Open formats in AAC: Part 4, what I want.

This is the fourth part in a series of posts on the subject of open formats in Augmentative and Alternative Communication (AAC). You can read the first part here, the second part here, and the third part here. You’ll be pleased to know that (I think) this is the final part.

I’m a big fan of open data (as you should be able to tell from The Domesday Dataset), but I’m also a big fan of all manner of different manifestations of open information. I use Firefox for browsing and I write my code in Vim or occasionally Eclipse. I ask lots of questions on Stack Overflow and I poke at Wikipedia when I see a typo. So where is my open-source AAC solution?

Why do I not have a tablet and framework that loads AAC setups in some common format. Come to think of it, why do I not have a common format? At the moment, a man might be using an AAC device for four years, it’s got all his common phrases on it – stories from his childhood, personal data of all types. It breaks – he’s got a backup, but they don’t make the device anymore. His new device doesn’t co-operate so someone has to setup each picture, each icon, each page, each phrase by hand – and that’s if they can find something to read the old data. Just a common format.

So what do I think are the next steps? How do we get to a common format? To more use of open information in AAC?

There are big bits of the vision, and there are little bits, one of the little bits is that I think that conferences like Communication Matters should have a hackathon next year and every year after that, where anyone who can write code gathers around a table with an SLT and, more importantly, a couple of users and we see what cool stuff we can build in a couple of hours.  It’s pretty surprising what can be put together by a half dozen motivated people in an afternoon…


Now let’s look at the serious parts of the vision. Clearly we can write little bits of software that translate between formats, (I demonstrated some at Communication Matters this year), but if we have to write a translator between any two formats that’s going to get old fast… If there are over dozen manufactures, then to fill in all the blanks then I have to write 150 different programs and that’s a bit of workload, it would be worth it…

…it’s also trivially silly. Of course, you are all thinking: why don’t we just agree a common format? Because that buys us a lot. Instead of 150 different bits of code, we write one piece of code that goes between the common format and your favorite format of choice, and suddenly transfering between any two machines is easy.

…and if all that is working, then not only do we get the ability to rescue a wordlist from an out-of-date device, not only can users wander around the stands at conferences and exhibitions with a USB stick and discover the devices that *really do* work for them, but we also get the ability for research groups to work much more closely with users. Both at the level of adding content and in terms of finding out the content that should be added.

Now I should be clear here – this isn’t about taking a virgin Dynavox setup and running on a Liberator, Companies work hard on building their initial page sets and there is a lot of intellectual property tied up with them. I’m talking about the hundreds of other phrases that arrive on a well-used device. The stories from childhood, the in-jokes, the medical information. In my little brother’s case It’s also the pages of Star Wars characters and the full listing of every game ever produced for the SEGA mega drive. That’s the stuff that sets a user apart and that’s the stuff that I’m interested in transferring.

There’s one more thing, if we’ve got a common format… then someone will get around to writing a viewer for it, someone might get around to writing an analysis tool for it, someone might get the hang of an interface enough to write an editor for it… and because those people have all been facing the right direction at the right time, suddenly we might be looking at a proper open source solution that can take its place in the ecosystem. We need to invite talented people, with specialties in programming and the various fields into our community but I think this is the structure that we can start building on, this is the structure that I think we’ll hit a critical mass on. But the open source goal is a pipe dream for now. For sure, I don’t believe it could happen without a common format. But a common format is valuable now, it’s important now, and it’s achievable now.

This isn’t a ‘me’ post, if it was a me post, I would have come up with a name for the format, for which the file extension would be “.joe”, and I’d have the SQL diagrams up for your general approval… but I don’t want this to be a me post, because this sort of thing only works at the grassroots level. The point of this series of posts was to show you three problems, it was to start a conversation about common formats, and possibly also to show some of you how easy it might be. All I want is for the people who work in AAC to have the idea of a common format on their minds.


Leave a Reply