Date: Tue, 6 Dec 1994 16:35:46 +0000 From: Simon Crosby Subject: DragonDictate and programming To: Multiple recipients of list SOREHAND I've been involved in an interesting discussion on sci.med.occupational, with somebody who believes that speech technology, in particular DragonDictate, should not be widely recommended for use in people with RSI, when they need to use it for programming. I enclose the original article and my response. Apologies in advance for the length of this posting, however I thought it would be of general interest. Any comments? Join in (on Usenet)! Article: 2278 of sci.med.occupational From: viking@narqlinq.net23.com [Simon.Crosby@cl.cam.ac.uk originally wrote] |> hmm... I can see why -- DragonDictate has no inherent support for |> programming. What you do to get it to work with a programming |> environment is to construct sets of macros in the DragonDictate |> vocabulary, and also add as much intelligence to the editor which you |> are using (hopefully something clever like emacs) and then it works |> fine. Unless you do something like this then I would be inclined to |> agree -- difficult to use. |> |> -- |> Simon |> Simon, what you described is *exactly* what you have to do. Dragon Systems *discourages* people from doing this, because it is *extremely kludgy*, slow and unworkable. Dragon has found that programmers with RSI are most likely to use a combination of typing and voice in their work, in an attempt to get around this problem. For most programmers, DD does not reduce the typing required by 70% but to around 50%, which is not enough to safely rehabilitate a programmer with RSI (This means a possibility of legal problems for Dragon if programming becomes an 'approoved' use of DD! Dragon Systems *strongly* discourages programmers from using it!) To figure out whether DragonDictate is appropriate for a programmer with RSI, several factors have to be taken into consideration: a. What's the health of the RSI programmer? How much typing can the programmer do without making matters worse? What's the functional level of the programmer's disability? (ie: What tasks of daily living does the programmer have difficulty doing, like cooking, cleaning, bathing, washing clothes, driving a car, etc.) Is the programmer still symptomatic? If so, is the programmer's condition basically improving, remaining constant or deteriorating? b. What's the nature of that programmer's job? How many different programming languages does that programmer use? Does the programmer use languages that are difficult to dictate, like lisp or assembler? How large is the universe of code that programmer responsible for or codes for? Does the programmer physically use many boxes or can the programmer be tied down to only one workstation? What are the nature of the supporting tools the programmer uses? How many tools, and what do they look like? Who shouldn't even think of using DragonDictate? If very casual typing increases the intensity of symptoms, don't even consider Dragon Dictate. If the person has a very serious level of disability: can't cook-- generally eats finger food, has difficulty doing day to day tasks, can't drive a car, the person should obviously not be working. If the programmer uses a very rich language, or has made a language very rich, programs in many languages, programs in languages that have utterly no resemblence to natural languages whatsoeve r (like assembler), or who programs in many languages regularly will have great difficulty with DD. If the programmer uses many different workstations. A programmer I know with severe RSI used to program control systems for heavy machinery. Naturally, DD could not help this person do his job. There are probably more contraindications than the ones I've stated above. Dragon Systems has discovered that for *most* programmers with RSI, DD is indeed contraindicated. It's *very* important for a programmer with RSI to try DD before purchasing it. Dragon Systems experience is that most programmers can not tollerate the limitations of the system and it does not allow them to effectively return to their line of work or something similar. It *is* useful for some people; for their work can be modified to avoid DD's serious faults. >From an ethical perspective, I would *strongly* advise that someone tone down the advise on the UK web server about DragonDictate. What's being said there is overly optomistic about the usefulness of DD for programmers. This web server should also include the advise and warnings made by Dragon employees about trying to program computers using DD. ----------------------------------------------------- My response follows: Thanks for the very detailed response. Clearly you are well informed about the DragonDictate position, and I'm glad to see that DragonDictate are looking into this seriously. However... In article <3bvoc3$mka@senator-bedfellow.MIT.EDU>, root@athena.mit.edu (0000-Admin(0000)) writes: > Simon, what you described is *exactly* what you have to do. Dragon Systems > *discourages* people from doing this, because it is *extremely kludgy*, Not really. > slow and Not at all - I'm faster than anyone I work with when putting code into a file. > unworkable. Absolutely not. On what experience do you base this comment ? [much other vaulable information cut] > > It's *very* important for a programmer with RSI to try DD before purchasing > it. Yes, I agree completely. > Dragon Systems experience is that most programmers can not tollerate > the limitations of the system and it does not allow them to effectively > return to their line of work or something similar. It *is* useful for some > people; for their work can be modified to avoid DD's serious faults. > In all my time of using DragonDictate I've never heard this position stated by the Dragon Resellers with whom I've had contact. All of this is true in the general case, except I've no insight into the findings of the DragonDictate people. Where did you get hold of this information? In my experience (admittedly based on my own ability to program my environment to help me) I've found DragonDictate to be a positive factor in my work. I'll not state it all again, however I use DragonDictate for absolutely everything, including assembler, C, shell and document processing. There are several others who do likewise. Without DragonDictate I would be without a job and without a career. I cannot type, and this is my only input mechanism. I do, however use a trackball to position the mouse cursor, which is very slow to do by voice. For me there is no option -- DragonDictate (or some other speech system) or no career. > From an ethical perspective, I would *strongly* advise that someone tone > down the advise on the UK web server about DragonDictate. What's being > said there is overly optomistic about the usefulness of DD for programmers. > This web server should also include the advise and warnings made by Dragon > employees about trying to program computers using DD. What is said on the server is simply the experience of people using DragonDictate. Sure, I would agree that people should try DragonDictate (or any other speech system) before investing in it. I also agree wholeheartedly with all of the points you listed above. The a2x group, RSI-UK and sorehand regularly see requests for help from people who are at a dead end (such as myself, about 18 months ago). People who are at a loss for help need to know the options, and all that the RSI groups do is let people know what others have experienced. The reason there are no negative comments in the faq is that I have *never* seen a posting which says that DragonDictate cannot be used for programming before. Obviously I will add your comments to the faq, and I would appreciate hearing from anyone who has found it to be unusable or contra-indicated. Presumably you are aware of such cases and I would be very interested to hear of them. I will also forward your post to the sorehand mailing list so that it is widely circulated. Thanks for your detailed comments which were extremely valuable. Please note that I have *absolutely* no financial interest in DragonDictate or any other speech technology. (this text all composed by voice) -- Simon [Simon.Crosby@cl.cam.ac.uk] [http://www.cl.cam.ac.uk/users/sac/]