This was written in late 1986, it is in RDC Newsletters 45, 46, 47, and 48
While going over some old papers, I stumbled on an interesting document written over 6 years ago. It was the first time I set down anything in writing about my plans for writing software to work with sensory aids equipment. The remarkable thing is that virtually all the issues and ideas critical to Raised Dot were barely mentioned. I thought that it might be fun to excerpt portions of this document for the Newsletter. However, it is important to give some more background about the history of Raised Dot Computing.
In the summer of 1980, I saw a flyer describing a VersaBraille paperless brailler. I was very familiar with personal computers: I had been involved with two firms (ECD Corp. of Cambridge and Mohr Labs of Madison) that tried to make early personal computers. (They both went bankrupt, but not because I hung out there.) Both machines were based on the 6502 chip, the processor at the heart of the Apple computer. I had a friend from MIT who had recommended that I look into the Apple computer. He had a vested interest: Alan Baum, together with Steve Wozniak, designed the Apple 2.
The possibilities of combining the VersaBraille and Apple computer intrigued me. When Caryn and I were MIT students, we'd dreamed about designing a device we called "the Homework Machine". It was critical that the device be able to process mathematical braille since most of our homework was mathematical equations. The Homework Machine was to have an electronic braille keyboard to accept grade two braille and Nemeth code braille. The computer-based device would process the braille into printed form that could be handed in as homework. In 1972 we figured that such a device would cost around $75,000.
You can understand my excitement in 1980 when I first read about the VersaBraille. What was so amazing to me was that this braille entry device was portable, so notes and material could be entered where most convenient. I figured that a VersaBraille, a personal computer, a dot matrix printer and assorted peripherals would only cost $10,000, which was a lot cheaper than $75,000.
Caryn and I bought an Apple system in August of 1980. We ordered a VersaBraille unit (the State of Wisconsin paid for 60% of the VersaBraille). I decided that one way I could defray all these hardware costs would be to develop Apple software specifically to work with the VersaBraille. I didn't consider speech output at the time, because this was before the Type-'N-Talk or the Echo synthesizer were available.
I realized early on that I needed cooperation from Apple Computer, Inc. and Telesensory Systems, Inc., and I was fortunate to get it. Apple provided me with disk copies of their manuals to run into VersaBraille tape. TSI provided me with very early technical specifications--I was the first person outside TSI to see the VersaBraille I/O Manual. I was very eager to get started on the project, so I tried to write software to make another computer act like a VersaBraille. (We couldn't afford another Apple, so I cobbled together a Mohr labs computer, which we dubbed the "Orange.") Unfortunately, the VersaBraille was so much more sophisticated than the Orange that the simulator didn't help at all. I decided not to apply for any grant money. It seemed silly since we could buy equipment using our savings. My aversion to grant money was critical to the later success of Raised Dot Computing.
Here are the applications that I listed in 1980 for an Apple/VersaBraille system:
A pretty ambitious list, eh? In my proposal to TSI, I mentioned still more elaborate projects: video disk direct access, translation of braille music into printed symbols, map display digitizers and more. All this before I had worked with a VersaBraille and barely before I had unpacked the Apple computer! Since I was working full time as a computer programmer for the University of Wisconsin, it is hard to see how I thought I could do more than a small fraction of those projects.
It is interesting to note that the project that I really wanted to work on--a translator from Nemeth code to print math symbols--was barely mentioned. I suppose I did not want TSI to think that I was so interested in a project with such a small commercial potential. Although I'd set some high goals, I felt pretty confident. I was a fairly skilled programmer and Caryn had a good sense of what was useful. Besides, my housemates at the time (Jesse Kaysen and Nevin Olson) believed that the venture just might fly!
In February, 1981, Caryn received her VersaBraille. (Thus began several years of wrangling over who got to use it when.) After a few months, I had written a crude text editing program for the Apple. The program could move text from Apple to VersaBraille and VersaBraille to Apple. The program could print out material on a printer as well. The program did not contain any braille translation, could not work with any voice output devices, had no configuration program, and was virtually unusable by a blind person. I called the program "BRAILLE-EDIT."
In the spring of 1981, Caryn received her Ph.D. in Mathematics from the University of Wisconsin. She had spread 70 copies of her resume across the land. She got two offers for a teaching position: one from Auburn University in Alabama, and one from Bucknell University in Lewisburg, Pennsylvania. Auburn sent a free tactile Rubik's Cube; Bucknell offered cheap housing. We were going to Lewisburg.
Moving to Lewisburg meant breaking up our household: for over a year we had been renting a house in Madison with Jesse Kaysen and Nevin Olson. On moving day I made a prediction in jest: the work I was doing in my spare time with small computers would expand into a business that would eventually employ all four of us. We laughed.
Actually, my plan was to get a regular computer job. It turned out that Lewisburg was quite small with few opportunities for an over-excited computer programmer. I blew the one good job interview I had by not wearing a tie and then explaining in detail to my prospective supervisor why his plans would fail. I spent a few weeks making houses of cards and watching the Munsters on TV before committing myself to home-based computer operations full time.
In the fall of 1981, I taught the Apple to mimic another VersaBraille in what the real VersaBraille thought was a VersaBraille to VersaBraille data transfer. One month I wrote a grade two translator, the next month I wrote a back translator. I put in some crude speech output capabilities into my software. By late fall, "BRAILLE-EDIT" almost looked like a practical program. It was strictly designed to augment a VersaBraille--provide translation and reformatting capabilities. The first few customers found out about me through TSI.
Ted Glaser bought the first copy of BRAILLE-EDIT on 12/20/81. To switch from a voice output version to a screen output version, you had to change a few lines in some BASIC programs. Ted suggested a "configuration file" set by the user to simplify customizing the disk. I rejected his suggestion as too complicated. After he made the suggestion two more times, I began to work on a configuration program.
Customer number four made Raised Dot possible. Harvey Lauer saw some potential in the software that I was selling. He sent me a list of 12 suggestions. No sooner had I revised the software then he sent another list of 12 suggestions. I got suspicious and asked him about it. Harvey said that he feared that if he presented the entire list of 24 suggestions that I would get discouraged and stop working on software for the blind. He molded BRAILLE-EDIT so that it was actually a useful tool.
Many people think that since my wife is blind, that she was giving me detailed feedback. This was not really true--she was too busy! Caryn was devoting her energies to teaching calculus (and bringing home enough money to support my programming pipe-dream). Caryn did provide me with an essential role model: she taught me how blind professionals live and cope with the sighted world. She also pointed out that a back translation program was needed to properly work with a device like the VersaBraille.
In February of 1982, I started work on a translator from Nemeth code into printed mathematics. Caryn needed a way to quickly generate handouts, tests, quiz solutions, etc. In early 1982, BRAILLE-EDIT was such a low volume seller that I had plenty of time to devote to this exciting project. After a month's work, we got some crude results. After a few more months, the program became a reality. Caryn could make the Apple automatically transfer a file from the VersaBraille, translate the various braille codes, and print out the result (with three levels of fractions, 16 levels of radicals, Greek letters, integrals, etc.) on a dot matrix printer.
Caryn's first major design effort was working out the system of double-dollar-sign formatting commands. I led Caryn to believe that these formatting commands would also be worked into the mathematics software. When it came time for a unique characters to distinguish the format commands from surrounding braille text, she picked "$$" since this did not conflict with anything in Nemeth code. At the time, neither one of us thought about the word "needed" ( _- ne$$ _l ) in grade two braille. But I lied: I never did integrate the BRAILLE-EDIT system of "$$" formatting indicators into the mathematics software.
So many of the early customers gave so many useful ideas. I was quite free in distributing updates, because I didn't like the idea that someone was not using the very latest version. According to the records which have survived, Michael May, customer number 12, bought BRAILLE-EDIT on 5/28/82. I sent updates on 7/30, 8/24, 1/27/83, 2/11, 3/3, 4/21, 5/30, and 7/18. The following early customers (customer numbers in parentheses), provided very useful suggestions, and helped to make BRAILLE-EDIT a real product:
After 18 months of operation, Raised Dot Computing had 37 customers. Revenues were about $11,000, and expenses were $14,000. The economics did not bother me a bit. I never expected that the sales of software would support me. As long as Caryn provided a monthly subsidy to the enterprise, there were no problems.
January of 1983 was a month of many changes: Apple was switching from the Apple two plus to the Apple 2e; TSI was switching from the VersaBraille P2B to the P2C; Street Electronics was switching to a new speech chip that required a new TEXTALKER; and I was trying to switch from exclusively supporting the CCS 7710 card to also supporting the Super Serial Card. Keeping in touch with all my customers about these issues by phone was overwhelming.
Besides, it seemed that I was on the phone 20 hours a day. On several occasions I moved the telephone into the bathroom and conducted business from the bathtub. I'm sure that the folks on the other end of the wire were a little surprised to hear the sloshing when I got up to get names and addresses.
In an attempt to keep customers up-to-date with all the changes, the first issue of the Raised Dot Computing Newsletter was published in February, 1983. I just wrote a BRAILLE-EDIT chapter and used "Multi-copy print" to produce 40 copies on my dot matrix printer. [If anyone has one of these originals, I would appreciate a copy for the archives.] For the March issue, I got a tape duplicator.
The Newsletter was useful in cutting down on phone time, but it also complicated my life. Very soon, I realized that I was spending about half my time doing the Newsletter. The other half of my time was spent opening mail, writing letters, and answering the phone. A third half of my time was spent trying to write programs. I was quickly running out of halves!
In July of 1983, I went to the American Council of the Blind annual convention in Phoenix, Arizona. Between preparation, going to Arizona by way of Baltimore, and recovery time, it took a whole month of effort (and over $2,000). But it was worth it. More than anything else, the ACB convention involved me in the sensory aids swirl. I made many, many contacts, and sold quite a few Newsletter subscriptions.
The ACB convention gave me a chance to show off my latest item: a guide to interfacing devices to the Apple computer. By bundling detailed interfacing notes with each sale of BRAILLE-EDIT, I was able to cut down on some of the lengthy technical conversations.
I hired Cindy Peltier in August. She had the perfect background for my first employee: having worked in insurance, libraries, and retail, she had no experience with computers. We quickly became a very effective team. She did all the invoicing, shipping, newsletter subs, audio recording, record keeping. My grandfather had told me that the hardest transition for a business is when a second person is involved. There were so many things that I just kept in my head. Raised Dot's continued success depended on Cindy's ability to deftly extract this information from me.
One night the phone rang at 4 o'clock in the morning. Steveland Morris (also known by the stage name "Stevie Wonder") was having some problems with one of his VersaBrailles. I told him that one way to regenerate a worn out tape was to transfer the entire contents to an Apple computer and then transfer back to a fresh tape. His organization eventually purchased about $1,000 of software and hardware for RDC. They instructed me to put everything in a box and ship it out on the next flight--I had to drive 70 mph to get to the nearest airport. The whole experience was a fascinating glimpse at life in the fast lane.
Around the time that I hired Cindy, RDC started to sell the Cranmer Brailler as a dealer for MCS. At a time when Raised Dot was not making enough money to pay me a salary, signing a contract saying that Raised Dot would be buying $150,000 worth of equipment for resale was quite a gamble. I learned many hard lessons about business. I paid the invoices from MCS promptly in 30 days. Often a school or agency would wait three or four months to pay me. My profit margin was slim: one invoice past due complete consumed the cash flow cushion for seven units. The supply was unreliable: only rarely was MCS able to ship units promptly. It was very frustrating to deal with customers wondering when they were going to get their unit.
Despite all these problems, I realized how valuable it was to sell embossers. I was able to respond to folks inquiring about both hardware and software. By offering more elements of a "complete package," my product line was more attractive to schools and agencies. Such groups were often unsophisticated about computers, and didn't know how to put the pieces together themselves.
The Cranmer Brailler has its faults, and we were careful not to misrepresent it. In fact, our very low key sales of the Cranmer was a source of irritation to MCS. In the spring of 1984, Dan Gorney and Dean Blazie of MCS drove up in their three-piece suits to ask why we were not heavily promoting the Cranmer. I explained how the low margin meant my cash flow limited my ability to sell more units. They sweetened the deal by increasing the margin.
Starting in the fall of 1983, I began working on a very new version of BRAILLE-EDIT. I wanted to centralize all the best input/output routines in a large assembly language program which would always be present in memory. [These resident input/output programs were later expanded to become the core of BEX].
I finally finished BRAILLE-EDIT version 2.45 in the spring of 1984. Despite a number of virtually crippling bugs in the first few versions, BRAILLE-EDIT 2.45 was a significant improvement. It was possible to scan for chapters on most options. Large print on the screen was available in the Editor. This was the first two disk version of BRAILLE-EDIT.
One weakness of Raised Dot had been the lack of braille documentation. With my usual optimism, I decided that what we really needed was a Thiel embosser--at that time, it cost $15,000. It never occurred to me to get a loan for this capital equipment: in October 1983, I began a series of nagging articles in the Newsletter. "Buy more software so I can get a Thiel!" I harangued my readership. Much to my delight, it worked: in late December I was able to send MCS a check for $10,000 (the remaining $5000 was on 30-day account.) Looking back, I'm still proud of that stunt.
Thanks to the customer financing of the Thiel, the new version of BRAILLE-EDIT came with braille reference cards. I believe the popularity of version 2.45 of BRAILLE-EDIT had more to do with the availability of a braille reference card than with the features of the software. Having a Thiel made it possible to produce several manuals and other large documents in braille. However, BRAILLE-EDIT Version 2.45 did a very poor job of braille page format (braille page numbers at the end of a text line, print page indicators, running heads, etc. were not available). RDC's interest in improving this situation lead to BETTE and eventually to TranscriBEX.
I believe this is one of the strengths of Raised Dot. I was actually using our software for production tasks. That's why it was constantly changing. The main user was also a programmer: when I needed a particular feature to print or braille my documentation, I simply modified the program. It's also why RDC software is particularly good for producing print and braille versions of the same document. It's often seemed to me that many other sensory aids firms don't rely on their products as I did. For example, all of MCS's Thiels were travelling around the U.S., being demonstrated by sales personnel. Few MCS employees were able to gain expertise in using the equipment for production work. Raised Dot Computing ended up fielding a lot of technical phone calls on the Thiel and several other braillers.
The last year in Lewisburg was increasing hectic. RDC was completely taking over our rented house. I filled 20 boxes with my books so I could use the living room bookshelves for RDC products. Everytime we got a shipment of braille paper, the pile of boxes replaced our kitchen table. The dining room was dominated by the phone and the customer files. In addition, it was not easy running a business out of a small town. A modest mailing could clean out both stationary stores of a certain size of envelope. A broken print wheel meant an hour's drive to buy a replacement. Making more room in the house for more personnel would mean lessening the quality of our life.
Every week our old friends Jesse Kaysen and Nevin Olson would call and extol the virtues of Madison. I caved in.
Caryn decided (or maybe I persuaded her) to leave teaching and join Raised Dot full time. Jesse and Nevin were willing to help find a real office for RDC outside of someone's home. Jesse offered to work for RDC "for a few months" until we got settled. Unfortunately, Cindy could not relocate in Madison, since her partner was a tenured professor. I was going to be the only RDC employee to bridge the transition.
And what a transition it was! In addition to the logistics of packing up thousands of boxes, there was a minor cash flow crisis. In early June, I calculated I needed $37,000 to finance the move. The checking account had $5,000. I panicked. I didn't order any Cranmers for June to minimize my accounts payable. I attempted to use up every sticker and staple of office supplies, with plans to re-order once we landed in the Midwest. Most importantly, I began a marathon of phone calls to any customer who was out more than 30 days. I'm sure my daily phone calls made some enemies, but we did manage to scrape together enough cash to finance the move.
The RDC office in Lewisburg closed on July 18th, 1984. On July 24th, Caryn, Jesse, and I found ourselves in a spacious suite of grubby offices, surrounded by dozens of boxes. We had arrived.
It was July of 1984; we had arrived in Madison, Wisconsin. The initial staff was myself, Caryn Navy, (whose husband I happen to be), and our friend, Jesse Kaysen. After unloading the U-Haul, the possessions of Raised Dot Computing comfortably filled half a room. There was an Apple 2 plus and a 2e, a well-worn letter quality printer, and a Thiel Braille embosser. An old kitchen table in an otherwise empty room served as our conference center. Oh, but we had great plans! Raised Dot Computing was incorporated under Wisconsin law; we retained an attorney and an accountant.
The building we purchased had been used as a union hall for decades. It seemed that not a single bingo card, "dues-paid" button, annual picnic photo, nor grievance had been thrown out: our first job was to cart all this stuff out to the curb. A few weeks later, a frantic union leader called to ask us if we had found his gun somewhere in the building--we hadn't.
Between then and now, Raised Dot Computing has grown from "David Holladay's pet project" to a real business. The building on Baldwin Street had 5 offices upstairs, a completely open ground floor, and a wet bar (complete with euchre tables) in the basement. We started out using only a small part of the upstairs. Now we're beginning to feel the building's limits: we recently carved out a 6th office downstairs.
By 1984, BRAILLE-EDIT had developed a good reputation. It was known as a good braille translator that was relatively easy to use. Many people thought it was a good way to introduce blind people to computer use. From 1981 to the move, I had sold a total of 450 BRAILLE-EDITs. In the next 12 months, we sold 450 more. The increasing volume of business demanded a more formal business structure.
All along, I had cadged business advice from my friend, Nevin Olson, a small-business-development consultant and industrial engineer. Jesse had always wanted to work in the same business as her husband, so Nevin bowed to the inevitable and joined the RDC team in December.
It was a wonderful team, but none of us were particularly good at typing invoices. We placed an ad for a "mail order fulfillment coordinator." Kristi Seifert, who'd spent many years as a systems analyst at a large hospital, wanted to meet the strange people who'd invented that job title. Whiskey (Caryn's guide dog) played an important part in the first round of job interviews, wandering in to check out each applicant. All four of us knew that Kristi was the one when Whiskey ambled in and curled up at her side. So Kristi became the first employee not related to another employee.
Up to this point, RDC's administrative functions had clearly been designed by a programmer. Cindy or I had coped as well as we could, but neither of us really had the time to develop a coherent system. Both Nevin and Kristi brought valuable experience gained working in large corporations: they stabilized the volatile RDC without killing its spirit. For the first time, every transaction--even Newsletter subscriptions--was invoiced. Kristi, Nevin, and Jesse attempted to organize the chaotic customer files. The "S"s were particularly unruly--there are almost 50 "State Commissions for the Blind." During my father's visit that fall, he helped out with many mundane tasks. His suggestion that we file material by ZIP code marked the turning point in our customer records.
When convention time loomed in June of 1985, we realized that we needed more help. Becky Rundall heard through a mutual friend that we had an opening, but no details about what the job involved. Fortunately for us, she braved our five-on-one interview. She impressed us with her positive attitude, her unwillingness to accept "imperious" supervision, and her ability to cope with stressful situations. One thing she could not cope with was our phone system, so we got a new phone system with real hold buttons.
Kristi and Becky insisted (and still insist) on keeping good records. These records allow us to understand and reconstruct each customer transaction. Since our customers are widely distributed over the country, ZIP codes are an ideal identifier. We have retrieved files fast enough that callers have praised the speed of our "computer system." The hardware for the system? Three four-drawer filing cabinets.
One way RDC copes with the large number of transactions is to handle them quickly and correctly the first time. (In the first seven days of 1987, for example, we processed 100 invoices.) Becky ships most orders within 24 hours of receipt. Even in unusual backlog situations (updates are the prime offender), orders never take more than a week to process.
Caryn started work at Raised Dot answering tech calls so I could spend more time programming. When she started, she had a $15 garage-sale desk, her own VersaBraille, and no computer. Caryn always insisted on completely tracing the bug or "feature" that caused problems. In the fall of 1984, she got many calls about braille page numbering problems. Thanks to her perseverance, she isolated a nasty bug in the Cranmer brailler. To get around this problem, we issued BRAILLE-EDIT version 2.50.
As Caryn gained more experience and became more involved in program design and coding, she needed freedom from constant phone duty. After a national search, we hired Phyllis Herrington in the Winter of 1985. Her teaching background helped our technical support become even more accessible. She also instituted the tech support log, which helps us develop a better understanding of what causes problems. Phyllis is also becoming involved in product design and evaluation, particularly for the IBM-PC. Phyllis, Caryn, and I now rotate answering the technical line. We've recently begun weekly technical meetings to evaluate new devices and to work out better procedures for dealing with technical questions.
We're always happy to answer questions, but we try to anticipate the most common ones in our manuals. I wrote the first RDC manuals--the BRAILLE-EDIT version 2.45 manual was written in one weekend. In the fall of 1984, Jesse revised the BRAILLE-EDIT manual. Jesse and Caryn also rewrote my early draft of the BETTE manual. In 1985, Jesse wrote from scratch the BRAILLE-EDIT User's Guide and most of the BEX Dox (Caryn and I wrote the Master Level and the Interface Guide).
While Jesse has great talents as a technical writer, it became clear that she couldn't write manuals and edit the Newsletter and produce RDC promotional materials and do all the other things that Jesse does at RDC. In 1986, we hired Andrea Botts as a full time technical writer. Her first project was the TranscriBEX manual, which is far and away the best manual we've ever produced. (It's getting rave reviews from transcribers around the country.) She insisted on getting each feature working for herself before she would write it up. She is currently working on software for the IBM-PC and other product testing.
Raised Dot Computing has been entirely self-capitalized. An Apple 2 plus computer and a VersaBraille bought by Caryn and I became the basis for Raised Dot Computing. The sale of software and hardware has financed the purchase of our equipment and our salaries. This funding method has meant that we must provide our own direction: no outside funding agency supplies us with a business plan. So how do we make decisions about the future?
Most days the RDC staff has lunch together, where we discuss the general approaches to the sticky situations we've encountered that day. The details are thrashed out later by those who are directly affected. For example, when someone wants a quote on a large quantity of software, it's an issue for Nevin (business manager), Kristi (office manager), and Becky (production manager). No one else needs to be involved. After all, it is Nevin who makes the deals, Kristi who types up the paperwork, and Becky who does the shipping. We have weekly staff meetings to develop long term policies. There are many good reasons for having the entire staff figure out its own rules. It's easier to remember and work with policies that you have helped to establish.
I'm proud of how well our system works, and I think that an understanding of it can assist those who have contact with RDC. When you call or write someone at RDC, she or he speaks for the company. The question, "Gee, does David approve of what you're saying?" is not really appropriate. Whether the person you're speaking with approves of what I'm doing is more to the point. As RDC has grown as a company, I've realized that software itself is a tiny part of a software enterprise. The technical changes I may make in software greatly affects documentation, product testing, production, customer service, and scheduling.
If you called me to try to develop a package price on software, I'd just have to put you on hold and consult with the front office staff. Besides, if I were calling all the shots at RDC, I would have to take myself more seriously. My door would not proudly display the "Mickey Mouse Executive Set" (complete with Mickey Mouse credit card, play money and phony keys).
Since RDC has been self-capitalized, we haven't been rolling in money. We've always been very cautious about purchases. We're unlikely to jump right in and buy new devices when they're announced. RDC did go out on a limb for a Thiel and for an Apple LaserWriter. Jesse, with her typesetting background, wanted a LaserWriter the moment she saw it. She had a hard sell convincing the rest of us. In September 1985 we started to phase in the Macintosh/LaserWriter system, and now we use it for just about everything--the Newsletter, manuals, promotional materials, etc.
As mentioned before, our customer record system is basically manual. For mailing lists and product data bases, we use the simple AppleWorks software. Kristi feels that our most significant purchase has been "Mr. Option" (her fancy memory typewriter).
To the dismay of many, Jesse has not yet explored naval architecture software for the Macintosh. If she did, it might help us do better in the annual "Dairy Carton Regatta." Our first boat made entirely of milk cartons did manage to cross the finish line without disintegrating. Our second craft, "Dairy Goes" displayed a pronounced tendency to pitch Kristi and Becky into the lagoon.
All in all, it's been an exciting six years. My hobby has grown into a real business. In 1984, callers on the tech line always asked to speak to "Dave Holloway." Recently I answered the tech line and the caller said, "David? Are you new around there?" To me, that's proof that we've developed as a team. We enjoy working hard and we enjoy our jobs. We definitely plan to be around long enough to add more chapters to this story.