The following narrative was edited from a tape recording made during the Disability Action Committee for X (DACX) Meeting held recently at the Closing-the-Gap Conference October 20th, 1994 in Minneapolis, MN.
Meeting content and speakers are identified and reported herein "as-best-as" they could be determined from the recording.
Mark Novak (menovak@facstaff.wisc.edu)
Mark: We don't have a long agenda tonight, but we have a couple of interesting topics to discuss and I am very happy to see some of our developers, as well as people from around the world in attendance. First thing on the agenda is to get an update regarding the built-in X server features regarding physical access. I would like to have Earl Johnson give us a quick synopsis.
Earl: Backing up a bit, there is a version of the X Window Server that comes out every couple of years and the current one that has just now come out is called the X11R6 Window Server. What we (DACX) have done, is the AccessX features have been incorporated into a server extension as a part of the window server which will be coming out probably in the first patch, known as the XKB keyboard extension. Probably when companies such as SUN, DEC, and IBM, etc. start shipping their products with the X11R6 Window Server, what will be incorporated into it will be the AccessX features. Right now XKB is a work in progress which means they almost got it to a standard, they are almost ready to include it as part of the standard binaries for the window server, but it is not quite there. As an interim, what SUN and DEC have done, is they've included and bundled in their windowing system AccessX as a part of their product line. As of today, both the SUN and the DEC products, if you buy the latest release, AccessX is there by default on the systems.
Mark: Just to reiterate. At this point and time both DEC and SUN are shipping the latest version of the window server, X11R5 version, which has the AccessX included. For those of you new to the DACX group, AccessX is a clone of Easy Access, or AccessDOS, but does not yet include the SerialKeys function.
Earl: That is correct. When the X11R6 window server comes out and X vendors start shipping it in the product line, HP, DEC, SUN, IBM, etc. will all have this capability incorporated in their work.
Mark: One of the first things DACX as a group decided to work on, was to look at some of the mobility related needs and get some of these straight-forward access features built right into the X server where they belong.
Earl: Let me add a little bit more. AccessX is kind of an end user solution. It is kind of a homerun hit where we put the capabilities in the server, and we built a user interface that the user interacts with. There is not going to be any adaptive technology developers utilizing it most likely. But in some of the other things DACX is involved with, like RAP and the screen magnification extension, DACX is essentially developing the infrastructure, so that the adaptive technology people like Berkeley Systems, Telesensory, Flipper, Arctic, Microsystems, etc., etc., so they can all develop solutions and port their solutions over to the X platform. What DACX is doing, is providing them with a strong, clean, and a stable hooking point so that their solutions will work robustly with all the applications on the X Windows System.
Peter: Because X Windows is not owned by any one, it is rather to some extent, owned by everyone. It is developed by consortia of individual companies that ship UNIX based workstations that include X, take from this consortia derived version and modify it as they see fit and then ship workstations using their version of X. Just as SUN and DEC are shipping and modifying versions of X11R5 that contains keyboard modifications, AccessX. Some of the things that DACX has done is take those modifications that were developed through DACX and implemented first on SUN and DEC hardware, and are rolling it into the next release of X Windows. These same future versions of X Windows are containing more and more hooks that will make the job of the making screen readers and screen magnifiers significantly more attractable then they were in the past.
Question? Regarding screen readers, content in X Windows?
Peter: There is zero on X Windows that a user can use today out of the box regarding screen readers. There is significantly more in X Windows for developers such as myself to use to create a screen reader.
Question? Those are already in now?
Peter: Those are in X11R6.
Question? Are they finished?
Peter: There are hooks in X11R6 that make it easier. There are more hooks that we need in other places that we need to make it easier still. It is fair to say that some of the work that DACX has done is in X11R6. But no one is shipping X11R6?
Question? In X Windows, I use UNIX, and I really love it, it is really powerful. What does X Windows, give me apart from the GUI, what everyone else uses. (More text in here that was not clear) What am I losing out on right now by using a character based approach?.
Peter: The big thing initially, it gave people when they started developing it, was better access to large screen real estate and multiple windows that they could do things in. Imagine, here you are under editing your program, so instead of having 24 x 80 character cell screen that you are editing your text in, you can have two windows, each window is over a page tall and a full 80 columns wide so we can see a lot more of your code at once. This is not something of specific benefit to you necessarily if you can scroll rapidly with your screen reader, then it doesn't matter. This is a big first thing. After that, the ability to move away from commands that you had to remember, to menus that would appear at the touch of a key, that would give you all of your options, and to do it consistently, so that the same menu interface was from one program to another to the third to the fourth. Another thing that GUIs give you is the ability to see visually on the screen what the page will look like. It is very hard to do proportional space text on a monospace character 80 x 25. It is very hard to do superscripts. It is very hard to do multiple fonts or italics. All of these things were the initialize benefits of GUIs. The initial benefits of windowing under X on your UNIX based systems. Stepping away from that people started to involve graphics more. Here you are, you are setting the breakpoints in your debugger. One way of doing that is to type line such and such break, using the debugger you might do it that way. Another way is that you see the line of code on your screen, you move your mouse to the left edge of that code, and you click to the left of the line in a special column, and all of a sudden a little icon of a bug appears, or an icon of a stop sign appears or whatever icon the debugger chooses to use. That is saying that whenever the program comes to this line it will stop, or you can have a different icon to indicate that this was a conditional break or whatever. So you get a lot more feedback using these icons rather than having lots of text to describe what this breakpoint is.
Peter: Those are the primary reasons for X. That is why it was developed, in my opinion. The benefits to none sighted users are consistency across applications and the ability to get at all of the commands without having to know them ahead of time. In the Macintosh, where you have many more consistencies then even MS Windows or X, every program has several menus that are sitting at the top of the screen. File menu that will always contain the quit command, always contain the save and print commands. The edit menu always contains the commands for editing your text and so on. These menus are broken up by group, so you can immediately go to a menu title and know that all of the commands relating to that category are there. It is very, very useful to learning new programs.
Paul: Regarding the Windows "usage" tutorial. Doug Wakefield in our office has developed a tutorial called "Opening Windows", and it is a tutorial to teach concepts of GUI interfaces geared primarily towards windows, the Windows Operating Systems, but also teaches general windows concepts. It is distributed free to anyone who needs it and it is a series of audio cassettes, and a diskette to run on DOS systems and was developed by a blind computer user who has much experience in writing tutorials, so it is very well done. If anybody wants it, give me your contact information. It is going to be awhile because we went through the first run of them very quickly.
Comments regarding the concept of multi-tasking versus multi-processing, and the understanding of these concepts for the visually impaired as well.
Earl: By having multiple windows, you get a number of command shells up and just cycle through each one of the windows by calling up a window, you essentially open up another window to multi-task with. It kind of does it automatically. The other thing, on a more practical level, is the application developers are not generally developing command line applications anymore. They are moving away from the command line interactions. So what is happening is that the greater capabilities are opening up in GUI land and command line land is kind of staying flat. That is not entirely true. That is from the users point of view.
Curtis: Suffice it to say GUIs are here, whether we like it or not. And unfortunately soon...
Peter: I think there are tangible benefits of GUIs to people with visual impairments and people who are totally blind.
Curtis: Its tough to convince somebody of that when their job is on the line as you know, Peter? I agree with you. In the long run absolutely.
Peter: I am less convinced, even in the state of consistency or lack there of in X Windows applications, the benefit to UNIX users of X other than employment. That is where the future is going. Now that we have Motif getting adopted by more and more people, more and more products are being written in Motif. Now that Motif is being adopted fairly widespread, we are going to start to see the benefits of consistency from application to application because they are all going to be the Motif standard interfaces and Motif standard menu commands, etc.
Question: Do you suppose there will be a product in five years?
Comment: A screen reader for X Window in the next five years.
Mark: We are going to touch that a little bit as we move ahead on the agenda, so can we save that question?
Earl: Yeah, I will wait to address that. I also want to talk a little bit about the Motif question.
Comment: I guess IBM is the only company I consider to be seriously working on this.
Mark: Can you hold that thought and bring that up again at the end. Tell us if you are still not convinced. I am going to move on to the agenda so we can get to some of the new issues. We kind of did a quick review of what X Windows, did a quick review of what some of the development DACX has been involved within X Windows regarding access for mobility challenged. I would like to move into some of the areas which the DACX group is working on currently. That involves some of the things we just discussed a little bit in the area of some of the hooks that have gone into X11R6, which all the vendors now currently have. Those hooks were targeted to make it easier for third party developers to write screen readers and to develop other products, such as screen magnification. I would like to turn the meeting over again to Earl and ask Earl to brief us, because SUN has been leading the effort to provide the APIs .
Earl: I think what I will do is just kind of a quick history of what we have gone through. We have a proof of concept for screen magnification. We brought a SUN system here, it is in the Trace booth. We have the systems set up with both AccessX and this proof of concept that tests the server extension that we are writing for X screen magnification. Regarding screen magnification, what we have done to date, is we sat down and we sent out to a number of disability aliases to solicit for users trying to get feedback to find out, with the screen magnification systems that they are using right now, what kind of features are missing that you would like to have, what type of tasks do you want to do that you can't do, etc. To try and get back from the user to find out what type of capabilities may not be available in screen readers currently today. Unfortunately, there wasn't really too much user feedback. If it makes sense to send it out again, however, we can send it out. At the same time, we have been doing a competitive analysis looking at various screen magnifiers to find out what features were available in current systems out there. We developed a feature list and then from there, we started developing a prototype server extension. The capabilities that the server extension will do right now is essentially kind of pixel duplication making everything bigger. Again, those of you who checked it out will be able to see the system at the Trace booth. There are still features that need to be added to it and what we are in the process of doing right now is identifying the features that we need to provide so that it makes it easier for adaptive technology developer of the screen magnifier to port their solution to the X platform. We also need to be careful so that as we are doing this, we want to provide as much as possible but we also want to allow them to distinguish themselves from the competition too. So, this is kind of a hard part in okay, what areas should we make sure we provide functionalities in screen magnification. What areas can we look to the adaptive technology developers to provide. One of the things we know we are going to be adding and it is a big challenge in the X Windows environment, is something I am calling the "coincidental" challenge. In X Windows the system that we've got is there, if you seen UnWindows and PUFF, it is essentially similar to that right now. We see the problem for the user being, if the user wants to say click on an icon, but wants to be able to see that icon before they click on it, we think that the magnification window should be able to follow the pointer around. They need to have that opportunity. Well, if you provide that capability in X Windows, if you double click on that icon in this magnified region, you are double clicking on the magnified window, you are not double clicking on the icon underneath. So what the challenge here is that we need to make that window transparent so you can double click on the icons underneath. You can interact with the system underneath and that is kind of the biggest challenge. There is a couple other features we will probably be adding to the server. We are working closely with Will Walker at DEC and one of the main reasons we are here at this conference is to talk to some of the people like Berkeley Systems, etc.. Have them review what we have come up with, review the feature list and give us an input for the types of capabilities they want to see in the server extension also. That is kind of where we are with screen magnification. We're working through the X Consortium, who is the keeper of the Windows server. They are very interested in what we are doing, they are behind what we are doing and they said go ahead and we would like to see it standardized. We just kind of got to wait to see what occurs.
Mark: Then to summarize what Earl said regarding work that went on in the DACX group. There was a lot of heavy email traffic back in early May and June and contacts were made by people from DACX, mainly SUN employees, talking to some of the different developers from both the Macintosh and PC world to get some of their ideas for what they would want as a third party developer to encourage them to port their products to X Windows. What Earl is also talking about is trying to walk that line, what can DACX provide as an X system extension, to provide the basic ground work, such that developers would want to create their own products and create unique products to port them for their other platforms to X. The DACX effort is up to that line. To develop an extension and work with third party developers to make it easier to create the actual screen magnifiers. DACX is not trying to develop a screen magnifier for X.
Earl: Let me add a little bit about the beauty of DACX. DACX is a vendor neutral organization, we've got IBM, we've got DEC and we've got SUN that are members of DACX. So we believe that we need to take a standards approach with both the X Consortium as well as the Motif tool kit. By going through the vendor neutral approach, we do away with the kind of political battles that can occur as to who is submitting it, what company is submitting it. By brining it in the vendor neutral solution, what happens is the X consortium or OSF, judge what we (DACX) are doing. Judge the work based on the merit of the technology versus who is presenting it. That is why DACX has had a lot of successes. AccessX is one example and we will be getting into others in a little bit.
Earl: The last thing I will say is that the screen magnification opens up all sorts of challenges just as the screen readers.
Mark: That is what is going on with the screen magnification area. Like Earl said there is a SUN workstation in the Trace booth. Please stop by there tomorrow during exhibit hours and take a look at this proof of concept screen magnifier.
Mark: The next item on the agenda is going to be addressing screen readers. Unfortunately, some of the people who are doing most of that work and who are most technically versed in that work area, were not able to be with us tonight. I am going to speak for a couple of them and speak very softly, if you will, because I am not at the technical level of the individuals doing this work. Will Walker from Digital has been heading up the RAP development at this point. For those of you who are new tonight, RAP is the Remote Access Protocol. That is the communication link which DACX is trying to develop which works between what are called the special agents and X clients. Special agents in terms of DACX are the screen reader at this point. Backing up just a little bit more. There are three actual aspects to solving the screen reader problem, if I can put this into three simple aspects, which do not do this topic justice, but perhaps make it more understandable to those of you new here tonight. First, there are the hooks if you will, built into the underlying libraries of the X system. That is hooks into the Xt Intrinsics, and a hook into X Lib. This work has been completed in X11R6. Second part of that puzzle was developing a communication protocol, like RAP, which would work between the special agents or the screen reader and the client which is running on the screen. The third and last piece of the puzzle was what DACX needs to do to modify the graphics user interface, Motif itself. There is still some information, despite all the hooks put into X, despite the great protocol going on back and forth, there is still a small bit of information in Motif itself, especially when Motif uses its convenience widget calls to try to by-pass some of the toolkit hooks. These are the three pieces. Essentially, at this point DACX has worked to get the hooks into X11R6. DACX is currently trying to develop RAP and needs to yet look at the Motif work, which Earl can speak about in just a second. I have a handout, which is a print copy of a paper which Will published on the x-agent@x.org mailing list. The title of this paper is, "An Ice Rendezvous Mechanism for X Windows Systems Client." In this paper Will proposed three different scenarios to try and solve the biggest problem at this point in time with the RAP protocol, and that problem has to do with how does the special agent initiate communication with the client. Will is looking for feedback. This is something that has come through the X Consortium as being an approved project, something that has the approval of the X Consortium to be worked on. Will has taken the lead on it for DACX. From reviewing the mailing list, in the process of discussing this topic, a fourth scenario has come along, which is to use some stuff that was already available and there has been some good discussion on the internet back and forth about some of the pros and cons related to this fourth approach. So if you are more interested, I suggest you get on that email address and contribute.
Earl: Why don't you tell them what it is again.
Mark: x-agent@x.org
Peter: I have a feeling that we may have gotten too heavily into details, instead of talking about some of the intermediate steps in between.
Gerhard: I would be very interested in hearing an explanation for some of the stuff between the spring and now and why RAP was not part of the X11R6 distribution.
Peter: We need to sort of look back at the whole problem of screen review. You have all of these applications that are running on some machine, somewhere, and you have your workstation with your X server that is serving its display, serving its bit maps to whatever other programs want to attach to it, all on to its screen. Each application gets one or more windows to draw into. The first problem is of course, intercepting the text that is getting drawn finding out what it is. There are several approaches to do this all of them have drawbacks. The first hooks that got into X11R6, is a hook right at the front door of the X server that allows the screen reader to get in there and get a copy of every piece of protocol going across the wire or from one process of the computer to the X server itself. This solves one of the critical problems at getting at the protocol. With that done, then there is semantic information that is hidden in the protocol that needs to be retrieved. A button on the screen looks to the user, for example, like an oval with some text in it and some shading. How it is drawn is just that way, an oval with some text in it and some shading. It is difficult to know whether this oval with some text in it, with shading, is the button, or happens to be part of a picture in a paint program, that is not in fact a button and is not in fact something you can click on. So the next step is, on top of the basic X protocol, you have a library X Lib. This library is there to allow easy creation of buttons and other widgets in the windowing systems so that any given application doesn't have to come up with its own way of drawing a button. Most of them use X Lib. So the next set of hooks that also got into this release of X that came out a few months ago, is hooks that allow another layer to get at all of the semantic information, buttons, check boxes, radio boxes, etc. However, this next layer that is getting to this information must be the same address space as X Lib, it must be the same process as using X Lib. This same process is very often but not necessarily Motif. Motif is thankfully becoming the major standard for doing graphical user interfaces on X. So this is yet another library because the X Lib library is deemed not rich enough for doing user interfaces. So one of the big pieces that is missing, are some extensions to Motif, so that Motif can go to X Lib, get all the information that X Lib will percolate upwards and give it to another program running in a different address space, our screen reader, so that our screen reader can know that this text with this oval around it happens to be a button. So the two pieces, modifications to Motif and a protocol to allow communications between this modified Motif and this screen reader. This protocol is called RAP, or the protocol that DACX is developing, and is headed by Will Walker. RAP is, of course, sitting on top of yet another protocol called ICE. And ICE is a network neutral way of transmitting information from X programs to other X programs. It doesn't necessarily assume you have TCPIP or any other particular network protocol running underneath and that is what all of these terms are and that is what is going on here. So in order to develop a screen reader that gives you access to something other than X term, that actually works with the programs that are out there today like Frame, like these graphical debuggers, you'll need access to the semantic information. What is a button, what is a check box, what is the state of the check box, is it checked, it happens to have two line drawn through it, does that mean it is checked or not or maybe it means it is unavailable so that is what the work in DACX is all about. Sort of ground work for companies like mine, efforts like GUIB to come in and do a screen reader.
Question: What do you think that screen reader is going be. The reason I ask that is what specific features are going to included. When I ask it to write Braille output, it is not something that they really talked about. Certainly a speech and Braille approach is going to be necessary. I think the assumption can be made that you are going to have dual forms working together on Braille and speech. And that with those two dual forms, you probably are going to get in some text a very different kind of information out of it. There is some changing, happened with OS/2, certainly to make use of Braille. I certainly would like to see a dual approach to access to X Window being built, to get on the right track.
Peter: All of the groundwork that DACX is doing is in no way bias toward or against speech or Braille. This cake has so many layers, its sometimes hard to understand. All of the DACX work thus far is getting the information, presentation is completely left up to the screen reader. Berkeley Systems takes it one level further. We abstract our screen reader into two pieces, a toolkit that gets the information of any GUI way across multiple platforms, and the presentations to the user. Our screen reader is speech biased, absolutely. Because speech is much cheaper than refreshable Braille and many more people can afford it and that market is bigger, therefore, speech is the first thing most vendors have done, speech first and Braille second. The market doesn't have the money to support Braille, as much as it has to support speech. I will assure you that everything that is being done here in no way bias toward or against speech or Braille.
Paul: Another thing that is important to understand too is the philosophy under which DACX was created, was that DACX was not going to create a screen reader. The work with DACX is to put all of the framework in place in the X Windows environment, so that a screen reader developer could go off and develop their own proprietary screen readers using all the hooks and protocols and everything else that has been developed by DACX. In fact, not only has speech and Braille been considered, but the work that Beth Mynatt is doing in her screen reader, she is also doing audio work, trying to incorporate audio representation of where icons are on the screen and things like that.
Peter: Types of things like when you are in an edit text box sounds one way to the user, and another type of meaning since it sounds different when you hear text.
Jeff: She uses a concept which brings to those of us who don't see, the ability to translate the pictures quickly much in the same way that somebody sighted, sees the garbage can and translates that, I can relate to.
Paul: The idea is that, it is not only the speech and Braille that it is designed to work with, but work with any kind of output modality that is developed.
Mark: Also in answer to Gerhard's question, a couple of things I would like to pass on tonight have to do with some of the delays that have been holding RAP development up. DACX has always done work through the X Consortium, since we do not wish to develop something that is not acceptable. Also, several other things have been very high on the X Consortium's list of priorities, such as the CDE/COSE effort. That has put the RAP effort way back as far as priority, unfortunately. I am not saying that is an excuse, it is just reality at this point.
Question: What is that?
Mark: CD is the "Common" or some say the "Collaborative Desktop Environment" and COSE is the group of UNIX related companies trying to develop a common user interface.
Earl: There are a couple of other things that are higher on the list also for the X Consortium as well.
Mark: Will assures me that they are now on target to do a large development effort to push RAP forward, and the goal is that he is going to spearhead this effort and he is going to work with Keith Edwards, I believe, at Georgia Tech, and they are targeting to have the RAP protocol completed by March or April 1995, to make the next drop date for X11R7.
Joe: Does that mean that by March 1995, Peter will have enough information in X to write a screen reader?
Mark: If the corresponding work that needs to be done in Motif goes along with that, I would say that Peter probably has a pretty good shot at it.
Jim: As an interested integrator, and not one who is out trying to stuff my own pocket. I can honestly say that every manufacturer software developer has a better chance of convincing their management, if they find out what is the real market out there, how many users are clearly identified. If you look at the US Federal Government, you can clearly identify the number of blind people in the IRS and that is all. You cannot identify any quantity in the VA, in the Dept. of Defense, in Agriculture, it is unknown numbers. What about the universities and the state agencies. Nobody knows, and people like Peter and Beth Mynatt and all those other screen reader developers, they need statistics to say, what is real. They are not doing this for free, and if all of a sudden you come up and say I have 17 people in this organization and another one says I have 12 and another says I only have 3. If you add them up, if you can come up with statistics to justify expenditures to management and make them say, hey
Joe: I don't want to take this off track but the Federal Government spends dollars every year investigating many things. I would think that some of the screen reader companies might want to investigate obtaining some of this federal money and do just such research. Obviously, it would be a terrible price to pay if people don't get jobs or they lose jobs.
Jeff: Unfortunately the problem with us, DACX, trying to be able to identify some of those markets is self-identification.
Jim: The privacy act says, we can't ask you to tell us, you have to volunteer it.
Jeff: Unless you are willing to volunteer?
Peter: That is really one of the more useful things that the none technical members who are here, or part of DACX, or aware of this effort, can really do to help and that is to help come up to solid numbers. But just getting a sense of the market, research on how many people are there? Is there a market of 50 a month, 100 a month, 10 a month? What is the market for this product? I know technically, I can do it. Berkeley Systems did a prototype of X Windows, we didn't have the X11R6 hooks, so we weren't getting widgets that were outside of X Lib, but we had a product that was working with X terms, speech synthesizers, etc. Not a product, a prototype. We know that we can do it. The question is one of manpower and resources. Is this a viable enough market to hire people to do?
Joe: Also, you may find that if you develop an environment that is more accessible than MS Windows, you may find that your market grows.
Peter: The Macintosh was more accessible than Windows for quite a long time, and we didn't see that.
Comment: But it wasn't more accessible than DOS. We recognize that there are a lot of people, that their needs are not the same, but there is a lot of overlap, it's a huge market.
Joe: That is a really good point. A lot of screen reader developers are taking subsets of their system and marketing them to the LD market.
Paul: It makes it one level more blunt than what Peter was saying and that is if people in this room write a letter to Peter's, Earl's, Will's, and Jeff's boss, and say X Windows development work is important to us because I represent the MA Commission, or whatever, and I know 50 people that would buy this product. You are not committing to buying.
Jim: As it stands right now, there are companies like ours that, we are not a software development company, but we need these services for our own internal use and that is why Jeff is saying include his management.
Comment: Is the 508 law helpful at all?.
Jeff: The problem is commercial availability.
Paul: Right, it's kind-of a catch 22 situation. For the benefits of people that don't know section 508 is a rehabilitation act, says that all federal information systems have to be equally accessible by people with disabilities, but in the guidelines there is a "loophole" that says within the range of available technology. So within the X Window environment, it is easy to say technology doesn't exist to provide blind users access to the X Windows system. So, we the Federal Government, cannot demand it from the suppliers. So, if you can't demand it, then the developers are not going to develop it.
Comment: Can you put that into a position statement and make that statement to these developers?
Comment: No.
Earl: Speaking from somebody in a company and trying to sell and develop and keep an accessibility program, which is something that is common throughout the companies I think, 508 hasn't been a good sales pitch so far because sales haven't been effected. In X land, we are still selling products, and we are using system integration. Because contracts haven't been effected, then 508 doesn't mean a lot to management, there is no teeth in it because it doesn't stop sales.
Curtis: Believe me I understand marketing problems.
Joe: Just to expand on what Paul said a minute ago. I think it is a very good idea to try to expand the user base of this by looking for other disabilities of mainstream applications.
Earl: I would also extend on what Paul is saying. You guys can help us do our jobs. Make sure that we keep our enabling technology projects or accessibility projects or whatever it is you want to call them, whatever they are in the companies, by identifying to us, a market for people not just people with visual disabilities, but with various disabilities, so management can see that there are users out there and that there is a market out there and there is a reason for providing the infrastructure or making sure that adaptive technology is bringing technologies to our platform.
Curtis: But you know I have to say...from you guys,...that finding blind people so we can in effect can get it going.
Jeff: Why not put an article in the Braille Monitor to have a letter to be written.
Curtis: I have absolutely no problem with doing that and I have a little influence over the Monitor. Somebody from ACB could do the Braille Forum, but I sure be glad to try to push something if I could work with somebody. I think I am on the DACX email list. If I could work with somebody to try to put something together and then to get it out. I think the issue is important to get marketing numbers, that we ought to try to at least try to get groups of people to come up a to a line, and say yes there are "x" number of us people and we need to have access to X Windows.
Peter: Let's do that. Draft the half page of whatever we are saying and list five addresses, 20 addresses.
Earl: Let's make it an action item. Just going back to the screen magnification thing. The companies are willing to go out there and get feedback from the users. I am not certain that we necessarily thought of it from the point of view of getting us some marketing numbers.
Curtis: One thing that ought to be done, I suppose is what we kind of have to begin to think about doing as a spur, because once you get the hooks in place and you have to have the commercial development done, and as a spur to get them, we now have to reach all the state agencies for the blind, query all the organizations, go into the mainstream of the blind community.
Comment: Actually the only people surviving in UNIX are those who can hide from the by graphical user interfaces.
Curtis: Or they are hiring a reader full time.
Comment: That's not practical.
Curtis: I agree.
Jim: I want to emphasize again you can't go to the state agency. You have to get to the end user to get numbers.
Curtis: You have to go to the end user but you can also go to the state agency in one sense. They can't give you names. That is absolutely true. They can give you rough numbers. The state agencies across country are beginning to have technology specialists. I would bet you are running into this very problem. They may say I have 20 clients or I have 50 or whatever number they can come up with and you add them all up and hopefully you get somebody that is big enough to make it worthwhile.
Jeff: The thing to be able to clarify here is also the fact, to be able to clarify the difference between MS Windows, the MAC, and X Windows. We have to make sure that the population we are looking at is strictly X Windows.
Curtis: I understand that.
Jeff: And that is something that is going to be a tough sell. It is a tough sell for any state agency to do. It is going to be a tough sell for us and we are all in a general market to get a feel for it.
Earl: There are on-line aliases we can tap into a little bit. That is one number instead of.
Paul: Sounds to me like this is a good time for a task group, that can go up and do maybe some kind of systematic market study. Tapping the state agency. Maybe just get a group of interested people that can get together and figure out what's the best way to come up with some kind of a number.
Joe: Blind people are much more savvy now than they were years ago and they know it isn't accessible. I could call some people and say I just brought a PC and I need speech for it. They know it's there, they aren't calling up anymore and saying I think I am going to buy a computer, do they talk? They say I just brought a MAC I need a copy of OutSpoken. On the other hand, someone isn't going to even bother with X Window now because they know it doesn't work. So it is an automatic closing of the door.
Curtis: For example now, if some of this is part of the market effort. You can go and get some kind of a crude marketing study but that is another way to do it. I think you have to put all these things together. You have to do that kind of stuff to get the numbers.
Joe: The thing about programs like mine around the country is that, these are VR programs so that the two types of populations they target on are #1 job people, and #2 students. How many millions and millions of X Window machines are there in universities? I can see if this was available, at least half if not more of my consumers are students, so I would buy it. I have an advisory committee that recommends purchases, but I am basically in charge of purchasing. I could tell you right now I would by this if my clients wanted it in a heartbeat.
George: The universities may be a very interesting approach because, we deal with this because publishers are not required to make their books accessible, but the university is. The university is also required to make their computing systems accessible, and if it is feasible that the laws is going to support. You can't ask the university to do something that is impossible, but if a screen reader exists for UNIX, then the schools are going to be required to purchase it. Maybe the real valid method of determining something is through the universities, in this to be going to the head which is the Disabled Students Service Organization. Finding out how many screen readers they would be expected to come up with.
Comment: Others, Project EASI, Norm Coombs group. You might want to send him some email on this.
Mark: I think we need to create a little task group. There is a DACX alias for screen reader.
Paul: Why don't we get some volunteers why we are here.
Mark: Would somebody like to volunteer to at least creating a half page of information and getting that disseminated. Beyond that, the group would brainstorm a lot of ideas and pull some of them together and have some parallel efforts.
Curtis: I would be glad to participate actively.
Jeff: I will take the lead.
Phillip: I was just going to let you know that universities that have engineering programs like Purdue. The university is spinning its wheels trying to figure out how to get the students out of that SUN Digital lab and replace that computer with something DOS oriented or ? We are standing right in the spot, where if it is going to happen, how are the electrical engineers going to get through the curriculum without using X Windows?
Tom: I'd like to comment in some way for European involvement and hopefully the rest of the world involvement, be encouraged. Inevitably sitting where we are with this discussion, I am not kidding about that, I recognize reality, but there are 350,000,000 members of the European, Community, more than the population of the United States. I think that European countries organization needs to get information like this. So please, those of you who are involved in this marketing effort, keep that in mind.
Mark: Jeff volunteered to take over the lead on this. Curtis wants to be an active member.
Jeff: What I committed to do, is to work with the individuals.
Mark: Who else wants to be involved. Curtis, Phillip, Jeff, George, Joe,. ....
Joe: Can I ask a quick technical question. I am using character based UNIX. Can I think of that of that as DOS, X Windows-Motif that is the GUI that sits on top, like Windows sits on top of DOS? Is that correct.
Comment: Yes.
Joe: So you guys are talking about building a screen reader for the GUI portion. What happens when I do Alt FX, or the equivalent, and go back to DOS.
Peter: In general, you much more likely will see windowed DOS sessions like X term, and those will be accessible by your screen reader and you will almost never go back to the "DOS" prompt.
Joe: And that is generally accessible?
Peter: Not if you are at the console of the workstation. Some workstations, have there own character driver that is not going through X, drawing in multiple fonts potentially on the screen. Now, if I turn on my SUN workstation at home, it goes straight in X. It goes to its DOS bootup sequence, its UNIX/DOS if you will, then it launches into X and I am given the login prompt and already many programs, including the screen reader could have been launched.
Question: Regarding how we got to DOS?
Peter: No, he was using DOS to Windows as the analogy UNIX to X. So that is where DOS was coming in as an analogy for UNIX.
Joe: So if a person wanted to get access to character base UNIX, we would still be in the same boat that we are in now. Basically, you would use some other kind of PC to attach to it as a terminal.
Peter: Right, if you want character base access to UNIX, use a DOS screen reader, or a windows screen reader and a terminal program.
Peter: The point is, it is likely that the screen reader for the UNIX monitor for individual workstations will not be written and its not such a bad thing.
Joe: Why do you say that. Would that be hard to write?
Peter: The trouble is there isn't a standard there. You would have to write one for DEC, one for SUN, etc.
Joe: I heard that years ago, I didn't know if that was still true.
Phillip: Why not do refreshable Braille display? I talked with most people, somebody has just got to write the driver.
Peter: The issue is you will need to modify. Earl's company will need to modify the SUN boot ROMs of every SUN so that they echo everything they do out in serial port as they boot up, so that can then be read by Braille display.
Earl: There is somebody that is doing some prototype investigative work in being able to use refreshable Braille, and that is Sue Hearten from the NSA, and she has had a certain amount of success.
Joe: That sounds so simple. Have a special set of chips.
Earl: Well, the part of the challenge is getting the refreshable Braille manufacturers to kind of provide some sort of API to send the right stuff so you can develop a serial port driver.
Peter: Just another answer to your question, about why Braille is taking so long and speech synthesizers are so popular. It is very obvious how you send speech to a speech synthesizer. I send it text over the serial port, but a lot of Braille displays have their own weird protocols where you are sending it six dot none ASCII text, etc., etc. so you need a special driver and each company that makes refreshable Braille displays have different protocols from each other companies. So that magnifies the burden on the software developers of screen readers to support Braille because I don't have straight text with the odd escape sequence to change. I need to figure out their whole six dot, eight dot representation. How do I send this escape sequence. Here is my escape sequence table, look up the synthesizer, look up the escape sequence, send it.
Mark: I think we need to draw the group back together, we have one agenda topic left to cover. I'd like to ask Earl to comment on that.
Earl: SUN has been supporting Mercator which has lead to the RAP work and the hook work and has lead to the Motif toolkit investigation work. We are making an effort. We don't know how to make screen readers.
Comment: I have a very quick question. It may be a bit off the side, but I am interested in and I think a lots of blind people here in programming jobs may be very interested in this. When programming Motif, is there a special or different kind of language that you do that with? Is it some kind of Visual C?. The reason I ask that, is that is it something like Visual Basic or Power Builder where you have to like link things together?
Murali: Motif is a language and what it means is there are some calls you need to incorporate in your work within a C program.
Peter: There are two important changes. One is sort of a mental paradigm change because you are doing user response code, some sort of main event for waiting to see what happens and take action on that. The other thing is that a lot of tools that are very graphical in nature for creating user interfaces. I am sure they exist for Motif, there is a Motif dialogue box design, etc. Same with Power Builder, Resource Editor for Microsoft Windows. It is a graphical tools for graphically laying out the user interface, but what those graphical tools all boil down to is buttons.