GLEN GORDON: On FSCast 194, Matthew Horspool is here to show us how he uses JAWS Speech and Sound Schemes. He’ll also talk a bit about the challenges he faces working as a braille transcriber. Then I’ll be joined by Joe Sullivan to discuss the early years of developing the Duxbury Braille Translator. All for Braille Literacy Month, January of 2021.
Hello, everybody. Glen Gordon with you for our first podcast of 2021. One of the nice things about having hosted this podcast for over nine months now is that I think those of you listening are becoming a little more familiar with me and realizing that I am not an ogre, even though I play one in real life. And I say that because I’ve been getting lots more email from folks. And this is one of my favorite things to receive. Even if I can’t directly help you, I always like hearing about what’s going on for you, some of the issues you’re facing, some of the things you’d like our products to be doing.
I’m not tech support; but if you’re feeling particularly frustrated by something, maybe I can help grease the wheels a little bit. So feel free to drop me a line, email@example.com, with questions, comments. I’ll do my best to answer you. And again, I’m not just saying it. It is one of my favorite things, to communicate with some of you who I’ve never spoken to before.
One of those people is Timothy Jones. I talked to him for the first time at our Vispero NFB presentations back in July. But he wrote to me last month and sent a Power Tip that he discovered by accident.
TIMOTHY JONES: Have you ever gone to press a certain keystroke on your keyboard and accidentally pressed the wrong keys and then discovered a new keystroke that you didn’t know existed? Well, that’s what happened to me this week, and I would like to show you something really cool that you probably didn’t know you could do in your web browser. We all have used the traditional way of navigating from one tab to another by pressing the keystroke CTRL+TAB or CTRL+SHIFT+TAB to go forward and backwards through our tabs. But did you know that you could press numbers on your number row to move to tabs, as well? I’ll show you now how this works.
Let’s suppose you have more than one tab open in your web browser, and you know exactly what tab number you want to go to, and you don’t want to have to press CTRL+TAB a zillion times to get there. Well, the easiest way to get there is to press the CTRL key with the corresponding number on your numbers row. To move to tab number one, I will press CTRL+1.
JAWS VOICE: CTRL+1. The SnowMan’s Scripts for JAWS for Windows, Google Chrome.
TIMOTHY: If we want to move to tab number three, we can simply press CTRL+3.
JAWS VOICE: CTRL+3. Get iTunes Into Audio. Hartgen Consultancy, Google Chrome.
TIMOTHY: If we want to go to number five, we can do the same thing.
JAWS VOICE: CTRL+5, FSCast 186, Google Chrome.
TIMOTHY: And of course we can always get to these tabs using CTRL+TAB or CTRL+SHIFT+TAB, but then you’ve got to go one by one; whereas this method allows you to move directly to the tab you want with one simple push of a button. Of course the downside of this method is, if you have more than nine tabs, you won’t necessarily be able to use this method to get straight to the tab you want. But if you’ve got, say, 12 tabs open in your browser – which I’ll be honest, that’s pretty rare – you could navigate to tab number nine using the method I described earlier, and then you can use CTRL+TAB to go one by one until you get to the number you want to be on. Or the other alternative is you can just navigate the tab bar. But if you have nine tabs or less, this method can really speed up your browsing.
GLEN: Well, thank you, Timothy. I did some checking, and that works not only in Chrome, but also in Microsoft Edge and in Firefox. So assuming you’re using a modern browser, that feature has you covered. And for sending it in, we’ll extend either your JAWS SMA or your JAWS Home User License for another year.
Now, you may be saying that’s not a tip about a Freedom Scientific product, and you’re absolutely right. But we reserve the right to have some Power Tips about things that Freedom Scientific product users would take advantage of, and sometimes might not know about it. So part of a Power Tip is selling us on the utility of it and the likelihood that most people don’t know about it. And if you want to give that a try, either with a Power Tip in writing or with an audio tip, as Timothy sent in, do write to us, firstname.lastname@example.org.
GLEN: Joining me now on the podcast is Matthew Horspool. He’s been a user of JAWS for a really long time, and I’ve wanted to talk to him ever since Eric Damery told me the story that he was off in the U.K. doing some sort of JAWS education session, and there was this young, six-year-old kid who was answering all the questions that Eric posed. And then, as the session continued, older folks said, oh, yeah, I know who you are. You’re this kid who’s helping us with our questions on the JAWS mailing lists. And it turns out that that is Matthew. He’s no longer six, but he’s still a JAWS user. Matthew, welcome to FSCast.
MATTHEW HORSPOOL: Hello, Glen. It’s a real pleasure to be on the podcast. I should say I was more like 12 when that story was told. But that particular story that Eric tells happened in 2004, to give you some context. I had been given a computer through a charitable grant two years previously. And because people didn’t know what these computers could do, and people didn’t really understand the power of the Internet, my parents certainly didn’t, what would happen was I would just be told, yeah, you can go on the computer, and you can play. And I don’t think my parents properly realized until it was too late just how involved I’d become on these mailing lists and things like that. And it was great fun.
GLEN: I happen to know that your favorite JAWS feature is the Speech and Sounds Manager. Will you take us through some of the customizations you’ve made, show us how it makes your life easier?
MATTHEW: Sure, absolutely. So I am in the JAWS window.
JAWS VOICE: JAWS Professional.
MATTHEW: And if I go to the Basics dialog you’ll start to see things work. So if I press ALT+O for options...
JAWS VOICE: Options Menu, Basics, dot dot dot.
MATTHEW: And press ENTER.
JAWS VOICE: Leaving menus.
MATTHEW: So the first thing I’m going to demonstrate is a dialog box. I’ll use the Basics dialog box from within JAWS. So I’m in the menu, and I’ll just press ENTER now to open the dialog up.
JAWS VOICE: Leaving menus. Basic settings dialog. Tutor messages, announce custom messages only. Three of three. ALT+T.
MATTHEW: And you’ll notice it didn’t tell me that this is a radio button. I know it’s a radio button for two reasons. First of all, it said “three of three,” so I know that there’s a choice of three items in this case. And secondly, if I arrow up and down you’ll hear it. There’s a door knock sound which means checked.
JAWS VOICE: Tutor messages announce menu, and tutor messages turned off menu and CTRL+L. Tutor messages announce custom messages only, three of three, ALT+T.
MATTHEW: And this comes into it. So that if we skip past the next one, there’s another group of radio buttons.
JAWS VOICE: Automatically start JAWS dot dot dot, ALT+J.
MATTHEW: So it’s a fairly quiet click, but it’s a definite click nonetheless. And, I mean, I know my way around this dialog box fairly well. So I know what this is anyway. I’m going to press the SPACEBAR on this button and see if we can find any more interesting controls.
JAWS VOICE: Start JAWS options dialog. Start JAWS at the logon screen, ALT+J.
MATTHEW: So this is a checkbox because we didn’t get a “one of three” or a “two of three.” We just got the door knock.
JAWS VOICE: Start JAWS after logon for all users, ALT+...
MATTHEW: That’s another checkbox. But we should have a combo box coming up.
JAWS VOICE: Start JAWS after logon for its user, using the All User setting, three of three, ALT+T.
MATTHEW: So this was three of three, but it was three of three with an ascending tone.
JAWS VOICE: Combo box. Start JAWS after logon for its users. Using the All User...
MATTHEW: And if I get really confused about what the control is, I can press INSERT+TAB, which will tell me the control and play the sound at the same time. So I’ve now ALT+TABbed to possibly a familiar web page to listeners of this podcast.
JAWS VOICE: Navigating web pages. Google Chrome. Matthew. Main region.
MATTHEW: So we’re on the Surf’s Up navigating web pages page. I put perfect pitch to my advantage here. If I press CTRL+HOME to go to the top of the page and then move by heading...
JAWS VOICE: Navigating web pages.
MATTHEW: This is a heading at Level 1. And I can tell this because it’s played a piano note, C. And if I go down to the next heading...
JAWS VOICE: Reading text.
MATTHEW: This is a heading Level 2. And it’s played a note, D. If I move between the headings quickly, you’ll probably hear what’s going on here.
JAWS VOICE: Navigating, reading text. Navigate, reading text. Navigate, reading text.
MATTHEW: So you can hear that the tones are at different pitches. You might not be able to hear that. If you don’t have perfect pitch, you might encounter a heading and wonder what level that’s at because you haven’t got it in context. But I know because I’ve got perfect pitch. There should be a Level 3 heading on this page somewhere.
JAWS VOICE: Hyperlinks.
MATTHEW: Yeah. Know, you know, that’s a note E, and it’s a Level 3 heading. And you can play “Mary Had a Little Lamb” on this if you want to, to see how it works.
JAWS VOICE: Reading hyper, reading nav, reading hyperlink.
MATTHEW: Well, sort of, anyway.
JAWS VOICE: Skip lists.
MATTHEW: But it will go all the way up. So Level 4 would play an F. Level 5 would play a G. Level 6 would play an A. So the last thing that I’m going to do is walk through a Word document that’s got some different stylistic information. And this is where voices will come into their own. I tend to use voices more for style and sounds more for controls. And it’ll become more obvious why that is in just a bit. So I’m in this Word document. Let’s read the first line of this Word document to give us an idea of what the speech rate is like.
JAWS VOICE: I thought I would write this quick document to show how my JAWS settings are configured.
MATTHEW: So that’s the same rate as we’ve had all the way through the demo. If I drop to the next line...
JAWS VOICE: I am going to write something in bold.
MATTHEW: Notice that that line was considerably slower. If we go down to the next line...
JAWS VOICE: Notice how JAWS slowed down.
MATTHEW: So that’s gone back to the normal speed because we’re not involved anymore. And what will hopefully become apparent on the next line is that it will slow down at the point at which bold or underline or italics starts. So the first half of this next line is in normal type, and the second half is in emphasized type.
JAWS VOICE: It will do the same if something is underlined or in italics.
MATTHEW: So what I decided to do fairly early on was to use one voice to represent everything that was different. And then I could drill down. So if I look at this line...
JAWS: It will do the same if some...
MATTHEW: I know that something changes halfway through this line, and I can just go word by word. The final thing that I wanted to demonstrate because it’s kind of cool and not many people seem to know about it is that you can adjust how JAWS reads capitals. Normally when you navigate by character, it announces the capitals, and that’s it. But you can extend that so it announces it by word. And you can have different levels of pitch for single caps and all caps. So let’s have a look at this sentence.
JAWS VOICE: I am using JAWS for Windows, not JAWS for DOS.
MATTHEW: So there is a capitalization mistake in this sentence, and it will become apparent as I navigate through it where it is. So I’m just going to CTRL+RIGHT ARROW through this sentence.
JAWS VOICE: Am using JAWS for Windows, not JAWS for DOS.
MATTHEW: So notice how the second JAWS had a lower pitch than the first JAWS. Here’s the first JAWS.
JAWS VOICE: JAWS.
MATTHEW: And here’s the second JAWS.
JAWS VOICE: JAWS. JAWS. JAWS. JAWS. JAWS.
MATTHEW: And notice that the JAWS with J A W S all in capitals has a higher pitch than just the JAWS that has initial capitals in it.
JAWS VOICE: JAWS, JAWS, JAWS.
MATTHEW: And so that’s very helpful for me if I’m navigating, and I want to know how somebody’s capitalized something. It’s just a useful thing to know. Well, I find it useful generally; but particularly in transcription I find that quite a useful thing to be able to do.
GLEN: If Matthew’s demos have gotten your appetite whet for playing around with Speech and Sounds Manager, there are a couple of places where you can get a little more tutorial-based help for setting it up for yourself. One of them is back in FSCast 148. And if you go to your favorite search engine and type in “FSCast 148,” you’ll get to the archive of our older episodes, and you’ll find there that Jonathan Mosen does a really complete job of demonstrating setting up some of these features. Also, if you go to our Surf’s Up web page on there, there’s a little bit of a written tutorial on using Speech and Sounds Manager. So those are the two most obvious places to look. Matthew, other ideas for ways people can get started with this?
MATTHEW: I think I would probably not rush into it too quickly. I would maybe change one or two behaviors, get used to how those behaviors feel, and then change another one or two behaviors. All of those sounds in the dialog boxes, for example, I didn’t start off with all of those sounds. I started off with turning checkboxes off, I think, and radio buttons, and just having the knock-knock sounds. And then I had a different sound for edit boxes. And then I had a different sound for password edit boxes. And it sort of gradually grew. And then as I got used to what the different sounds meant, I would introduce more of them. So I wouldn’t rush into it when you’re actually ready to start making changes. Otherwise you will very quickly find yourself overwhelmed and probably not benefit very much.
GLEN: When you were doing your demo, you mentioned a couple of times that you have perfect pitch. Is that a blessing or a curse?
MATTHEW: It’s a mixed blessing. And I’ve had to tame it over the years. Which was made easier by singing in choirs because what happens when you sing in choirs, especially if you’re singing unaccompanied for a long time, is that somebody starts to go flat, and then somebody else follows, and then somebody else follows. And before you know it, you’re singing flat. And this would really bother me, and I would want to, when we got to the end of each section, I would want to keep resetting the pitch. And a bit like being told not to rock, I was also told not to keep resetting the pitch because most of the audience wouldn’t notice that the choir had gone flat. And so to reset the pitch makes it obvious to the rest of the audience that something’s gone wrong.
And so, yes, something’s gone wrong. But just as you would, if you make a mistake, you pick it up and carry on, you need to take going flat as a mistake and just pick it up and carry on. I found it very difficult. I knew that the next note was a G. I didn’t know that the next note was one note lower than previous note. So I then had to transpose in my head all the time. And so over a period of several years, I learned out of necessity how to have relative pitch alongside perfect pitch. And so if you give me a tune now and ask me to transpose it, I can transpose it in my head very quickly and sing it in the new pitch. But that took an enormous amount of work to get to that point where I think somebody who didn’t have perfect pitch wouldn’t necessarily have that problem.
GLEN: And so do you go through the world hearing things? And you know, you hear a doorbell, and you say, oh, that’s off. That’s off a little bit.
MATTHEW: I used to. I’ve turned that off. I made a conscious effort to turn that sort of thing off because it was just a distraction. Where it really came into its own, and what really helped me to turn this off, was for a while when I was at college – I went to a College for the Blind down in Hereford that was a boarding college. And they had this thing called “acoustic air rifle shooting.” So I don’t know if anybody’s used a light probe, you know, as the light gets brighter the pitch gets higher. It was the same sort of thing, but with target shooting, so that the closer to the center of the target you got, the higher pitched the sound became.
But what happened was I would start off, and I would say, right, I need to aim at the note C. Whenever I get to the note C, I need to shoot. And this worked very reliably until you got to the note C and shot at it and then realized that you had to put two shots per target, not one. So once you’d shot out the center of the target, that note C had gone missing because you’d shot it out. So I’d be looking for this note C and not find it, when really what I needed to do was look for the highest pitch that I could find.
And if I wanted to get better at the sport, it was advantageous to switch off all of the perfect pitch completely and just every time find the highest sound and not pay too much attention to what the highest sound was. And that sort of trained me out of doing that sort of thing. So no, I don’t listen to a police siren and go, oh, that’s a B and an A or whatever, you know. That doesn’t happen anymore. It did used to.
GLEN: So how did you go from being in University – I hear rumor you were in University for computer science – to becoming a braille transcriptionist?
MATTHEW: I was. By accident, as so often happens. So I had been doing a lot of braille transcription for my own purposes as a child, getting things ready for church and choir and things like that. While I was at University, I needed some what we would call in the U.K. “mobility training.” And I struggled to get hold of it because of the way the funding models work over here and things like that. So in the end I got hold of somebody who worked at a school for the blind in Coventry, which is where I was at University. And so an arrangement arose where I came into the school for half a day a week to help with braille maths. And in return I got orientation and mobility. And I enjoyed this a lot more than I expected to.
The school eventually realized that there might be a job for a braille expert, and so they put together a job description, and they suggested that I apply. So I applied, and I had an interview and was offered the job. And I had to make a decision at that point about whether I took the job, because I wasn’t really expecting to get it. I thought somebody more qualified would apply, and they didn’t. And I decided that I’d do it for maybe two or three years, give myself a bit of a break from education, and then go back to it. And actually I found I enjoyed it so much that I’m not still in that job anymore. But I am still very involved in braille and transcription and standardization and all of that sort of thing.
GLEN: Why did you become sort of disenchanted with computer science?
MATTHEW: I basically didn’t want to solve other people’s problems. The appeal of computer science to me was it would solve my problems, and it’s done that. And it continues to do that, and it does that very well. But when I’m put in a situation where I’ve got to solve somebody else’s problem, and they probably want to solve the problem in a different way to the way that I want to solve it, and I just didn’t want to be dealing with that.
So I’d worked out that I didn’t want to be a programmer. What I hadn’t done was worked out what I did want to be. So I was very directionless, and I just felt that the statistics were so swayed against blind people getting jobs that, yes, I’ve got the potential to do it. But unless I’ve got a real sense of direction about what sort of job I’m going to get, I’m probably not going to get one.
GLEN: Was it challenging to be a transcriber and not be able to look at the original print layout of what you were transcribing?
MATTHEW: It would have been if I was transcribing from hard copy print. I call myself a “transcriber” because that’s all I can call myself. And I think, broadly speaking, it’s a fair assessment of what’s going on. But I think if we wanted to be particular about what I was doing, it was predominantly editing, rather than transcribing. So what would happen is documents would originate as a Microsoft Word document, or sometimes as a PDF file, or sometimes as a PowerPoint presentation, but generally as a Microsoft Word file. And all of the text would be there, but there would be problems.
For example, instead of using heading styles that were built into Word, teachers would be lazy, and they would just put it in bold or in underline or whatever they wanted to do to their headings. Lists were not styled up as lists all of the time. You would have setups where, for example, one particular one I remember very clearly was there was a table, and the table had two rows and about 15 or 16 columns. And that sort of table, of course, would be far too wide to braille.
But if you know a thing about tables, and you know a thing about braille, and you know a thing about technology, you can combine all that knowledge and essentially flip the table round 90 degrees. So instead of having 20 columns and two rows, it had 20 rows and two columns. And that way the table was able to be brailled properly, you know, you could braille it as a table. And I believe a transcriber’s note at the top to say this table is being transposed. So it was really editorial things like that, restyling it to put the styles in the right place, adjusting tables so that they read properly. Sometimes shortening tables. You’d have a column that was called something like “Number of children in class.” That would be the title of the column, “Number of children in class.”
GLEN: Yeah, that’s a whole braille line.
MATTHEW: Exactly. That’s a whole braille line. But the numbers would be three, six, 17. And, yeah, so you understand the problem. The column header was way longer than the actual data in that column. And that doesn’t really matter so much for print, but it matters enormously for braille. So you’d do things like you would shorten the header to “Children,” and then you would put a note at the top of the table that says, “When you see ‘Children,’ the print actually says ‘Number of children in class.’” And sometimes, if that didn’t work out, you’d have to work out how to present the table in a linear format or something like that. And so it was all that sort of work that I was doing.
But the source documents were electronic to begin with. They were in Word generally. There was only a very minimal amount of print, hard copy print that I needed to scan in. And it was generally very simple. And I’d got used to doing this sort of thing through all of the church stuff that I was doing. I’d got used to scanning in hymn books and scanning in service papers and things. So I knew already the sorts of mistakes that OCR software might make and generally found it fairly straightforward to correct them.
GLEN: Kids growing up today I think have much more of an option, at least potentially, of completely avoiding braille and using speech exclusively. What are your thoughts about that?
MATTHEW: Fundamentally, I think if you’re going to teach a sighted child how to write with a pen and paper, I think it’s only fair that we teach blind children how to write with a braille machine of some sort, and read braille. And I think braille encourages spelling, good spelling. It encourages good grammar. I think if we get beyond the spelling and grammar arguments, though, and onto writing, just writing style, you’ll notice, or at least I certainly notice, if I’m reading in braille, and there’s a long sentence, I will notice that, hang on, I’ve read six lines of braille, and I haven’t found a full stop yet. Versus listening to speech at, say, four or 500 words per minute, those six lines go pretty quickly, so you don’t always realize that you should have had a full stop. You know? Things like that.
Huge long paragraphs. I’ve written documents on a computer, and I’ve thought, yes, yes, these paragraphs are perfectly acceptable. And then I’ve gone and embossed the document and have tried to read it with my fingers and gone, actually, no, these paragraphs are too long. And that is a useful process to go through because if you, as a reader in hard copy are starting to go, look, these paragraphs are a bit too long, I think there’s a good chance that a sighted person reading your document onscreen is also probably going to be starting to think these paragraphs are a bit long. And then they don’t read your writing because it just looks like this mass of letters that they can’t skim very easily, and they’re not terribly interested in.
Reading braille gives people the opportunity to experience page numbers. I know plenty of blind people who say, well, I don’t know why I should bother putting page numbers in my Word documents. And I used to agree with them until I read a braille document that didn’t have page numbers. And then I realized how much of a problem it was, and I’ve never forgotten to put page numbers in a Word document since. Running headers, running footers, a sense of spatial layout. Okay, we can’t do in braille what people can do in print in terms of making headings stand out. But we can make them stand out sort of. We can center the heading. We can leave a blank line before the heading. We can leave a blank line after the heading if we want to.
You know, if you’ve seen a braille document with this degree of spatial layout, I think even if it’s not an exact science, you know, that this in braille means this in print, it gives you enough of an appreciation, I think, of what spatial layout might look like that you’ve got a better chance of making a print document that looks visually appealing because you’ve got some sense of what it is you’re trying to achieve.
GLEN: Well, Matthew, I thank you for joining me on FSCast. It was great getting to chat with you. You are both animated and have interesting things to say, which is a great combination.
MATTHEW: Well, it was so much fun. Thank you very much for inviting me, Glen. And long live JAWS, and long live braille.
GLEN: Even if you don’t read braille yourself, you’ve probably learned the basics over the years, that each character is made up of from between one to six dots, arranged in two vertical columns. And it’s different combinations of dots that form letters and symbols. But every braille code, in addition to having a one-for-one representation between letters and braille combinations, has its own set of contractions. Things like G in the U.S. English braille code means go. H means have. Z means as. Even combinations of letters form words. GD means good. WD means would. And then things like the letters D I S in English at the beginning of the word have a braille symbol to represent them.
There are lots of rules, codified in some form, for all languages. And these are things back in the day that, if a blind person needed someone to transcribe a book for them, the transcriber needed to memorize all the rules because they had to create braille on a Perkins Brailler. But over the years, as computer technology has evolved, the idea that someone needs to know the braille code continues to be useful because there are also ambiguities that occasionally you’ll want to resolve manually. But it’s not 100% necessary anymore. So schools and other places that need to produce braille have been able to delegate people who are good workers, but don’t necessarily know braille really well, to producing some of this material.
And all of this is due in large part to a man named Joe Sullivan. Joe and his partners, as we’ll hear, formed Duxbury Systems to create the first really commercial braille translator. And fortunately for all of us, Joe has joined us to talk about those early years of Duxbury and the translation software that has enabled so many of us to take on the tasks of doing braille production on our computers. Joe, welcome to FSCast.
JOE SULLIVAN: Oh, thank you very much, Glen. It’s a pleasure.
GLEN: So what was your initial contact with braille? You’re not blind, at least not to my knowledge. Nobody in your family is blind. How did all of this come about?
JOE: Well, a coworker of mine at MITRE Corporation was a blind person whose name is Bob Gildea, was my first introduction. MITRE Corporation is a military, mostly Air Force-oriented think tank. And we were both employees there, obviously working on military projects. And I asked my boss at one point, name was Bill Amory, fortunately a very sympathetic guy about these things, if I could work on something not military. And he said, you know, Bob Gildea has just got this project started in conjunction with MIT and the Atlanta Public Schools to see if it would be possible to build something that could actually automatically translate from print to braille.
I should say, by the way, that a number of people before me had become involved, including Bob. So I really joined the project. But it was later Bob and I that, together with a friend of his, Anne Simpson, who was a professor at MIT, who decided to take the project out into the commercial world because MITRE being a not-for-profit could not actually produce a product, you know, in a commercial sense. In fact, officially we were just supposed to examine the feasibility of doing braille translation.
GLEN: And at the time did you think there was a doubt that it could be done?
JOE: The machine’s capacity at that time was a bit limited for something like this. It was thought to be complicated because you had to, you know, to do it perfectly you would have to understand the meaning of words in some cases, the sounds of words and meanings of words. So it was kind of on the level of text-to-speech, the complexity of the project. And a lot of that was just starting to be done at that time. This is like in the late 1960s – ‘69, ‘70. So there was concern that, you know, it might not be possible, in a perfect sense. But when you set aside perfection in an absolute sense, you can actually do something useful, which is what we set out to do.
GLEN: Now, I read on your website the other day, much to my surprise, that I think it was the American Printing House for the Blind had some kind of braille translator in the early ‘60s.
JOE: That’s right. They were probably the first to have something automated. It was, from my understanding of the architecture, it was essentially a kind of a dictionary lookup sort of thing that would spit out any word it had not seen before. And then a human would intervene and say how that particular word would be handled. So I would think of it as, in a sense, semi-automated. But apparently the dictionary was big enough that they were able to run most books through.
What we set out to do was primarily to, first of all, to write it in a higher level language so it could be used on other, you know, on various kinds of hardware, and also get away from straight dictionary lookup because we wanted it to be able to handle words it hadn’t seen before because it could be used by somebody who didn’t have access to a transcriber who could alter the dictionary.
GLEN: Now, at that point were there braille printers that this could be sent to, once the translation was done?
JOE: I think it was called Braille Emboss, the one they had at MIT. And that was, you know, designed as a page embosser. There were others that came online that actually embossed on paper tape. And so what you had was a strip coming out that you could read. And so it was almost like you had a braille display. And I suppose you threw the paper tape in the wastebasket. So it was ephemeral to read one of these things.
GLEN: Well, actually I wanted to stop you because when I went into high school, a friend of a friend’s father was in the Lions Club, and they bought me something called the Banks Pocket Brailler. It was this little Perkins-style braille keyboard that you could emboss on paper tape. But it was really small. And so I would carry it to class, all the way through college, and I would have these long rolls of tickertape.
JOE: That’s right. Well, it works; you know?
GLEN: Yeah, apparently.
JOE: I mean, it’s whatever works.
JOE: And but a fellow by the name of Guy Charbonneau came along later and produced something called the LED, Line Embossing Device. That was also a page brailler. That was probably the first one in commercial use. The other thing we did in the original DOTSYS project was configure the program so that it could use the old line printers that were used commercially in those days. If you put an elastic band behind the paper, and just used the period in a certain way, you could actually create braille.
GLEN: Was it around this time, after you had proven that this works, that you decided to go out with your partners and try to commercialize it?
JOE: Yes. That happened in 1975, actually, we had our first meeting at the home of Anne Simpson. And Anne contributed some subroutines and other things, but the main contribution was a computer for us to work on, which was one of the data – a Data General Nova. But it was pretty meager by today’s standards, to say the least. That was housed in her house at Duxbury, which is where the name came from.
GLEN: So was this for the thrill of getting it done? Or did you sort of see the future, that people who needed to produce braille wouldn’t necessarily need to be really great transcribers because a lot of this stuff could be automated?
JOE: We thought of it, first of all, as an aid to transcribers, that our main customers would be existing braille transcription houses. And it would let them get a lot more done in a shorter time, but not without the aid and oversight of real transcribers, human transcribers. Which is really where the first sales were.
In fact, you know, it was like one sale per year for the first few years. First was Canadian National Institute, and then the Clovernook School for the Blind, I believe that was ‘78. And then the third one was in Australia, actually, the Royal New South Wales Institute for Deaf and Blind Children, who had heard about us from the Canadian group. I think we had – I thought at the first that this would be a way of launching a business that would probably move out into other things because we didn’t see this to be an enormous market. Of course everyone was then, as they are still, thinking in terms of braille becoming obsolete in 10 or 15 years or so.
GLEN: Well, it might happen in 10 or 15 years. We never know.
JOE: Yeah, it always seems to be 10 or 15 years that it’s going to go away. But it never does.
GLEN: So in those early days, how were you able to verify that you got all the rules right? Was there any automation? Or was it just brailling a lot of stuff and having someone review it by hand?
JOE: Pretty much the latter, yes. The first installation, as I say, was at the Canadian National Institute for the Blind. And one of the main bodies of text that we used was a set of problem words that Krebs had devised called the – it was 5,208 words. And when I went up and installed the program at the Canadian National Institute for the Blind, the transcriber there, you know, of course they had to do kind of an acceptance test.
So they ran a few samples through, but then they also took this file that I gave them. I said, “Okay, and here’s a set of words that are known problem words. You could want to have a look at those.” Well, I’ll never forget the transcriber running through and stopping on one of them and saying, hey, is this right? And I said, oh, I think so. And I looked it up quickly, and I said, “Oh, that’s the way Krebs does that.” That was kind of my first clue that this is not, you know, absolutely cut and dried, the way that it’s sometimes presented; you know?
JOE: There are some judgments involved, and not everyone thinks of it the same way on a given word and so on. And it’s still the case, by the way; you know.
GLEN: Yeah. When did you move into the realm of doing languages besides English? And was that a challenge not being a native speaker of those languages?
JOE: Well, yes, it is a challenge. That is true. And actually that happened very quickly. People in Spain, a gentleman by the name of Pedro Zurita – that is still around, and I still communicate with sometime, he’s a great, great guy – he speaks eight languages, reads contracted braille in all of them. And he contacted me and said he’d really like to do the same thing for Spanish. So just to, I don’t know, just to prove a point, it’s one of those things, you know, you want to see what you can do with it.
I put together a table to do contracted Spanish. And Pedro basically had me over to Madrid, and we put it in and got it going. And so that was like ‘76, I think. It was only about a year later that we worked on Spanish. And then not too long after that we did French and Arabic. And Arabic was a real crunch because of course it’s written very differently. I had to learn even how to read the Arabic alphabet. But that’s still way short of learning the language. But you have to learn, again, something about a language, about how the language and the contracted braille relate to each other. Which is interesting. I found it just as an interesting intellectual exercise. I always wished that it would lead to learning how to actually understand eight languages. But I’ve never got there.
JOE: You know, it’s too much.
GLEN: I mean, this would be a lot easier now, in the world of the Internet, because you could easily pass files back and forth. You could get someone to review. You could do really quick turnaround. Did you have to, like, go to all those countries or bring someone to the U.S. to work with you to get this stuff done with any speed?
JOE: Yes and yes. We mostly – I would go. I went to Saudi Arabia a couple of times. I went to France a number of times. South Africa, same story, to work with people there on Afrikaans and later on other languages there. So, yeah, there was a lot of interesting trips, especially in the early days, to go and work with people.
Of course it was still possible even then to send files back and forth, became possible with tapes and so forth. But usually you’d make a lot more progress by going and working with the people in the remote place. Nowadays, though, as you say, not necessary. We just send files and do all of our communications through the Internet. So it, you know, is kind of sad. We don’t have go see these interesting places. But we can get a lot more done, you know, in that way.
GLEN: You probably don’t want to weigh in on Nemeth versus UEB Math or UEB versus Nemeth?
JOE: Well, of course, I was deeply involved in the UEB project, having chaired the committee that did the design, in fact, and worked with Abe Nemeth, great guy, and also Kim Cramner and Emerson Foulke and people from other countries. Some people think of UEB Math versus Nemeth code as being kind of in competition. And I don’t think of them that way. Nemeth code has the title of – the book says it’s a code for math and science. And it does a great job of that. But that’s its purpose.
UEB was set out to be a general code that was really a literary code base – because that’s what 99% of literature is, is what you have to call “literary” – but a code that would also encompass mathematics so that, particularly beginners and general readers and so on, who may not be interested in becoming technical wizards, would not have to learn a different code, but could still read something like 1+1=2 or even something a lot more complicated, if need be. But there would be a seamlessness between the literary portion of a text and the math.
So it’s a different purpose. And what we set out to do, we feel we did. And we do see other languages now picking up that same philosophy for a general code. So I don’t think of them as in competition. It seems to me perfectly logical that somebody who becomes particularly a professional mathematician or whatever might well want to work with Nemeth code when working on mathematics.
GLEN: Well, Joe, I don’t think you’ve been on our podcast before. I did look back. I have missed Jonathan Mosen interviews in the past, but I don’t think you’ve appeared, and I’m really glad to have had the chance to chat with you.
JOE: Well, it’s been my pleasure to talk to you, as well. So anytime.
GLEN: That does it for FSCast 194. I’m Glen Gordon. Thanks as always for listening. We’ll see you next month.