

It is my belief that other screen review programs will increase
their use of intelligent screen processing techniques to a much
greater degree.  One topic that was recently debated on the
bulletin board circuit was configurability versus automatic
programs.  I don't think it will ultimately be an either or
situation.  I suspect that all the programs will meet somewhere in
the middle.  In the past year, we have seen programs including
Vocal-Eyes and JAWS (just to name two) add some automatic features
to their lineup.

For Further Information

For further information or to see either of these products, contact
the International Braille and Technology Center for the Blind at
(410) 659-9314.  You can also contact the two companies mentioned
above as follows:

RC Systems, Inc.
121 West Winesap Road
Bothell, WA  98012
(206) 672-6909

MicroTalk 
337 S. Peterson
Louisville, KY  40206
Voice:  (502) 897-2705
Modem:  (502) 893-2269
Fax: (502) 895-3022


Managing Your Memory or
Is There Life After 640K?
by Steve Jacobson

     From the Editor:  Steve Jacobson serves as vice president
     of the National Federation of the Blind in Computer
     Science.  He is also employed as a Systems Design Analyst
     by the 3M Corporation.  As a long-time PC user and an
     evaluator of a variety of screen access systems for the
     blind, he has acquired a wealth of knowledge and
     experience in, among other things, memory management
     software and its compatibility with screen access
     programs for the blind.

As screen reading software becomes more and more sophisticated,
there just seems to be less and less memory left for anything else. 
One sometimes gets the feeling that the PC is nothing more than a
talking terminal with some extra memory left over for word
processors and spreadsheets.  If you add device drivers for optical
character readers and CD-ROM drives, you have even less memory
left.  Fortunately, there are several solutions to this problem. 
Having a general knowledge of memory management may assist you in
determining how best to get the most from your computer.

To begin with, you never get something for nothing.  You must have
extra memory in your machine before anything can happen.  There was
a time when those of us having a whopping 640K of memory in our
machines thought that we had small mainframes.  Those times are
long past.  Now, even an 80386 system having 2MB of RAM is thought
to be on the small side.  Although a lot of flexibility can be
achieved with at least an 80386-SX microprocessor, there are some
memory tricks that can be performed with an 80286 and even with
that old IBM-XT with the right software and hardware.  Of course we
have all heard how Microsoft Windows and IBM's OS/2 can make use of
megabytes of memory, but what of our old 640K DOS applications?

Before going any further, let's clarify a few terms.  The type of
memory into which data and programs can be placed is called random
access memory or RAM.  Some data and programs are permanently
stored in your computer in memory that cannot be altered--that is,
in read only memory or ROM.  An example of this kind of memory is
the collection of routines that handle your keyboard, screen, and
disks called the basic input output system or BIOS.  The specific
read-only memory that contains these routines is often referred to
as ROM BIOS.  Every piece of information stored in Ram or Rom is
stored at an "address."  The number of characters that can be
stored in memory is expressed in terms of the number of characters
or "bytes."  Since computer memory usually holds thousands or even
millions of bytes, the terms kilobytes (KB) and megabytes (MB) are
more commonly used when giving statistical information about
memory.  Kilobytes are often referred to with the single letter
"K."  For those of you who enjoy such trivia, one kilobyte of
memory actually represents 1,024 bytes.  (We computer people like
to do everything in terms of powers of two, and 1,024 is actually
two to the tenth power).But let's get back to the subject at hand.

Contrary to popular belief, it isn't precisely true that DOS
applications can only use 640K of memory.  In fact, DOS can
directly access up to 1024K.  In a sense, DOS accesses memory above
the 640K line whenever information is written to your computer's
screen or hard disk.  In fact, memory addresses greater then 640K
are typically used to refer to hardware devices.  Your computer's
screen, hard disk, and BIOS all use addresses between 640K and 1MB
as do some terminal emulator cards, optical character readers, and
network adapters. Unused addresses (those not associated with any
hardware) above the 640K line do not have any real memory
associated with them.  Writing data to these addresses is like
pouring water into a bottomless cup, referred to as a "bit bucket". 
Even if real memory was installed in your computer at addresses
between 640K and 1MB, this still does not enable a DOS application
to address memory above the 1MB line.

The first attempt to overcome this limitation was undertaken
jointly by Lotus Development Corporation, Intel, and Microsoft. 
This common approach to accessing more memory was named the LIM
standard after the three major players involved.  From this
standard comes the two most frequently used memory management
terms, extended memory and expanded memory.  Although you don't
need to be an expert to use either expanded or extended memory, a
general understanding of these terms can prove helpful.

Using extended memory is a little like using a hard disk.  To
illustrate, when a program like WordPerfect saves your document to
disk, it generally needs to know only the name of the document. 
WordPerfect doesn't have to be concerned with where the document
physically is placed on your hard disk.  It lets other software and
hardware figure that out.  The same is true of extended memory. 
Even though a program can't directly address more than 1024K of
RAM, it can pass data to other software in a standard manner and
let that software and hardware deal with the problem.  In DOS 5.0,
the software that handles extended memory is supplied in the form
of a device driver called HIMEM.SYS.  DOS 4.0 calls it XMAEMS.SYS. 
Placing a line in your CONFIG.SYS file such as DEVICE=HIMEM.SYS
installs the software that allows extended memory to be used.

Expanded memory goes one step further.  It takes pieces of extended
memory and moves them in and out of the addresses below 1024K so
that programs can get at the extra memory more easily.  These
pieces of extended memory in this context are called expanded
memory pages, and the memory addresses that contain these pieces of
extended memory are referred to as the expanded memory frame. 
Generally, this frame occupies up to 64K of the address space
between 640K and 1024K, and it must avoid addresses used by other
hardware.  The DOS 5.0 software that enables expanded memory is
EMS386.EXE which, in spite of its "EXE" extension, is loaded as a
device driver in your CONFIG.SYS file.  It can also be run as a DOS
program to provide status information and to exercise some control
over expanded memory.  The DOS 4.0 software that enables expanded
memory is XMA2EMS.SYS. 

It is generally agreed that extended memory is faster while
expanded memory is safer.  However, as we shall see later, it can
sometimes be undesirable to give up that 64K of address space that
the management of expanded memory requires.  Moreover, in order for
extended or expanded memory to be of any value, your software must
be able to use one or the other.  Although many programs do, there
are many that don't.  This leads us into a discussion of other
types of memory managers.

As was mentioned earlier, it does us no good to refer to addresses
for which there is no memory assigned.  Memory managers such as
QEMM386, 386-To-The-Max and DOS 5.0 solve this problem by filling
in the holes between addresses used by hardware with extended
memory.  In other words, pieces of extended memory are connected or
"mapped" to the unused addresses between 640K and 1024K resulting
in several new areas of memory above the 640K boundary becoming
available.  These areas are sometimes called "upper memory blocks,"
particularly in DOS 5.0.  By contrast, the memory below the 640K
line is often referred to as "conventional memory."

For a number of reasons, upper memory blocks are not suitable to
run programs such as WordPerfect.  However, they are well suited
for programs that remain in memory such as speech screen access
programs, print spoolers, and certain DOS functions.  They can also
hold device drivers such as those used for CD-ROM players, optical
character readers, and some speech synthesizers.  This can free up
significant amounts of conventional memory for your word processor
or spreadsheet program.  DOS 5.0 even permits most of itself to be
loaded into a special kind of extended memory called the "high
memory area" or HMA, freeing up even more of the memory below 640K. 
This feature is enabled by placing the command DOS=HIGH in your
CONFIG.SYS file after the HIMEM.SYS line.

To decide whether DOS 5.0 can do everything you need to have done,
one other aspect of memory management must be touched upon.  As we
have already seen, it is possible to access small blocks of memory
above 640K.  These blocks can hold various memory-resident programs
and device drivers, but each one must fit inside a single memory
block.  On the other hand, each block can hold as many programs
and/or device drivers as can be made to fit.  Because of these two
factors, the order in which programs are placed into upper memory
blocks can greatly affect how well memory is used.  

Let's take a simple example.  Assume that we have two available
memory blocks of 30K and 15K.  Further assume that we want to place
two programs into these blocks.  These programs need 25K and 10K. 
Since DOS 5.0 starts with the largest block (in this case, the 30K
block), loading the 10K program first leaves no place to put the
25K program, except, of course, in conventional memory along with
everything else you want to run.  This is the case because even
though there are 35K left, there are only 20K remaining in the
first block and 15K in the second.  Remember, our 25K program must
reside within one block.  If the 25K program is loaded first, there
are 5K left in the first block and 15K in the second.  Our 10K
program will slide nicely into the second block freeing a total of
35K of conventional memory for more important things like that huge
adventure game you want to run.  So what does this have to do with
anything?

Placing programs into these blocks of memory in an effective manner
can be a little like fitting that square peg into a round hole. 
You can't always change the order in which you load your programs. 
For example, what good would it do to load the module that drives
your speech synthesizer after your screen reading program.  If you
find that you just can't do what you would like with DOS 5.0, or if
you do not have DOS 5.0, then it is time to look at other
commercially available memory managers.  The two that are currently
leading the pack are QEMM-386 from Quarterdeck Office Systems and
386-To-The-Max from Qualitas.  As their names imply, these two
programs are designed for computers having at least an 80386-SX
microprocessor.  Other programs are available for 80286 machines
such as Quarterdeck's QRAM.  Both Quarterdeck and Qualitas have
software specifically tailored to work with particular IBM models. 
This type of memory management software has the ability to examine
your batch files to calculate where your memory-resident programs
can best be placed to get the most free memory.  It allows memory
blocks to be filled in any order, and it will work with earlier
versions of DOS.  In some cases, QEMM-386 and 386-To-The-Max even
permit the use of addresses occupied by ROM's to be used for
software.  Although DOS 5.0 can provide some information to aid in
solving problems, 386-To-The-Max and particularly QEMM-386 go a
good deal farther to provide problem solving tools.  Only DOS 5.0
can relocate itself beyond the boundaries of conventional memory,
however.

QEMM-386 also provides a utility called VIDRAM that converts up to
96K immediately above the 640K line to conventional memory.  Since
this memory is used for CGA and EGA graphics, they are disabled
while VIDRAM is active.  However, VIDRAM can easily be turned on
and off, and besides, those of us using speech and braille can do
nicely without graphics.  Of course, programs that magnify the
screen often need this memory to operate.

Now, let's get down to specifics.  On the machine I use on the job,
before I acquired DOS 5.0 or any memory management software, I had
only 340K of memory available for conventional programs after I
loaded the IBM Screen Reader along with a terminal emulator.  Using
expanded memory and DOS 4.0 only improved this figure to 383K.  By
converting my graphics memory to conventional memory with VIDRAM
and relocating my memory-resident programs, I now have 558K in
which to run my software.  If I were to upgrade to DOS 5.0, I could
likely free up another 50K or so.  Periodically, I have experienced
keyboard lockup when switching from terminal emulation to DOS. 
However, this occurs only at the DOS prompt and in every case,
pressing the ENTER key resolves the problem.  I have yet to loose
any data; knock on wood.  The only time I had to reboot was when I
forced Microsoft WORD into graphics mode to see what would happen. 
(Computer people are like that you know.)

If you decide to take the plunge into the world of memory
management, you can bet it won't go right the first time.  Even
with the claims of "foolproof" installation procedures, you will
likely need to do some fine tuning.  It is my hope, though, that
the information set forth here will at least give you a running
start.


Using the Graphical User Interface
by Christopher J. Chaltain

     Editor's note:  Christopher Chaltain is employed by IBM's
     Control Systems Engineering Department as a Senior
     Associate Programmer.  At the 1992 meeting of the
     National Federation of the Blind in Computer Science, Mr.
     Chaltain delivered a speech entitled "Supporting
     Manufacturing with PC Software."  He is one of the few
     blind individuals in this country who is actually
     programming applications that run under IBM's OS/2
     operating system and therefore knows a lot about how to
     use the graphical user interface.  Copies of the speech
     presented by Mr. Chaltain at the 1992 NFBCS meeting can
     be made available by writing to Curtis Chong, President,
     National Federation of the Blind in Computer Science,
     3530 Dupont Avenue North, Minneapolis, MN  55412.

The graphical user interface (GUI), when combined with a
multi-tasking operating system, such as Presentation Manager (PM)
on OS/2 or X-Windows on UNIX, provides a powerful and productive
computing environment for the computer user.  The GUI provides an
intuitive and easy-to-use interface to a multitasking operating
system capable of running many applications at once.  It is easy to
see why these systems are becoming so popular among sighted
computer users.

The blind computer user needs to face the GUI for two reasons.  The
first is that with the growing popularity of these systems in
today's business office, the blind computer user will no longer be
able to find suitable text-based alternatives in the familiar DOS
environment.  The other and less understood reason is that the
blind user stands to gain just as much from the graphical user
interface and a multi-tasking operating system as does the sighted
computer user.  I have been using a prototype of IBM's Screen
Reader/2 program with OS/2 and Presentation Manager (PM) for over
a year now.  Not only does this technology allow me to use the same
software as my sighted colleagues and  develop software for OS/2, 
but it also gives me access to an environment much easier to use
and more powerful than I ever had in DOS.

The problem, as many now know, is that the screen is no longer
"character based" but is now "pixel based."  Instead of 25 rows of
80-column character lines stored in the PC's memory, there is now
a matrix (measuring about a thousand dots wide by a thousand dots
high) of different colored dots stored in memory.  To the sighted
user, these dots make up the windows, icons, and characters of the
screen.  To the blind user, however, these dots are not readable
since the screen access program has no text in memory to access and
send to a synthesizer or braille display.

This problem is remedied with Screen Reader/2, which uses the
concept of an off-screen model (OSM).  OSM was first used by
Berkeley Systems in Outspoken, a product for the Apple Macintosh.
In the off-screen model, system calls that write graphical images
to the screen are intercepted, and a text-based database called the
off-screen model is constructed. This contains all the information
on the screen available to a sighted user, such as icons, windows,
and of course the text. Furthermore, information about the text,
such as font, size, color, and position, is also stored.  It is
this off-screen model which is read by the screen access program.

Another problem for the blind computer user is the wealth of
information now contained on the screen.  The Presentation Manager
screen, called the Desktop, is filled with windows and icons of
currently running applications. Furthermore, each application
running in a window can contain its own set of icons and windows. 
Also, the text can now appear in any one of a number of fonts, come
in any size or color, and be positioned anywhere on the screen. 
There is no tidy arrangement of rows and columns any more.  It is
up to the screen access program to determine what information is
important to the user.  For the blind user, too much information
can be just as detrimental as too little.

Typically, the sighted computer user will navigate around this
crowded screen with a mouse.  By moving the mouse pointer and
clicking on or dragging objects, like icons or window borders,
windows and icons can be resized and rearranged, icons can be
expanded into their windows, windows can be shrunk to icons, and
background windows can be brought to the foreground. Fortunately
for the blind computer user, an IBM standard called the Common User
Access (CUA) provides that any action done with a mouse can be
achieved by key strokes.  For example, holding down the control and
escape keys from anywhere on the system will bring up the "Task
List."  This is a list of all the currently running applications on
the system.  Using the arrow keys and the enter key, any of these
applications can be brought to the foreground.

The separate keypad used by IBM's Screen Reader/2 is also able to
emulate the mouse in several ways.  A pair of keypad key sequences
allows one to cycle through the windows and icons on the Desktop. 
For example, it is possible to set up the user's guide for an
application in one window and the application itself in another
window and then flip between them without going through the Task
List every time. There is also a key sequence on the keypad that
simulates a "double mouse click."  So, as you are exploring the
Desktop, an action bar, pull down menu, or some other set of icons,
you can always choose the last one announced by simply using this
key sequence.

Screen Reader/2 handles all of this action by restricting its view
to the currently active window and announcing the title of newly
active windows as well as newly highlighted objects, such as icons
or menu items.  Thus to start the system editor from a window
containing an OS/2 command prompt, which is equivalent to the DOS
prompt, I would hold down the control and escape keys to bring up
the Task List, and I would hear "Task List" followed by "OS/2
Window," since this item would be highlighted as the last active
window. Next, I would use the down arrow key and hear announcements
like "Communications Manager," "Print Manager," "DOS," and so on,
which are the applications currently running on the system. Upon
hearing "Group - Main," I would hit enter to bring up a list of
programs that can be started.  I would hear "Group - Main" as this
window came up, followed by the highlighted item. Pressing the
arrow keys would lead to announcements of things like "File
Manager," OS/2 Window," OS/2 Full Screen," etc.  Pressing enter
when I heard "OS/2 System Editor" would bring up the system editor,
and Screen Reader/2 would announce "OS/2 System Editor." Now I
would have an empty window, where the cursor and text would be
located; a title, which is OS/2 System Editor; an action bar
containing items like "File", "Options," etc.; and a few icons like
scroll bars.  I could now start entering text into the file, or I
could strike the "F10" key to check out the action bar. The
synthesizer would announce "File," since as the first item in the
action bar, it would become highlighted. The ENTER or arrow keys
would bring down the associated pull down menu, and you could
select "Open" to enter the name of the file you would like to be
editing.

This is not unlike many text-based applications running in DOS. 
These applications (e.g., word processors, spread sheets, and
database applications) used concepts like menus, dialogues, and
windows.  The difference is that under OS/2, these objects are
graphical and not text-based.  These images can be rearranged or
resized on the screen.  Smaller or larger text fonts may be used.
Furthermore, this is true of every application in the system and
not just the word processor or database application.  In fact, the
entire operating system itself is made up of windows and icons. 
Therefore, I don't have to consult the command reference for
obscure system commands; I can move through a series of windows and
menus to format disks, change the time, etc. In addition, OS/2 is
a multitasking operating system.  Thus, in the example above, I
could start a compile in the window with the command prompt, and
while I am waiting for it to finish, I can go to the editor and
start on the next round of changes.  I could also go to the
emulator, and so on.

The move from the more familiar text-based world of DOS to the
graphical world of OS/2 can be gradual for the more hesitant
computer user.  OS/2 supports a "dual boot" option, so that with a
single command, the PC can be rebooted and come up in DOS instead
of OS/2, or back into OS/2 from DOS. There is also a "DOS
compatibility box" in OS/2 that looks just like DOS, and where most
of your DOS programs will run. However, there is no multitasking
when you are in the DOS box; for example, if you start formatting
a diskette in the DOS box and then go to another session in OS/2,
your formatting will stop until you enter the DOS box again.

The OS/2 Full Screen Command Prompt gives you a text-based screen
which looks similar to DOS, but you still have all of the
advantages of multitasking. Therefore, if you start formatting a
diskette in a Full Screen session and then move into another
session or window, the formatting will continue in the background. 
These programs must be written and compiled for OS/2 to run in
"protect mode."  This mode refers to how the 286 or higher Intel
chip operates to allow an operating system, like OS/2, to take
advantage of all the machine's memory and provide the methods for
programs to run at the same time without interfering with one
another.

The DOS compatibility box first appears as an icon on the Desktop,
and is listed in the Task List. When this icon is chosen, the
Desktop disappears, and a DOS-like text-based screen appears,
complete with the DOS prompt.  Similarly, the icon for the OS/2
Full Screen can be found in "Group - Main," the list of
applications available on the system.  The icon for "Group - Main"
can always be found in the Task List.

When I first began to work in OS/2, I often used the dual boot
option, the DOS compatibility box, and the full screen prompt.  Now
that I am more comfortable with the graphical environment of
Presentation Manager and have found more applications which run
under PM, I seldom use anything but the Desktop.

Besides the multitasking, OS/2 and Presentation Manager offer
several advantages to the blind computer user.  Since applications
build their user interfaces by using calls in the system, the
off-screen model automatically knows which window or pull-down menu
is active.  In addition, the model always knows what item is
selected or highlighted.  Since Screen Reader/2 has access to this
information through the off-screen model, there is no need for
special profiles or artificial intelligence to track windows, menus
and selectors.

Furthermore, since PM applications conform to the Common User
Access (CUA) standard mentioned above, all PM applications have a
similar "look and feel."  For example, the Alt key together with
the F4 key always closes the process, the ESC key cancels a pull
down menu or dialogue box, and the F1 key always brings up a help
panel.  This, along with the automatic handling of selectors and
menus, makes using an application as easy for the blind user as it
is for his or her sighted colleague. The blind user will also
benefit since "on line" or "context sensitive help" is also part of
the Common User Access standard.  With built-in help existing for
almost every application and online user's guides for many, the
blind computer user seldom has to scramble around to get
documentation into a useable format.  In the graphical or windowed
environment, it is possible to have the online documentation in one
window and the application in another window.  The blind person can
then quickly switch between the document and the application or
copy text from the document into the application.

The screen access program, as an application on the system, is able
to take advantage of the multitasking and the rich set of system
calls provided by OS/2.  For instance, the Profile Access Language
(PAL), which can be used to customize Screen Reader/2, has been
greatly enhanced in Screen Reader/2.  There is a function in Screen
Reader/2 that will watch the system; and when a particular
application is loaded, any profiles or commands you want will be
automatically run.  These environments switch automatically as you
move from program to program, providing a customized Screen
Reader/2 for each application. Screen Reader/2 can also react to
different events in the system, such as mouse movements, key
strokes, or input on the serial devices.

Screen Reader/2 also allows for multiple devices to be connected
simultaneously.  I am fortunate enough to have a EuroBraille
attached to my system in addition to an Accent SA synthesizer. I
control what text goes to which device through a profile written in
PAL.  For example, when editing a file, I may have the lines in the
file go to both the synthesizer and the braille display. Therefore,
as I hear the text, I am also reading it on the braille display. 
This is great for editing, and noting indentation and
capitalization conventions.  However, when moving down a menu or
across an action bar, I will just have the synthesizer announce the
highlighted text.  This gives the greatest speed and performance.

Sighted computer users have much to gain by using a graphical or
windowed environment, especially when it is part of a multitasking
operating system.  That is why these environments are becoming so
popular in the office and at home.  To compete with our sighted
colleagues, we need to be able to run and access the same
applications.  In addition, however, many of the same advantages a
sighted user has in a graphical interface are shared by blind
users.  I am able to back up my system, edit several files, run
multiple terminal sessions, work with spread sheets, compile
applications, and so on all at the same time. Furthermore, I can
switch from session to session and work in a custom environment
with a single click of a key combination on the key pad. As you can
see, the GUI is not something to be feared; it is something to be
embraced.


The Problem with Printers
by Vicky Winslow

The problem with printers
as I often rail
is the outrageous cost
when the print is in braille.
And although the price
is, in fact, rather dear
the noise these things make
is quite something to hear.
The quiet ones sound
just like popcorn machines
while the loud ones sound like
machine gunning marines.
And over this racket
your co-workers shout,
"We can't even hear what we're thinking about."
To which you reply,
that you'd move it,
you would,
to some other office,
if only you could.
But this dinosaur printer
just bought by the State
weighs at least thirty pounds
or at most sixty-eight.

While printers of ink
are set through a font
from a WordPerfect screen
for the print that you want
braille printers are set
through a menu or two,
four pages are used
while you choose what to do.

Or if it's not that,
then the waste that you get
is from trying to make sure
that the form top is set.

Your frustration is peeked
after sending a file,
that's at least fifty pages
and will print for a while.
Well after a couple of K have gone by,
the printer behaves as if it might die.
It starts brailling garbage,
and to your distress,
you have to word process,
the whole sorry mess.
To cut out the part,
that is already brailled
then reprint at the point
at which printing first failed.

Is it RAM that's the problem?
Is the memory too small?
When you call Tech Support,
they know nothing at all.
Now the bells and the whistles
you can take back and keep
I don't need it to talk,
I don't want it to beep.
But the one little thing
that I want it to do,
is to translate my document
into grade two.
But hold on,
better wait,
if they add what I ask,
it could prove to be so
overwhelming a task,
that the model would change,
and the price would increase,
and the maintenance on
my old printer would cease.
Of course "sorry, no trade-ins:
Who'd want that old thing."
the adaptive equipment
trades people would sing.
So, forget all the things,
I've just recently said.
I think I'll just keep my old printer instead!


National Federation of the Blind
in Computer Science

Minutes
Annual Meeting
June 30, 1992
by Michael Freeman, Secretary

The 1992 annual meeting of the NFB in Computer Science was held in
the  Convention Center in Charlotte, North Carolina, on Tuesday,
June 30, 1992. Registration  began at approximately 1:00 p.m.
President Curtis Chong opened the meeting at 1:30  p.m.

After welcoming remarks and opening announcements President Chong
introduced  the first program item which featured three blind PC
programmers.

SUPPORTING MANUFACTURING WITH PC SOFTWARE was presented by
Christopher  Chaltain, Senior Associate Programmer, Control Systems
Engineering Department,  IBM Corporation. Mr. Chaltain works in a
plant manufacturing IC'S and  maintains operator workstations which
interface with the computers that  control the manufacturing
process. He has found OS/2 and IBM Screen Reader  invaluable and
the multi-tasking capability of OS/2 helpful to his  productivity.
He feels that his colleagues and management respect him and that 
management has a good attitude toward his blindness.

THE BLIND AT WORK USING MICROCOMPUTERS was presented by Brian
Buhrow, Vice  President, California Association of Blind Students.
Mr. Buhrow works in a  university computer center supporting
various types of computers:  UNIX  machines, PC'S, Macintoshes and
Sun workstations. He has found it possible  (without special
equipment) to help people using "unfriendly" systems such as  the
Macintosh by memorizing the various menus and results of choices.
He  programs for various machines in both assembly-language and in
higher-level  languages. He is working on updating the firmware for
the Speaqualizer, the hardware-based screen access system for the
PC developed by the National Federation of the Blind.

Larry Skutchan of MICROTALK presented the program item entitled THE
ART OF PROGRAMMING ON A  MICROCOMPUTER. He began with the lowly
TI-94A, graduating to the APPLE II-E and  then to PC'S. He
concocted a method to review his work on the TI, wrote word 
processors and terminal programs for the APPLE and is the author of
ASAP, a  screen-reader for MS/DOS systems. He emphasized that if
one has the knowledge  and needs a programming tool, one can write
it. He also emphasized that many  of the gimmicks he used for
debugging (beeps and clicks at strategic points in  programs, for
example) have been left in programs because they proved useful  to
all. One of his biggest challenges in writing screen-review
software is that  he must debug the product using that same product
(that is, he is debugging in  an unstable environment). Questions
and discussion followed.

The next item, YOUR PC AS A TERMINAL, was presented by Steve
Jacobson,  Systems Design Analyst, 3M Corporation and by Curtis
Chong, Senior Systems  Programmer, IDS Financial Services. A
synopsis of this topic is omitted as  much of the material may be
found in a recently-published NFBCS article.

The NFB in Computer Science business meeting began at 2:57 p.m. The
minutes  of the 1991 meeting of NFBCS were approved as printed in
the Summer, 1991  issue of Computer Science Update.  Susie Stanzel
next gave the Treasurer's report, indicating that the NFBCS
treasury is in good health.  Elections were held; the following
officers were elected for two-year terms:  Curtis Chong, President;
Steve Jacobson, Vice President; Mike Freeman,  Secretary and Susie
Stanzel, Treasurer. The following persons were elected for 
two-year terms on the NFBCS Board of Directors: Jim Willows, Curtis 
Willoughby and Lloyd Rasmussen.

The "Technical Exchange" followed.

Next came two presentations on IBM'S OS/2 Screen Reader. The first,
entitled  OF GRAPHICAL USER INTERFACES, OS/2 AND WINDOWS
APPLICATIONS, was presented  by James Thatcher, Staff Researcher,
Thomas J. Watson Research Center, IBM  Corporation.  Dr. Thatcher
announced that Screen Reader for OS/2 (Screen Reader/2) will be
available  for delivery in September. He then proceeded to explain
some of the schemes whereby  the OS/2 and graphical environment is
navigated using this new software.

The second item, SCREEN READER FOR OS/2, A USER'S PERSPECTIVE, was
presented  by Cathy Schroeder, Systems Analyst, New Mexico
Department of Labor. She said  that although a new user would be
confused at first, the OS/2 environment  could be mastered.  She
said that Screen Reader did a good job of presenting material  in
a form that the blind could access efficiently. She also said that
IBM was  working on documentation to assist blind users in
accomplishing mouse  movements from the keyboard.

The topic entitled INDEPENDENT ACCESS TO THE WINDOWS PLATFORM was
next presented  by Don Dillin, Senior System Engineer, Government
Special Needs Access Project,  AT&T Federal Systems. After an
introduction outlining the search by his staff  for possible
programs and tools which might be adapted to be used by the blind 
to access the Windows environment, Mr. Dillin demonstrated
SLIMWARE'S  Windowbridge. The system appeared to work, and
information from MicroSoft Word  was indeed spoken. The program has
rough edges, however, and the participants  in the meeting did not
get a chance to try the system for themselves.

The final program item, TOOLKIT FOR ACCESSING WINDOWS, was
presented by Mark Sutton of Berkeley Systems, Inc.  Berkeley
Systems is working on a toolkit that can be used by other screen
access developers to make Windows talk.  The toolkit provides the
"hooks" into Windows while allowing screen access systems to
continue functioning pretty much as they do today.

The meeting was adjourned at 5:00 p.m.

Respectfully submitted,

Michael Freeman, Secretary


Using Your PC as a Terminal

The National Federation of the Blind in Computer Science has
recently produced a publication called "Using Your PC as a
Terminal," written by Curtis Chong (NFBCS President) and Steve
Jacobson (NFBCS Vice President).  This publication discusses a
variety of ways in which your PC can be made to function as a
terminal including communication with ASCII hosts, local area
networks, and IBM 3270 and 5250 emulation.  The publication also
lists some sources for terminal emulation software and hardware and
discusses compatibility between terminal emulation systems and
screen access technology for the blind.

Individuals wanting braille or print copies of the publication
should contact:

Curtis Chong, President
National Federation of the Blind in Computer Science
3530 Dupont Avenue North
Minneapolis, MN  55412
Phone: (612) 521-3202

"Using Your PC as a Terminal" is also available in machine readable
form on NFB-NET, the NFB's electronic bulletin board system (BBS). 
The publication is stored in area 1 as TERMINAL.ZIP and requires
that you have the PKZIP and PKUNZIP software on your PC.

For access to NFB-NET, instruct your modem to call (410) 752-5011.


Roy Zuvers Passes Away
by Curtis Chong

It is with sadness and regret that I must report that early on the
morning of Monday, October 19, Roy Zuvers of Kansas City, Missouri,
passed away as a result of diabetic complications.  Roy was a long
time member of the National Federation of the Blind in Computer
Science, an active participant on many national bulletin board
networks, an active and committed Federationist, and a pioneer for
the blind in the field of computer programming.  A number of blind
people who are currently employed in responsible positions at the
U.S. Department of Agriculture owe their success to Roy.

Roy truly believed in the ability of the blind to get and keep
good, high paying jobs.  He also believed in the ability of the
blind to live in the world as fully productive, first-class
citizens.  Roy worked tirelessly to promote our movement and to
improve public attitudes toward the blind.  I am sure that he will
be missed by all of us.