# Raised Dot Computing Newsletter: Exploring Microcomputer Applications for the Visually Impaired -- ISSN 0890-0019. September-October 1989 -- Volume 8, Number 80.

Published Every Other Month by Raised Dot Computing, Inc., 408 South Baldwin Street, Madison, Wisconsin USA 53703. Telephone: 608-257-9595.

Subscriptions: $18/year Print,$20/year Audio Tape, $30/year Apple II BEX data disk. (Kindly add$20/year for postage outside N. America.)

Submissions are always welcome, especially on diskette. All are subject to editing for style and clarity. All opinions expressed are those of the author. Editors: Jesse Kaysen & Phyllis Herrington.

Entire contents copyright 1989 by Raised Dot Computing, Inc., All Rights Reserved. Nothing may be reprinted in any medium--print, braille, audio, or electronic--without prior written permission from RDC Inc.

Table of Contents: the all-uppercase words name the disk chapters; the words after the equals sign are the actual article titles.

READ ME FIRST = Attention Old Subscribers--Format Changed Again! How To Read the RDC Newsletter on Disk.

CONTENTS SEPT-OCT 89 = This chapter (print page 1).

LASER LINES = Laser Lines from the Editor, includes Foundation Center As Resource for Fundraisers and Staff Changes at RDC (print page 2).

NOW SELLING ZOOMTEXT = RDC Now Selling ZoomText MS-DOS Large Print Access Software (print pages 2 - 4).

NEW CABLES AVAILABLE = Cables Now Available for Apple IIgs and IIc+ Ports (print page 4).

INDEX INTERFACE = Interfacing HumanWare's Index Braillers with BEX, pixCELLS, and Hot Dots by David Holladay (print pages 4 - 6).

ARKENSTONE READER = Arkenstone Reader: Low-cost Modular Reading Machine Technology by Jesse Kaysen (print pages 6 - 8).

COMPARE PB & BNS = The PocketBraille and the Braille Tn Speak: A Comparison by Robert Carter (print pages 8 - 12).

YOHEI INTERVIEW = To The Stars and Beyond: David Holladay Interviews Yohei Yagi (print pages 12 - 14).

BRAILLE REPORT DB = Creating Lovely Braille Reports from IBM and Apple Database Software by Jesse Kaysen (print pages 14 - 18).

DOING HD WITH BATCH = Automating VB File Transfer and Hot Dots Braille Translation with Batch Files by Robert Sweetman (print pages 18 - 19).

COMMENTS & CONREP = Putting Contextual Replace to Work: Non-printing Comments in BEX Chapters by Jesse Kaysen (print pages 19 - 20).

TEACH BEX WITH 3 - 5 = Teaching BEX with 3.5-inch Disk Drives by David Holladay (print pages 20 - 21).

BULLETIN BOARD = Bulletin Board, includes: Braille 'n Speak for Sale; Coalition to Improve Access to Materials for Print-Handicapped Readers; Handy Apple IIgs Graphic Conversion Utility Increases pixCELLS Options; Accessible Cookbooks as Fundraiser for Reading Service (print pages 21 - 23).

FACTS ON FILE = About the Authors; Addresses Mentioned; The RDC Full Cell; Production Notes; Trademarks. (print pages 23 - 24).

F1 DATA DB = Figure 1 from BRAILLE REPORT DB article

F2 TRANSFORM DB = Figure 2 from BRAILLE REPORT DB article

F3 CHANGED DB = Figure 3 from BRAILLE REPORT DB article

F4 BRAILLE ME DB = Figure 4 from BRAILLE REPORT DB article

STRIPBRACE = Transformation chapter discussed in COMMENTS & CONREP article.

## Laser Lines from the Editor

### Foundation Center As Resource for Fundraisers

Following up on Nevin's article last month on fundraising with dignity, Steve Mendelsohn writes:

... There is an interesting new information resource that has recently appeared. With the unprepossesing title of "Grants for Physically and Mentally Disabled," it is the Foundation Center's compendium of over 18,000 grants, totalling more than $71 million, made (mostly but not exclusively) to organizations by US foundations and corporations during 1987-88. The Foundation Center is a clearinghouse that has the largest library on grants and funders in the nation. It publishes a variety of indexes, sorted by geography, area of interest, and a host of other rationales. Their books are expensive; if you can't afford them as an individual, you may be able to convince your local library to purchase them. The "Physically and Mentally Disabled" index is now available for$47 postpaid (they also accept credit card phone orders). For more info, contact:

The Foundation Center

79 Fifth Avenue

New York NY 10003

Telephone: 212-620-4230

### Staff Changes at RDC

If the proverbial Martian asked me, "So, Jesse, what do you like most and least about the sensory aids biz?" my dual-purpose response would be "the rumor mill." While it has been thrilling to be a node on such an active and involved network, it's also been frustrating to see small errors magnified and reputations destroyed. If the rumor mill has been up to its usual standards, many of you already know that, after five fascinating years at RDC, I'm leaving at the end of October. I hope to contribute to the implementation of SGML, which I firmly believe can afford all people equal access to information in any medium.

I offer my heartfelt thanks to all of you for challenging me as a writer and editor. Anyone seeking to sharpen their tech-writing skills should try working within the intriguing constraints of all prose, no pictures. I'll miss almost all my co-workers, who created such a stimulating work environment. (I won't miss Nevin, since we're still very happily married.) Thanks are also due to David and Caryn for the wonderful opportunity to be a Dot. And now, it's Nevin's turn:

I have tried to write this article several times, but my harshest critic [That's me! JK], keeps rejecting it. Here's what I want to say: It's time for me to bid farewell to the many people I've come to know during my five years at RDC. I'll be leaving at the end of October to start on something different. Thanks to everyone who has made it an interesting and unique experience. My time at RDC has been filled with excitement, challenges, and many pleasant moments that I will long remember. I wish the best for all of you in the 1990s.

Both of us leave with this final thought: It's hard to say "au revoir," so we'll just say "au gratin!"

## RDC Now Selling ZoomText MS-DOS Large Print Access Software

ZoomText is a memory resident text magnification package for IBM and compatible computers. It magnifies the display of existing software while you run the programs, and it lets you design the way your display screen looks. You can choose to:

-- Magnify at 2X, 3X ... all the way up to 8X.

-- Magnify all the screen or portions of it.

-- Select one of four different fonts (text styles).

-- Read text at your own speed, or let ZoomText automatically scroll through the text for you at the speed you determine.

ZoomText is specifically tailored for the needs of the visually impaired, opening new opportunities for employment and education. It's simple to install and operate, affording magnified text with no special hardware. ZoomText is compatible with almost every character-based software currently on the market, including many word processors, spreadsheets, data bases and communications programs, as well as DOS itself. ZoomText only magnifies text. If your software mixes text and graphics modes, ZoomText turns itself off, allowing normal graphics display. When the main program returns to text-only mode, ZoomText turns itself back on.

Unlike most enlarging systems, ZoomText does not require any special hardware. You need the following computer equipment to run ZoomText:

-- IBM PC, XT, AT, PS/2 or compatible

-- MS-DOS or PC-DOS version 2.0 or greater

-- EGA or VGA graphics adapter.

-- Any EGA or VGA compatible monitor, like Enhanced Color, Multifrequency, TTL Monochrome, RGB Color Display or VGA Analog.

Please note: an IBM EGA graphics adapter must have a minimum of 128K graphics memory. ZoomText won't work with a CGA color graphics adapter; you need EGA or better.

### ZoomText's Features Let You Get to Work

ZoomText's magnification setting and character font can be changed at any time. The built-in fonts vary in character thickness and intercharacter spacing, resulting in highly legible display. With a monochrome monitor, all character attributes are supported, including reverse video, high intensity, blinking, and underline. With a color monitor, ZoomText supports all foreground and background text colors.

Since enlarging systems display only a portion of the full screen, keeping track of what you're looking at can be difficult. ZoomText offers three solutions:

Scrolling allows you to magnify any portion of the screen at any time, and move this magnified portion around with the arrow keys. Scroll bars indicate the relative position of the magnified window within the normal screen.

Viewing displays the rectangular outlines of the magnified window superimposed on the normal screen. With the arrow keys, you can move the view outline around relative to the normal size screen and then exit to see the text within the view window enlarged.

Tracking automatically scrolls the screen, based on the cursor location. You can control whether ZoomText follows the cursor as you move it in the application, and whether the magnified image is centered over your cursor.

ZoomText has four split screen modes that enable you to see normal size text and magnified text simultaneously. The relative size of the zoomed and unzoomed windows can be easily changed. The right and left zoom modes split the screen vertically, displaying magnified text on one side and normal text on the other. Vertical splits are useful in application programs that scroll the screen, since the entire height of the screen is shown in the unzoomed window. The top and bottom zoom modes split the screen horizontally. These modes are useful for word processing and data entry. A horizontal split can be sized down so only a single line is enlarged, to assist in confirming data entry.

### Powerful Yet Simple to Use

ZoomText offers three ways to control its features, making it easy for users at all levels of computer sophistication. A configurable "hot key" calls up ZoomText's pop-up menus, where you can control zooming modes, magnification, windows, and font styles. Of course, the pop-up menus are easy to see, since they're also enlarged. The "quick key" approach lets you control all modes without having to display the menus. Low-vision users can combine ZoomText with any voice output screen review program to get the advantages of both reading methods. While some screen review programs can't voice ZoomText's pop-up menus, you can still use the command-key interface. Finally, if you're comfortable writing batch files, you'll appreciate the ability to control ZoomText from a command line parameter when it's loaded.

## Interfacing HumanWare's Index Braillers with BEX, pixCELLS, and Hot Dots -- David Holladay

HumanWare sells two braillers carrying the Index label: the Index Classic and the Index Advanced. HumanWare kindly loaned us an Index Advanced so we could learn how it works with RDC software. While we have not worked with the Index Classic, HumanWare tells us that it interfaces just like the Index Advanced. The Index Classic does not do graphics, so you cannot use the Index Classic with pixCELLS, our brailler graphics software.

### Understanding Index's Resets and External Controls

The Index's "reset" and "escape sequence" features are unlike any other unit I've worked with. You can control the Index both through tiny "dip" switches and through commands sent from your computer. Usually, a device only interprets the meaning of the dip switch settings when you turn the power on. What's unusual is that the Index also reads the dip switch settings when you press the square "reset" button on the back. This allows you to change communications parameters on the fly, without turning the Index off and then on again. For example, you can switch from DTR to X-on/X-off handshakes by changing a dip switch and then pressing the reset button. Think of this square "reset" button as the "read the dip switches now" control.

Like most printers, you can send the Index commands that turn on particular features. But on most devices, when you turn the unit off, it forgets what you've sent it. Not the Index. It keeps information from escape sequences in non-volatile memory. For example:

[[<Esc>8 <Esc>M3<Return> <Esc>9]]

changes the built-in left margin to three. The [[<Esc>M3<Return>]] actually changes the left margin. The [[<Esc>8]] puts the Index in "set-up mode" while [[<Esc>9]] takes the Index back from "set-up mode." Even when you leave the unit off for a week, the next time you power up you have a left margin of three.

When you want to restore the Index back to its default value, you don't use the "read the dip switches now" control. Instead, you need to do a "system reset." First you turn the Index off. Then you hold the form feed and paper rocker switches away from you and turn the unit back on. Only this full "system reset" clears away any parameter changes established by escape sequences.

### BEX Serial Interface

The Index brailler offers both serial and parallel interfaces. As always, we prefer the serial connection for an Apple. If you do decide to go parallel, set the S/P switch to P (parallel), and then connect a parallel cable between the Index and your parallel card in your Apple.

For the serial connection, you need to set nine switches on the Index. First, set the S/P switch to S (serial). On the back of the unit, you'll find two banks of switches. For the serial interface, set the eight switches in bank one like this:

off on on off off off on on

This sets 9600 baud, 7 data bits, no parity, 2 stop bits, and DTR handshakes. We can't tell you exactly what the ten switches in bank two do, but our experience shows they should be set as follows:

on off off on on off off off off off

On the Apple side of things, you can use an Apple Super Serial Card set to RDC standard parameters. Connect this to the Index with a straight through male-to-male cable (RDC code 6M). When you are using an Apple IIc serial port, use a 2M cable. When you are using an Apple IIgs or IIc+ serial port, use an 11M cable.

There are four rocker switches on the top. The two right-hand switches control paper movement when the printer is off line. The two left-hand switches affect your BEX use. The left switch is for off line/on line. The next switch controls some built-in reformatting. When you print from BEX, both these switches should be pointing to the front (on-line and normal).

### pixCELLS Interface

These comments only apply to the Index Advanced, because the Index Basic doesn't have a graphics mode. To connect the embosser to your Apple, follow the BEX instructions above in regards to switches, cards, and cables.

You need to "define your own embosser" from the Preferences Menu. The Index requires a very long command sequence to establish the correct dot spacing, carriage width, and margins. When I limited pixCELLS's command sequences to 23 characters, no brailler required more than 16 characters. Unfortunately, the Index requires 24 characters. Fortunately, there is a way out: divide the command sequence into two halves. The first seven command characters are sent by the interface card command sequence, while the last 17 are sent with the embosser's "graphic command" sequence.

Go to pixCELLS's Preferences Menu. Press "I" for interface card. Supply the appropriate slot number. When prompted, make the interface card command:

[[ <Esc> 8 <Esc> W51 <Return> ]]

That's seven characters with no spaces. Finish the sequence with the caret key. When you already have an interface card command sequence, the Index commands must appear at the end.

Now press "E" for Embosser type and choose the highest number to "define your own embosser." Here are the answers to the five questions:

Full graphics: Yes

Dots per line horizontally: 102

Dots per page vertically: 99

Graphics command: [[<Esc> C15,15 <Return> <Esc> D6,45 <Return> <Esc> 9]]

Text command: [[<Esc> 8 <Esc> C15,20 <Return> <Esc> D6,67 <Return> <Esc> 9]]

Yes, the Index's commands include commas. Now you're ready to use the Index to emboss brailler graphics. Always wait for the Index to finish embossing a graphic before turning it off. If you turned off the unit in mid-page, subsequent text output would be incorrectly spaced.

### Hot Dots

We recommend that IBM users take advantage of the Index's parallel connection. Set the S/P switch to P (parallel). The first bank of switch settings control the serial interface, so their setting won't affect this interface at all. The following settings for bank two worked well in our tests:

on off off on on off off off off off

If you want to use a serial connection, set the S/P switch to S (serial). Use the bank two settings listed above, and set bank one as follows:

off on on on off off off on

This sets 9600 baud, 8 data bits, no parity, 1 stop bit, and DTR handshakes. Use an RDC 8M cable to connect the IBM serial port to the Index. To set up the IBM serial port use the DOS command [[MODE COM1: 96,N,8,1,P]]. Hot Dots (and some other software) require you to also use the DOS command [[MODE LPT1:=COM1]].

When you emboss from Hot Dots's menu, use the "Thiel" choice. The Index works fine with the "Thiel" or "vanilla" drivers.

We've been hearing very positive things about the Arkenstone Reader, a hardware/software combination that offers print-handicapped people access to a wide variety of paper documents. Noel Runyan of Personal Data Systems was so impressed with its performance that PDS has become an Arkenstone dealer. Anyone pondering the purchase of reading machine technology should include the Arkenstone in their comparisons. The unit is shipping now, but we haven't tested one ourselves. The following info is based on Arkenstone's promotional material and interviews with Noel Runyan.

In March of 1989, Arkenstone, Inc. was set up as a non-profit corporation to provide technical solutions to charitable needs that are being poorly served by the for-profit and other sectors. The initial board of directors has been drawn from people closely associated with Calera Corp., a leader in OCR and document interpretation technology since 1982. Arkenstone made a special agreement with Calera to market the Calera TrueScan reading machine technology at a dramatic discount to the print-handicapped population. The Reader is available directly from Arkenstone, Inc., and they've also established dealer relationships with Personal Data Systems and other experienced sensory aids folks across the U.S. and Canada--addresses in Facts on File.

By design, the Arkenstone Reader is flexible, affordable, and accurate. Unlike dedicated reading machines, the Arkenstone Reader is a modular system, so you can tailor its capabilities to your needs. You supply the base computer: an IBM PC/AT or PS/2 with at least 512K memory, MS-DOS 3.1 or later, and a 20 MB hard drive, as well as the voice synthesizer and screen access software of your choice.

The Arkenstone Reader has three components: the TrueScan circuit board, its supporting software, and a document scanner. The scanner captures an image of the printed page and hands it over to the circuit board/software, which performs the magic of character and format recognition. Two TrueScan models are available: The Model S lists for $1995, providing 2MB RAM and a 75 character per second recognition rate. The Model E goes for$2995, with 4MB, a 100 cps recognition rate, and the ability to handle documents in both portrait and landscape orientations. (You'd pay 25 to 27 percent more for these Calera TrueScan boards on the general market.)

The Arkenstone supports scores of different scanners, including models from the AST, Canon, Chinon, Datacopy, DEST, Hewlett Packard, Microtek, and Panasonic product lines. You're not tied to a single scanner, so you can balance cost and capabilities to meet your reading patterns. The TrueScan board interprets the scanned information and saves it as text on IBM disk in any of forty formats. If you don't already own a scanner, you can purchase a Hewlett Packard ScanJet from Arkenstone for just $1245, which includes a "daughterboard" card that attaches to the TrueScan card (so you don't need a separate interface card). Personal Data Systems has put together a package deal for those just starting out with computers:$6000 gets you the Arkenstone Reader, the HP ScanJet and daughterboard, and the IBM-PC hardware with an Audapter voice synthesizer. In addition PDS includes their "EasyScan" voice-friendly user interface that allows true beginners to read documents at the press of a spacebar. Noel feels the Arkenstone compares quite favorably to the Kurzweil Personal Reader: similar accuracy, more format information, and you get an IBM to boot for half the price! All of the Arkenstone's components are mass-market items: you can turn to any computer dealer for service. With a KPR, you need another computer to capture the scanned text for further processing. The Arkenstone always saves your text on disk, where it's ready for later review, editing in your word processor, or translation to braille for hardcopy.

The Arkenstone Reader can handle typewritten, typeset, laser printed, and draft dot matrix print from 6 to 28 point, monospaced or proportional. It can "decolumnize" multiple columns and tables, work with pages that mix text and graphics, and store format information (like italics and headings) in more than 40 program-specific files, including WordPerfect, Microsoft Word, Lotus 1-2-3, and WordStar. To quote from the glowing review in the March 28, 1989 PC Magazine, the unit "blew through all our scanner test samples very quickly and read everything except text with gray background. We also gave it newsprint, magazine articles, and a variety of other materials to test its performance. The fuzzy answer to the performance question is that the TrueScan did very well with these samples; its success rates across all unshaded samples outperformed even the $15,590 XIS Kurzweil K-5000." Noel echoed this sentiment. "I really appreciate the Arkenstone's ability to read my magic books, which freely mix text and pictures. With this reading machine technology, I can develop new magic tricks independently and surprise Debbie! On a more serious note, we set up a machinist with an Arkenstone Reader, and he has loved it. When it comes to the detailed tables in the five-inch thick Machinery Handbook, audio tape versions are useless and getting the whole thing transcribed in braille would be out of the question. The Arkenstone enabled him to scan the tables he needed and generate his own braille references." But, Noel cautioned, a reading machine is not a panacea. "Some reading tasks simply don't lend themselves to an OCR approach. The Arkenstone needs time to interpret a page, taking up to 30 seconds before it begins talking. Finding one definition in the dictionary or comparing two works at once are frustrating tasks with any reading machine, because the machine can't 'skim' like a human reader." Recognizing the importance of customer satisfaction and real-world tests, Arkenstone Inc. insists that all its distributors offer a 30-day money-back guarantee. As they state in their flyer, "Arkenstone was founded as a non-profit organization to make reading technology as accessible as possible to the people who need it. Arkenstone is in existence to help you read easily!" From the sounds of it, they're meeting those goals. ## The PocketBraille and the Braille 'n Speak: A Comparison -- Robert Carter, Ph.D Editor's Note: A longer version of this article appeared in the SUMMER 1989 issue of Tactic, the braille quarterly on adaptive technology--subscription information in "Facts on File." My thanks to Deborah Kendrick for permission to excerpt it. When compared with today's powerful microcomputers, the pencil is not commonly considered a sophisticated information processing tool. However, for only a few pennies, the pencil user can store and retrieve information quickly, quietly, and reliably. That same pencil can be used to reorganize, add to, or delete some or all of the previously written information. Unfortunately, for persons who do not have sufficient vision to use the pencil efficiently, equally convenient tools have only recently come into existence. Portable microcomputers equipped with synthetic speech, braille, or large print output now enable blind people to manipulate information with perhaps even greater ease than their sighted colleagues. This article examines two highly portable, reasonably priced, extremely powerful computer systems that allow blind people to store, edit, organize, and retrieve information in ways far superior to pencil and paper. One of these devices is called the Braille 'n Speak and is made by Blazie Engineering of Maryland. The other device is manufactured by the American Printing House for the Blind of Louisville, Kentucky, and is known as the APH PocketBraille. There are a number of talking portable "laptop" computers on the market. The Braille 'n Speak and PocketBraille are the least expensive, most portable, and arguably the easiest to use of all such products. The PocketBraille is$905; the Braille 'n Speak ranges from $905 to$1200, depending on the options you get. Their portability is largely due to their compact seven-key braille input system, which requires much less space than a typewriter-style keyboard. They both have built-in speech synthesizers for output, and both have "nonvolatile" (permanent) memory for storing the data that is entered into them. Because their software is permanently stored and custom-tailored, these devices are easier to use than most other talking portable computers. On the other hand, the permanent software means neither computer can run off-the-shelf software. They both come with a built-in text editor, clock, and calendar. For the Braille 'n Speak, a calculator, stop watch, and countdown timer is also available. I have used both devices on a daily basis for a month, and will devote the remainder of this article to some important differences between them.

### Physical Design and Portability

When it comes to portability, both systems get high marks. The PocketBraille is 8-1/2 inches wide, 4-1/2 inches deep, 2 inches thick, and just under 3 pounds in weight. The Braille 'n Speak is 8 inches wide, 4-1/2 inches deep, 3/4 inches thick, and weighs 13 ounces. Because it is thinner and lighter, the Braille 'n Speak is a bit more portable than the PocketBraille. The Braille 'n Speak will fit into the pocket of most suit jackets or overcoats, or in a purse or small camera bag. The PocketBraille, being thicker, requires more space--for example, the supplied carrying case with shoulder strap. After carrying both products for several hours, the smaller size of the Braille 'n Speak becomes significant.

While both units place their keyboards on top, they are quite different in design. The PocketBraille's seven keys are small round rubber "buttons" that are indented in the center, with a wobbly feel that takes some getting used to. To the user's advantage, they are completely silent. The Braille 'n Speak's seven keys are long and narrow. Similar to a Perkins braillers', they have a sturdy solid feel. While this keyboard is not as silent as the PocketBraille's, it does not make enough noise to be distracting to others in a classroom or meeting situation.

Along its right side, the Braille 'n Speak has an earphone/external speaker jack and a slightly recessed on/off switch. On the left side of the unit there is a battery charger jack and an RS232 serial port connector. To save space, the serial port connector is smaller than the standard 25 pin connector. You must use the special adapter (provided with the unit) to connect with standard cables.

The PocketBraille has four ports, two knobs, and an on/off switch, all located on the back of the unit. The two knobs control the volume and pitch of the built-in speech synthesizer. There is a standard RS232 serial port, a Centronics parallel port, a port for an optional braille display, and a port designed to be used with either the Screen Door or additional memory modules. (Both the Screen Door and braille display are under development and not yet available. More details on these options appear in the "Interfacing Flexibility" section below.)

While it is convenient to have a volume and pitch knob, they protrude from the back of the PocketBraille and seem vulnerable. When placing the unit in a backpack, the knobs and ports tend to catch on the pack itself or items in it. The Braille 'n Speak's volume and pitch are controlled from the keyboard, which means no protruding knobs, and no snagging problems.

### Storing, Retrieving and Organizing Information

Both units save data in nonvolatile, or permanent, memory, so both can be turned off at any time without losing what's been entered. But as far as storing, retrieving, and organizing that information, the differences are significant. The PocketBraille has 224K of free memory, and the Braille 'n Speak has 180K. Assuming that the average word is 10 characters long, the PocketBraille can hold a maximum of 22,937 words in its memory at any one time, whereas the Braille 'n Speak's memory could accommodate 18,432 words.

At first glance, the fact that the PocketBraille can store more data would appear to be a major advantage. Unlike the Braille 'n Speak, however, the PocketBraille has no internal system to help the user organize and retrieve the information that is stored. In the PocketBraille, all 224K of memory is either treated as one huge block or as seven 32K memory "banks." Unfortunately, switching between banks is awkward. It's difficult to tell which bank is being used, and since the banks are identified only by numbers, it is difficult to remember what is stored where.

The Braille 'n Speak, on the other hand, has a real file management system. The user divides the memory into up to 30 files as needed. Because each file has a name, organizing and retrieving data is much simpler. A student could create separate "Math," "English," and "Science" files, allowing her to go straight to the data she wants to work with. The Braille 'n Speak also has a file management menu, where files can be created, deleted, opened, renamed, made bigger or smaller, protected so they cannot accidentally be modified, and so on. Moving in and out of files is extremely fast and easy, too.

In summary, both of these devices have an incredible amount of storage capacity. After extensive use, I have never come close to running out of memory on either device. The important issue here is not the amount of memory, but how that memory is managed. The fact is that the Braille 'n Speak has a memory management system and the PocketBraille does not. Without such a management system, the 224K of memory in the PocketBraille quickly becomes confusing and unwieldy.

### Word Processing Features

In many ways, the PocketBraille and Braille 'n Speak's text editors are similar. For the most part, the commands that are used to write, review, insert, delete, format, and find text are more or less the same on both machines. There are some important differences, however. Only the Braille 'n Speak has a cut and paste feature which enables the user to move blocks of text from one place to another. The PocketBraille, on the other hand, has the ability to number pages when sending text to a braille embosser or printer, while the Braille 'n Speak can format text into pages, but can't number them.

Unless the user is in insert mode, newly written text on either machine is always appended to the end of pre-existing text. This arrangement prevents the user from accidentally writing over existing text. For at least two reasons, this feature works significantly better on the Braille 'n Speak.

Because the user can divide the Braille 'n Speak memory into individual named files, it's much easier to group like information together. Suppose the user creates a file called "Phone" to store telephone messages. Each time the phone rings, he can open the "Phone" file and begin writing. The new information is stored immediately after whatever text was already present in the currently opened file.

The PocketBraille's lack of a true file manager means text is appended to the end of a large block of memory, not to the end of a specific file. Thus, if the PocketBraille user turns the machine on to quickly write down a phone number, that number will appear as the last thing in memory. If, between phone messages, the user switches on the PocketBraille to create a grocery list, write down the bus schedule, and take some notes for a class, phone messages will be scattered throughout these other texts--unless care is taken to write all messages in insert mode, thus grouping them together. Obviously, data can quickly become disorganized and sorting through it can become confusing. Since the PocketBraille provides no good way to organize the data in its memory, much of the advantage of having new text always appear at the end of existing text is lost.

The second reason that automatically appending text works better on the Braille 'n Speak relates to the different cursor design. Both machines operate with two cursors: writing and reading. The writing cursor moves each time a new character is written. The separate reading cursor moves as information is read by the speech synthesizer. For example, it is the reading cursor that jumps forward one word when the user gives the command to speak the next word.

In the PocketBraille editor, entering information with the keyboard only moves the writing, not the reading, cursor. The reading cursor only moves when the user issues commands to read a portion of the text. For example, if the PocketBraille user writes three lines of text and commands the system to read the current line, she will hear the first line, not the third line just written. To actually hear that third line, the user would have to command the PocketBraille's reading cursor to jump to the end of the text and then read the current line.

The Braille 'n Speak, on the other hand, has a feature called "cursor tracking." When cursor tracking's off, the Braille 'n Speak's writing and reading cursors are separated like the PocketBraille's. But when the cursor tracking feature is on, the reading cursor keeps up with the writing cursor. Thus, with cursor tracking on, when the user asks for the current line, the most recently written line is spoken. After having used this feature on the Braille 'n Speak, it was sorely missed on the PocketBraille. Without cursor tracking, the user cannot take full advantage of the conveniences of having new text automatically appended to the end of existing text.

Another difference in cursor control concerns where the cursor lands when moving backward and forward a word at a time. In the case of the Braille 'n Speak, the cursor always lands on the first character of a word. Therefore, when the user commands the Braille 'n Speak to speak the current word, the word that the cursor is sitting on is spoken. This is also true when the "delete current word" command is issued. The word that the cursor is on will be deleted.

Where the PocketBraille's cursor lands depends on which direction you're moving. When moving forward a word at a time, the PocketBraille speaks the word and the cursor stops at the space after the end of that word. When moving backward through the text, the cursor is then left on the space preceding the last spoken word. This leads to confusion when the "speak current word" and "delete current word" commands are used. The PocketBraille user must make certain that the cursor is positioned at the beginning of the word before issuing the delete current word command or the wrong word will be deleted. Taking time and care to position the cursor before deleting a word is awkward and inconvenient.

Both machines provide the user with excellent control over the speech synthesizer when reviewing material. The user can have the current character, word, line, or entire text spoken, as well as move around within text by lines or paragraphs. Both systems allow the user to select the amount of punctuation that is spoken and to choose whether numbers will be spoken as full numbers or individual digits. Unlike the PocketBraille, the Braille 'n Speak can be commanded to speak phonetic character equivalents. To ascertain whether a character is b or t, the Braille 'n Speak can be commanded to say "bravo" or "tango" for clarification.

In conclusion, both machines are quite good rudimentary word processors, providing the user with basic word processing tools. Its cut and paste feature and file management system makes the Braille 'n Speak more powerful and flexible than the PocketBraille.

### Interfacing Options

One important strength of both devices is their interfacing capabilities. Both can be used with other microcomputers, telephone modems, optical scanners, paperless braille devices, inkprint printers, and braille embossers. Further, they can serve as battery-powered speech synthesizers for other equipment.

The Braille 'n Speak has an RS232 serial interface, which means any equipment that one wishes to connect must also have an RS232 serial port. Serial ports are common, but some computers and printers have only Centronics parallel interface ports. In this case, a Braille 'n Speak interface requires either a parallel-to-serial converter box, or outfitting the other device with a serial port. Currently, the Braille 'n Speak has a second port which is not used.

In addition to its RS232 port, the PocketBraille offers three more interfacing options than the Braille 'n Speak. The PocketBraille's Centronics parallel port means an easy connection to both parallel and serial devices. Further, the PocketBraille has a port for an optional braille display, which is currently being developed by the American Printing House for the Blind. The fourth and final interface is the Screen Door port. It has this unusual and interesting name because it was designed to connect the PocketBraille to a "Screen Door" circuit card, inside either an Apple or IBM computer, that would send everything that goes to the computer's screen to the PocketBraille. This means that a blind computer user could use the PocketBraille to read the screen of Apple and IBM software. Unfortunately, like the braille display, the Screen Door is not yet available. The American Printing House has been advertising the Screen Door for the last couple of years but has not yet been successful at bringing it to market.

In addition to being used as the connecting port for the Screen Door circuit card, this fourth port permits the PocketBraille user to connect external memory modules. These external memory modules allow the PocketBraille user to expand the PocketBraille's memory. Each module will hold 32 kilobytes of data. I was unable to obtain any of these memory modules for testing, and cannot, therefore, comment on their usefulness. Assuming that they do work well, however, this is an important possibility for anyone who requires more than the 224K of internal memory.

One of the most exciting features both devices have is the ability to interface with telephone modems. The modem connection offers the user access to everything from computer bulletin boards to the major information services such as GEnie and CompuServe. Here a user can shop electronically, conduct research, read the news, use an encyclopedia and much more. The Braille 'n Speak's "interactive mode" makes it a talking terminal that works extremely well with a modem. Everything that the remote computer sends to the Braille 'n Speak is spoken as it comes in through the serial port. While the PocketBraille can be used with a modem, it lacks an interactive mode, forcing the user to review incoming text with the reading cursor commands. This is somewhat time-consuming and can be a bit awkward.

Finally, both devices work well with either braille embossers or inkprint printers. Both systems have reverse braille translators that can be turned on or off at any time. The user can write in grade 2 and the reverse translator, when turned on, will expand the grade 2 contractions to regular text. This means that the text can either be sent to an inkprint printer or be spoken properly by the internal speech synthesizer. The translator can be turned off, and the same grade 2 information can be sent to a braille embosser or paperless braille device in its native form. The reverse braille translators in both machines work extremely well.

### Conclusion

Because of their flexibility, power, and portability, the Braille 'n Speak and PocketBraille represent a major advancement for blind people. They allow blind people to process and manipulate information in ways that have never before been possible. I conducted my research in May 1989, and this technology is new and will continue to evolve.

Use the manufacturer's addresses in "Facts on File" to check if a feature you want is now available. At this time, the Braille 'n Speak is a fully developed, well integrated information handling tool. The PocketBraille has much potential but the pieces have yet to be put together in a fully functioning package.

## To The Stars and Beyond: David Holladay Interviews Yohei Yagi

While we were developing MathematiX, we were lucky to have Mr. Yohei Yagi as a beta tester. His considerable intelligence and newness to the Apple combined to isolate some thorny problems. I thought our readers might be intrigued to see how Yohei Yagi puts sensory aids technology to work.

Can you give us a little background on where you came from and what you're doing?

I'm from Japan. I've been in the United States for 20 months. I'm a graduate student in the Astronomy Department at the University of Texas. I just finished the first year. Now I'm not sure whether I should continue to study for a Ph.D. or whether I should stop at a master's degree. My interest is in Cosmology. The topic of my master's thesis is "The Magnetic Field in the Galaxy Formation". I lost my sight from age three to six gradually.

What equipment are you using?

I have a VersaBraille II Plus, an Apple IIgs with 1 megabyte expanded memory, an ImageWriter II, and an Optacon. For software, I use BEX and MathematiX from Raised Dot Computing. I'm planning to get a Kurzweil Personal Reader next month. Currently, I'm using the Kurzweil Reading Machine and the VersaPoint in a library at the University of Texas. I don't have a braille printer at home.

How did you learn Nemeth Code (the braille code for mathematical and science notation used in the U.S.)?

I learned Nemeth Code at the age of eighteen when I entered college to major in physics in Japan. In Japan, we have our own braille system for scientific notation, so I didn't have to use Nemeth code in high school. There are very few transcribers in Japan who can handle college level mathematics or science (and one of those transcribers uses Nemeth Code). So I had no choice and studied Nemeth Code by reading Nemeth Code transcriptions and by reading the Nemeth Code Book.

Describe how you go about producing technical inkprint documents.

Usually I write everything on the VersaBraille using Nemeth Code. When I produce inkprint, I just change my own personal style of Nemeth Code to the style required by MathematiX. I pay special attention to switching between literary and math modes, formatting issues, and spacing. I proofread what I have written carefully before I ask for an inkprint copy. I use the verbalization feature of MathematiX to find structure errors (errors caused by contradictory information in the input). Since English is not my native language, I prefer proofreading in braille to listening to voice output. It doesn't take me long to produce readable technical inkprint output, unless there is a structure error which I never previously encountered.

Does the MathematiX system contain all the inkprint symbols you use in your work?

MathematiX allows me to superimpose inkprint symbols or offset symbols vertically before superimposing them. For instance, in astronomy we use a capital M with a subscript of a dot inside a circle to represent the mass of the sun. While this symbol is not one of the symbols supported by MathematiX, I can create it by using MathematiX flexible superimposition capability. So far, I've had no problem producing symbols.

How would you describe the benefits of technology for you?

I find that I can have a great amount of information in small disks. Before I came to the United States, I typed my 18 volume English-Japanese dictionary into three hard disks. This meant I could leave the 18 volumes at home. With a Kurzweil Reading Machine and an Optacon, I am able to transcribe materials by myself. With MathematiX I can produce technical inkprint by myself. Before, I had to ask someone to transcribe into braille all of my materials, and I had to ask someone to write down my homework or the answers to tests. Of course, sometimes I asked someone to read material so I could transcribe it on my Perkins brailler, and sometimes I used a regular typewriter to write down long equations. Now with the VersaBraille it is much faster to write things down and as much easier to correct mistakes. I always hated to wait for braille transcription. Of course, it takes a lot of time to transcribe by myself, but I prefer doing something than just waiting around. I hope that in the future every book will be available in electronic form so that we can get a braille copy immediately using computers.

## Creating Lovely Braille Reports from IBM and Apple Database Software -- Jesse Kaysen

Database (DB) programs are wonderful for keeping track of your stuff, whether it's a record collection, mailing list, checking account, or the books for your business. You use a DB's reporting function to design and produce inkprint forms, lists, mailing labels, summaries and so forth. Few databases excel at formatting to the much smaller braille page. With the general techniques described here, you can use Hot Dots or BEX to make useful braille reports from any of the hundreds of DBs available for the IBM-PC and the Apple.

Since each program has its own terminology, let me briefly define some generic vocabulary. The common analogy for DBs is a name and address list kept on index cards. Your shoebox full of cards corresponds to one database file. Within the shoebox, there are many cards; within the DB file, there are many records. Each individual would have their own record in an address DB file. Getting down to the basic data unit, every record contains the same number of fields. The address, for example, is often composed of four fields: "Street," "City," "State," and "Zip Code." Even if you don't enter any data in the field for the program to remember, the field is still there.

### Field Names Become Format Markers

When you lay out an inkprint report with a database program, you directly determine the output format of the information. But creating braille from within the DB is trickier, because the Grade II will use a different amount of space than the inkprint. The solution is to get the DB to place predictable markers in your report, and then print the report to disk. Once this data is on disk, you use Hot Dots's or BEX's file-driven search and replace to change the markers to appropriate $$commands and punctuation. Fortunately, most DBs offer an easy technique to create these "predictable markers." You ask the software to print the field names as well as the field data, like this: [[ Name: Jesse Kaysen Street: 85 Easy Street City: Madison State: WI ZIP: 99901 ]] Suppose you wanted to center the "Name" information in braille: simply change [[<CR>Name:]] to [[p$$c]]. Many programs let you identify empty or blank fields in a particular record, by printing a distinctive character. When this is possible, you can make empty fields stand out in braille. In a membership DB, for example, you could change [[<CR>Dues Paid: -]] to [[*Owes money!]].

Following the hints presented here, you'll spend an hour or two designing the DB report and the rules file/transformation chapter. Once these tools are tested, making braille reports requires six simple steps:

2. Print the report to disk.

3. Run HOT DOTS or BEX. For BEX, use option R - Read textfile to copy the information to a BEX chapter.

4. Use GLOBAL or Replace characters with the "braille format" transformations.

5. Translate the transformed data to Grade II.

6. Print the translated data to your embosser.

### Designing a Braille Shortwave Log

I'll demonstrate the technique with a database of a shortwave-radio listener. Each record in the DB contains ten fields: Key, Log Date, Log Time, Frequency, Station, Program, Rpt Send Date, QSL Rcd Date, SINPO, and Comments. Radio listeners have even more abbreviations than computer users! The important thing here is the general technique, which works equally well for any database. The "SINPO" code uses numbers to show five reception characteristics. A "QSL" acknowledges a listener's accurate reception of a particular broadcast.

When you make your report, you can shuffle the output order of the fields within each record, or even omit some fields. The DB software uses the "Key:" field to uniquely identify each record--I don't need that information on my report at all. I want the Frequency information first, so I can check if I've heard this station before as I tune in. Keeping the actual "braille report" layout in the DB program as simple as possible, my "braille format" DB report puts each field on its own line. Figure 1 shows three records from the DB as they're printed to disk. [Disk readers: see chapter F1 DATA DB.]

The final braille format must balance the goals of space conservation and ready reference. You don't want the braille log so big there's no room for your receiver, but you do want to be able to find things in a hurry (before the station fades away!). Use centering and different indents to make each piece of information distinctive enough that you don't need explicit labels, saving space. Some information naturally belongs together, so one can "collapse" separate lines into a series of items separated by semicolons. Sometimes the fact that information is missing is of particular interest, and that's where "flagged" empty fields can be handy.

### Functional Braille Format

Here's an overview of the format I came up with (the exact commands appear in Figure 3). To make it easy to find each frequency, the braille skips a line between each record. Two new-line indicators are followed with the "keep the next six lines on one page" command to prevent any one record from crossing a braille page boundary. The first, centered, line contains information from the first two fields. The frequency data is followed by the SINPO code. Since both frequency and SINPO could be five-digit numbers, I enclose the SINPO in parentheses to prevent confusion.

The second line has the logged time data, followed by a semicolon and the logged date. I can't write a reception report if I don't have that information, so when these fields are empty, I flag them with comments. The third line contains the station name, call sign, or other identifier. The fourth line again collapses two fields, showing the program heard, followed by a semicolon and any comments. The fifth and final line begins with the date the reception report was sent, followed by a semicolon and the date the station sent back the QSL acknowledgement. I use the $$vr= command to flag an empty QSL field. This command fills up the output line with full cells, so it's easy to skim along the right margin to pinpoint any outstanding QSLs. Three commands at the start of the file establish the overall format for the report: [[$$np $$ml2$$i-2]]. The [[ np]] sets braille page numbering at the end of line 1. With a left margin of 2 and a negative paragraph indent of 2, braille paragraphs start in cell 1 and run over to cell 3. While you can always edit the final file to insert these commands before you emboss, I used another common DB feature to place these initial format commands automatically: the ability to define a "set-up" string for your printer or interface card. Instead of escape codes for the printer, I used this feature to place a "predictable marker" at the start of the file, in this case, the characters [[{header}]]. When a DB lets you define a "title" that only appears on the first output page, you can use that approach to place your initial format commands.

### Designing an Effective Transformation

Both Hot Dots and BEX execute your transformation rules in first-to-last order. For success, your transformation chapter or rules file should follow this pattern:

1. Make all the "predictable marker" field names begin with the same characters. This requires two rules:

1a. Find any new print page command and change to the end-of-line code.

1b. Find any pairs of line codes and change to a single line code.

2. Change multiple spaces to one space.

3. Find the empty field marker. Change "line code, field name, space, dash" to appropriate message and/or commands.

4. Find the full field marker. Change "line code, field name, space" to appropriate commands or punctuation.

5. Repeat 3 and 4 for each field in the record.

6. Find the "set-up" characters or title and change to initial format commands.

After replacing with the rules shown in Figure 2, the file looks like Figure 3. [Disk Readers: the chapter F2 TRANFORM DB is the actual transformation I used. F3 CHANGED DB is the result of running that transformation on F1 DATA DB.] It's ready to be translated and then output to an embosser. Figure 4 approximates the result. [Emboss the F4 BRAILLE ME DB chapter for a hardcopy version.]

### Final Hints

MS-DOS and Apple textfiles have different "line end" codes. In an Apple textfile, it's just one carriage return. For MS-DOS, it's the two characters carriage return linefeed: <Control-M><Control-J>. For Hot Dots 2.0, you use the verbatim command to type this in a rule: you enter one line code with control-V control-M control-V control-J.

When you're flagging empty fields, those rules must precede the rules that handle full fields. That's because the full field marker is a subset of the empty field marker. If the full field rule is executed first, then the empty field marker no longer exists in the data.

You may need to experiment (or dig deep in the manual) to find the correct way to print your report to disk. Almost every DB offers an "ASCII export" option, but generally that won't create the right sort of file. That's because exporting saves just the data in each field to facilitate transfer to another DB or spreadsheet. Without the field name markers, you can't create the braille format. What you're seeking is the program function that prints to disk. In AppleWorks, for example, you define a custom printer whose destination is a disk file. If you have a "data box" like Braille 'n Speak or VersaBraille, you can capture printed output in it and then send the data back to the computer for processing.

Aside to BEX Users: Replace characters operates one BEX page at a time. In the unlikely but possible case where a field name gets divided between BEX pages, Replace won't see the complete marker. After Replace is finished, edit the transformed chapter and check the beginning and end of each BEX page.

## Automating VB File Transfer and Hot Dots Braille Translation with Batch Files -- Robert Sweetman

I use RDC's Hot Dots program running on a hard disk both to translate ASCII to grade 2 and to translate grade 2 to ASCII. I generally compose letters and lengthy documents in grade 2 on the VersaBraille II+, and then transmit to the IBM using PROCOMM set at 9600 baud with X-on/X-off handshaking. I can use Hot Dots to back-translate the received file and then print it. When I want to do careful writing and review, I'll translate the MS-DOS print file into grade 2 and send it to the VBII for writing.

To save time, I use batch files with generic file names to run through the transfer and Hot Dots process. Once I supply a filename to work with, the following batch file removes control characters, inserts paragraph markers, translates to grade 2, sets up COM port 2 for the VersaBraille and then copies the braille file to the VersaBraille.

For this batch file to work, the program [[mode.com]] or [[mode.exe]] (depending on DOS version) are in the referenced subdirectory or are otherwise available via the path command.

[[rem brl.bat echo off cls copy %1 c:\dot\data.prt cd dot del final.brl mode com2:9600,n,8,1,p global data.prt data.br1 printbrl.rul del data.prt bt data.br1 data.br2 del data.br1 global data.br2 final.brl rmvcomma.rul del data.br2 copy final.brl com2 cd\ echo braille translation and transfer completed.]]

Using this batch file is simple. I set the VBII to receive from com2, then type [[brl filename <Enter>]] at the c> prompt. Suppose I have a bunch of sections from the Internal Revenue Code in ASCII form in a file named "code" on drive A. Then at the c> prompt I type [[brl a:code <Enter>]]. The necessary global replace and translation functions are carried out on the C drive and the original file is left undisturbed. The process is fast and requires no user intervention.

Here's how the batch file works. The [[%1]] refers to "filename", or [[code]] in the sample case. The directory where I keep Hot Dots program files and data files is [[c:\dot]] (the root directory "dot" on the C: drive). The command [[copy %1 c:\dot\data.prt]] copies the "filename" I've supplied to the file [[data.prt]] in my [[dot]] directory. The original file is left untouched, and the batch process uses [[data.prt]] and its children from now on.

The "global" is the Hot Dots global replace program and "bt" is the Hot Dots grade 2 braille translator. The files with [[.rul]] extensions contain the changes to be made by the global replace program. The first, [[printbrl.rul]], is my customized file that removes control characters and extra spaces and inserts [[$p]] paragraph markers. The second, [[rmvcomma.rul]], removes double caps signs, extra paragraph markers and any extra spaces from the translated version. I use similar files to format work for braille hardcopy from a commercial embosser or from Dipner Dots, to translate grade 2 to ASCII for printing in inkprint, or any other custom applications. The batch file approach works with Hot Dots and the VB classic or VBII series, hardcopy embossers and printers, etc. To modify the above batch file to get hardcopy braille, you must include a step to perform braille formatting, using the Hot Dots program [[ftext]]. For example, to get hardcopy braille through com2, you'd replace the single [[copy final.brl com2]] line with these three lines: [[ftext final.brl final.bfm /w=40 /f=25 del final.brl copy final.bfm com2 ]] When I use the [[brl]] batch file, I leave the VBII+ settings the same both for copying to the VB and for uploading to the PROCOMM modem program. They are as follows: [[serial: b: 9600 d: 8 s: 1 x: h p: none i: dc in yes o: dc out yes t: dtr hs r: dsr y c: com mode parameters: t: no c: yes a: yes v: off d: demo no r: no ]] I have had no trouble uploading from or copying to the VB with this configuration; in fact, this file was written on the VB and translated with a batch file that reverse translated, formatted to a carriage width of 72, and then inserted one blank space before each carriage return. This meets the requirements of the "fast.doc" message preparation method used to automate uploading to and downloading from the CompuServe system. When uploading, text is sent from the VBII unformatted; formatting is then done with Hot Dots when the batch file is run. Please let me know what you think of this approach--you can reach me at 70255,111 on Compuserve. ## Putting Contextual Replace to Work: Non-printing Comments in BEX Chapters -- Jesse Kaysen When you're sharing information on disk, it's convenient to include non-printing comments. Here's how we do it here at RDC, where blind and sighted people work together on a variety of writing tasks. We use a predictable pattern for our comments, which makes them easy to find, easy to read, and most importantly, easy to globally delete. The syntax for our comments is this: begin with a left brace, write the comments as sentences, and end with a space followed by right brace. Suppose I see a sentence like this: [[Her participation, however, does not imply their approval.]] I add my comment in right at the point of the error, like so: [[Her participation, however, does not imply their {Verb doesn't match subject! } approval.]] When reading the edited version, you can leap from comment to comment by locating for the left brace. The control-T read a sentence command lets you hear what's been written. The following short contextual transformation chapter deletes all comments in one swoop: [DISK READERS--also provided in the STRIPBRACE chapter on this disk.] [[....{}}}}}}}}}}.Xoooooooooo..{}}}}}.Xooooo..{}.Xo..{}.xx.....]] Since all the change to strings in this chapter are empty, it deletes the comments. Create a chapter like this to use with Replace characters, and you're ready to creating bracing comments! ### How It Works For the curious, here's a ten-cent guide to the tools I'm using--details are in the BEX Master Level. With basic replace, each transformation rule has two parts: you look for the find string and replace it with the change to string. The terminator character (in this case, the period) defines the end of each string. Contextual replace adds another element in the middle: the pattern string. Each character (or pattern code) in the pattern string corresponds to one character in the find string. The combination of the find string and the pattern string define a "match." With basic replace, your find strings must be typed exactly: lowercase a is different than uppercase A. With contextual replace's pattern codes, you can define the find string more generally. The case of the pattern code is very important. Lowercase pattern codes are "departers." When the combination of find string and lowercase pattern code is satisfied, the find string character is removed and replaced by the change to string in the target chapter. Uppercase pattern codes are "boundaries." They describe what characters must surround the ones you're actually changing. In the sample, we use three pattern codes: Uppercase X is "boundary exact match." When paired with left brace, it means: look for exactly the left brace character, but don't change it to anything. Lowercase o is "departing wild card minus one." When paired with right brace, it means: look for any character except the right brace. When you find it, replace it with the change to string. Lowercase x is a "departing exact match." When paired with left and right brace in the final rule, it means: find exactly the two characters left brace right brace and replace with the change to string. One could delete all comments in braces with just the last two rules, but it would be a lot slower. To see why, let's take a detailed look at how BEX executes these four rules on the little sample above. The first rule matches the left brace followed by any ten characters that aren't right brace, and deletes those ten characters. The first match is with [[{Verb does]]. BEX only changes those find string characters that are paired to departers: after the first replacement, the data looks like: [[{n't match subject! }]]. BEX starts again at the left brace and repeats the first rule, resulting in [[{subject! }]]. At this point, the first rule no longer matches the data, because the right brace is ten characters away from the left brace. On to the second rule, with the result of [[{ct! }]]. Now the second rule can't be satisfied, so it's time to run through the third rule, eating away at the data inside braces one character at a time. After four bites, all that's left is [[{}]] which is deleted by the final rule. ## Teaching BEX with 3.5-inch Disk Drives -- David Holladay The recommended system for a BEX beginner is two 5.25 inch disk drives. When you're asked to teach BEX on an Apple IIgs, you may often be confronted with one 5.25 inch drive and one 3.5 inch drive. With Master Level BEX 3.0, you can use the 3.5-inch disk drive for BEX data. But the Master Level of BEX is not a comfortable place to start off with beginners. I am assuming that you, the trainer, can cope with the Master Level of BEX. There is a trick you can use to make the best of this situation, which we employ when we demonstrate BEX at conventions. We configure at the Master Level, but display Learner Level prompts. This seems to be an effective way to show the power of BEX without confusing people with short prompts. It is time that we shared this most useful of tricks. ### Copying Learner Level Messages to Master Level Page Step one is to make several backup copies of the [[MESSAGES]] chapter on the Boot side of your BEX 3.0 disk. As detailed on pps M9:1-6, this chapter contains three pages, with the bulk of the prompts for the Learner, User, and Master Levels. Once you have backups, use the Clipboard to copy all the characters from BEX page 1 to BEX page 3. You must reboot BEX with a Master Level configuration to see the result: Learner Level prompts at the Master Level! ### Configure with Two Drives Configure the modified BEX disk for your student(s) at the Master Level. Choose the appropriate voice, inkprint printer, and braille embosser depending on your student and your equipment. After configuring any printers, you get the question "Do you have an extended disk system?". Answer "yes". Set up virtual drive one to be slot 6, drive 1. When you're working with one 3.5-inch disk drive, virtual drive 2 is slot 5, drive 1. Answer 0 for the slot of virtual drive three to cut short this line of questioning. When this configuration is loaded, the prompts are at the Learner Level. Drive 1, your program drive, is the 5.25 inch drive. Drive 2, your data drive, is one-half of the 3.5 inch drive. The system will look and feel like a two drive Learner Level system. ### Cautions on Initializing You should do one more step to finish the camouflage. Before class begins, initialize a stack of 3.5-inch disks. The prompts for initializing 3.5 inch disks are not stored in the MESSAGES chapter. While these prompts are appropriate for someone at the Master Level, they can be scary or confusing to someone just beginning. During the class, you can only practice initializing 5.25 inch disks the first time you get to the Starting Menu. When you move from the Starting to the Main Menu, BEX loads the AmDOS software for the 3.5-inch disks, and you can't initialize 5.25-inch disks until the next time you boot. Of course, your students will actually be running the program at the Master Level. This means that a student can accidentally call up an option that is "locked out" of the Learner Level; if someone presses H at the Main Menu, for example, they'll be able to use Heading test. You may need to closely monitor some students to make sure they don't go exploring into areas of BEX you are not covering in your sessions. When the training session is over, you restore the Master Level prompts to normal by copying the backup [[MESSAGES]] chapter on top of the modified chapter on the BEX disk. ## Bulletin Board We're happy to publish your 50 to 100 word announcements here, free of charge--please send them in print, disk, or braille formats to us directly at RDC. ### Braille 'n Speak for Sale Such a bargain: a one-year-old Braille 'n Speak pocket computer, with print, braille, and cassette manuals, still under warranty--for just$500 or best offer. Call Richard Hutcheson at 315-267-2587 evenings or 315-265-5757 days.

Improving accessibility to books in braille, large print, and recorded formats for blind, visually impaired, and other print-handicapped readers is the goal of a newly established group of service organizations, the Coalition for Information Access for Print-Handicapped Readers (CIAPHR). The critical need for an information bank of available titles in alternate formats was confirmed by surveys conducted in 1984 and 1985 by a task force spearheaded by the American Foundation for the Blind. (AFB). Survey results indicated that nearly 50 percent of the responding producers and disseminators do not list their available titles. (Complete survey results appear in the Huebner, Kelly, and Davis article in the June 1989 issue of AFB's Journal of Visual Impairment and Blindness.)

In response to the task force's findings, CIAPHR was organized in March 1988 to explore ways for accessing a comprehensive North American listing of educational, vocational, and recreational materials in special media. CIAPHR has met several times since then to develop plans for a follow-up survey to update information about producers of materials in alternate formats and make recommendations for developing a North American resource bank of information.

"Knowledge is power and so too is access to information," said Dr. Huebner, chairperson of CIAPHR. "The Coalition strives to enable more blind and visually impaired people to have easy access to both knowledge and information, thereby ensuring their full participation in all sectors of society." Dr. Huebner added that the comprehensive listing will minimize duplication of effort and provide materials more efficiently, effectively, and economically.

CIAPHR includes representatives from AFB, Recording for the Blind, the American Council of the Bind, America Printing House for the Blind, Association of Instructional Resource Centers for the Visually Handicapped, Canadian National Institute for the Blind, Charles Crane Memorial Library/University of British Columbia, National Braille Association, Inc., and the National Library of Canada. For more information on CIAPHR, please contact:

Kathleen Huebner

CIAPHR Chairperson

American Foundation for the Blind

15 West 16th Street

New York NY 10011

Telephone: 212-620-2045

### Handy Apple IIgs Graphic Conversion Utility Increases pixCELLS Options -- Jesse Kaysen

Roger Wagner Publishing, Inc. has produced yet another winning program. Their "Graphic Exchange" utility lets you pass virtually any Apple II graphics image between various applications. You can now take a Newsroom, Print Shop, Dazzle Draw, or Paintworks Deluxe file and convert it to the format that pixCELLS can recognize. In addition to converting among fourteen separate Apple II graphic files, the Graphic Exchange can directly read "paint" files from Macintosh disks.

While the "Graphic Exchange" is modestly priced (list $50, street$35), it does require some pretty heavy hardware: an Apple IIgs with at least 512K, an Apple mouse, and a 3.5-inch disk drive. If you want to transfer MacPaint graphics, you need at least 768K memory. The program won't work with speech. I've tried it and it works like a charm, although it took me a while to get comfortable with its interface. Carefully reading the manual solved all the problems. If it's not available at your favorite mail-order emporium, you can contact the publisher directly:

Roger Wagner Publishing, Inc.

1050 Pioneer Way, Suite P

El Cajon CA 92020

Telephone: 619-442-0522

### Accessible Cookbooks as Fundraiser for Reading Service

The University of Kansas Audio-Reader Network is a statewide system of radio reading services. It's selling a cookbook with more than 230 recipes from its staff, volunteers, and listeners. The recipes range from the wacky--dog biscuits, glass cleaner, and Dirt Cake--to the everyday--Main Dishes, Breads, and Desserts. The cookbook is available in four media: two soft-cover braille volumes at $12; professionally recorded, voice- and tone-indexed cassette with case at$10; regular print at $5.25; and Apple or IBM textfile on disk at$5.25. Checks payable to "Audio-Reader Network" must accompany orders, which should be sent to: PO Box 847, Lawrence, KS 66044.

## Facts on File

Jesse Kaysen doesn't have quite as many shoes as Imelda Marcos, but she's working on it.

Nevin Olson grew up on a farm in southern Wisconsin.

In his youth, David Holladay has been detained for questioning by the CIA, PLO, Israeli Security Forces, MIT Campus Patrol, and UW Protection and Security.

Robert Carter is a counseling psychologist with Texas A&M University.

Robert Sweetman's first contact with RDC was on March 20, 1982. We've both learned a lot since then!

Index Classic & Index Advanced Braille Embossers

HumanWare, Inc.

6140 Horseshoe Bar Road, Suite P

Loomis CA 95650

Telephone: 916-652-7253

Personal Data Systems, Inc.

100 W. Rincon #207

Campbell CA 95008

408-866-1126

Arkenstone, Inc.

2500-B Augustine Dr.

Santa Clara CA 95054

Telephone: 408-727-3434

TACTIC: Braille quarterly on adaptive technology. Back issues $2.50 each (when available). Annual subscription is just$10 from

Clovernook Printing House for the Blind

7000 Hamilton Avenue

Cincinnati OH 45231

APH PocketBraille

American Printing House for the Blind

1839 Frankfort Avenue

Louisville, KY 40206-0085

Telephone: 502-895-2405

Braille 'n Speak

Blazie Engineering

Street, MD 21154

Telephone: 301-879-4944

### The RDC Full Cell

Phyllis Herrington, Tech Support/Newsletter; David Holladay, President; Caryn Navy, Vice-President; Susan Murray; Order Processor; Jesse Kaysen & Nevin Olson, Alumni-to-be.

### Production Notes

Written & edited with BEX on an Apple IIgs. BEX commands changed to Microsoft's RTF/Interchange format control words with BEX's Contextual Replace. File transfer with BEX & Hayes's Smartcom II to an Apple Macintosh SE. RTF commands interpreted and then spell checked with Microsoft Word 4.0. Pages composed with Aldus's PageMaker 3.02, output on an Apple LaserWriter Plus, and printed at The Print Shop. Two-track audio edition mastered on APH Recorder & copied on high-speed Recordex 3-to-1 duplicators.