Ken Banks is the creator of FrontlineSMS, which is used in Ushahidi as a way to allow local phone numbers to be used for incoming messages. There’s a dependency that I’m not a big fan of though – you have to know how to download it, setup and activate it on your computer. That’s a huge barrier to entry.
As Ken just posted, we were roommates at last months Pop!Tech Fellows program. We’ve known each other for years, but this gave us a chance to talk at length around certain ideas that had been frittering about in the back of our heads. One such idea was how we could get rid of the need to own a computer to run FrontlineSMS (and from my perspective, sync with Ushahidi).
An independent mobile hub
If you have someone trying to run an operation in a developing nation, you don’t always have the luxury of having a computer and/or an internet connection. What if you could run this whole system locally from a microSD card, slotted into the side of a USB GSM modem?
“The software, drivers, configuration files and databases could all be held locally on the same device, and seamlessly connect with the GSM network through the ‘built-in’ modem. This would mean the user wouldn’t need to own a computer to use it, and it would allow them to temporarily turn any machine into a messaging hub by plugging the hybrid device into any computer – running Windows, Mac OSX or Linux – in an internet cafe or elsewhere.”
Doing this would effectively remove the computer (the largest expense) from the system entirely.
That’s a very powerful idea. By taking away something, you make it more powerful and more useful to the user. It means a lot for those who are trying to remain under the radar and it means they could do their messaging effectively, and with a lot less knowledge of the system, than is currently needed.
Building a prototype
Both Ken and I would like to hack a prototype of this together, so if there’s anyone interested in helping us do this, please let us know. If there are any MAKE fans out there, this might be right up your alley. We’ll both be in Nairobi from Dec 7-12, so if anyone there is game we’d love to do it then. This can also be done remotely too, so anyone want to work on getting FrontlineSMS native to a device such as this?
[Note: this type of device isn’t just useful for activists at all, I can think of a wide variety of businesses and individuals who could use it, it’s that in our context activism plays the largest role.]
November 27, 2008 at 4:29 pm
Erik — can you describe a few compelling use cases for this? It’s a great idea but detail some real-world uses cases where this would be useful? The problem with Frontline right now is the lack of supported phones which this would solve, of course (as would open sourcing the application, finally…). But lack of scale and speed of SMS delivery would still remain, the other big barrier to Frontlines’ usefulness in on-the-ground operations. It would be useful not to let the tech run away here but actually address the key reasons why uptake of Frontline seems to be lacking with some real market-research on what’s needed on the ground. Is this something you are planning on doing?
November 27, 2008 at 7:31 pm
Gosh, I wish I had some tech-ability, this is a fabulous idea.
November 28, 2008 at 1:30 am
If the ‘tech-savvy’ system developer doesn’t have access to a computer how will the clients have access to one? Unless the clients of the system are not local. Good idea; it should give way to better idea.
November 28, 2008 at 4:20 am
this tech stuff move so fast, blink and you miss the latest dev
November 29, 2008 at 12:12 am
We’ve just begun talking about adapting your systems for India and her multiple languages and scripts. We’re planning a meeting on this with concerned members of the interaction design community here in Singapore after the 7th of Dec. Any advice, tips and information on how to deploy and use your system (as well as the eventual localization for Indian needs) would be gratefully appreciated.
November 29, 2008 at 4:57 am
I’ve received a couple of really helpful and interesting emails and Tweets about this idea, and they’re probably worth sharing more widely in case other people are interested in trying something out. Yanokwa in particular suggested trying a Huawei e166/e169/e176/e230 GSM modem, paired with SMSTools3 – if anyone has those available it would be interesting to see what they come up with. As I mentioned in my own blog post, this ‘idea’ has potential not only in the work we’re doing but “also for many others working in the same social mobile space, making rapid prototyping and the dissemination of tools much quicker and easier”. Crucially what we build will be useful for any SMS-based tool – it’s not just about Ushahidi and/or FrontlineSMS.
Chancellor’s comment – “Good idea; it should give way to better idea” – is spot on. If discussing a hybrid device such as this leads someone to develop something much better, or come up with a more appropriate solution, than that’s job done I’d say. I doubt we’ve come up with the perfect answer, but it’s do-able and a good starting point for a v1.0. It would be great to see other people build on it if we proved that the idea had legs.
And one comment for Niti – I’m sure Erik has his own thoughts on Ushahidi, but in terms of FrontlineSMS I’ve started conversations with a team of mobile guys in India who approached me and are interested in doing just what you suggest (localisation, outreach, support, etc). I have a conference call in the next week with them, so if you want to drop me an email I can do some introductions, if that sounds helpful (you have my email address).
See you in Nairobi, Erik!
December 2, 2008 at 9:59 am
@Katrin – Ken should really answer most of your questions, but I’ll give a try on a couple of the technical/usability ones.
As you said, this would solve the problem that not all phones can run the FrontlineSMS software. An important thought for me in all of this is the simplification of use for people on the ground using it – the non-techies. I’m imagining a device that negates installation and drastically improves the setup and management experience.
You’ll have to ask Ken about some of the other details, as it is his software, but I would add that the speed of delivery is normally a function of the device and the network. It’s an actual device/GSM limitation. Speed of data transfer (even though it’s small kb), handshaking, awaiting acknowledgment, etc. any software which sends through GSM devices has the same challenge. For NGOs without any alternative, even slow is better than nothing. One interesting thing here is that the device could do this through GPRS too, which makes it lightening fast if you loop through a global SMS gateway… Very interesting idea in fact!
December 2, 2008 at 10:25 am
Fully agree with your comments, Erik.
At risk of making this a conversation about FrontlineSMS and not the actual topic of the blog post (!) I’ll provide a few answers.
You’re right about the speed issues. Any software sending through a local GSM device has the same challenges, which are largely unavoidable. Connecting to a gateway via GPRS would solve it, assuming GPRS is available and the NGO has an account with a gateway (two barriers in many cases).
With supported handsets, FrontlineSMS currently supports phones with modems which communicate using standard Hayes AT commands, which is a large number (sadly the Nokia 1100 – which you find everywhere in many developing markets – is FBUS and not AT. Ask Nokia, I guess!). The list on the website are phones we’ve managed to get our hands on to test, but no-one on earth can get access to all devices. Users have been informing us when they come across phones which work or don’t work, and we’re working with some companies in the UK to get access to professional test facilities. This all takes time and funding (which we now have). It’s technically hugely challenging supporting such a fragmented device market. MIT NextLab staff have been hugely impressed by what we’ve managed to do, in particular the ability for FrontlineSMS to detect and auto-configure devices.
As for uptake, we’re quite happy with where we are. As of this morning there have been 1,039 downloads since launch of the new version in late June 2008. Again, it takes time, spare hands and funding to really start to dig into the user base, but we’re getting stories all the time which we share if we have permission, and/or if we get enough information. FrontlineSMS is work in progress, not a polished product, and we welcome anyone to the growing community interested in helping us make it faster, more accessible and more relevant.
Turning back to the hybrid device, I’m looking forward to seeing what we manage to come up with in Nairobi. My belief remains that if a quick win is possible then there’s nothing wrong with going that route. If our hunch proves wrong then I’m a GSM modem down – not the end of the world. If we’re right then things could get very interesting, and we can start to plan from there.
December 4, 2008 at 12:38 am
What about using a cheaply available smartphone such as early Nokia Series 60 phones like the 3650? They can be had for ~$40USD and can run PyS60 which is an easier-to-use-than-Symbian-C++ dev platform that can read incoming SMS, send them over GPRS to some address…and that’s the main purpose of this, right?
It even has crypto functions in case that’s important in reporting from areas with questionable speech rights. The app could be packaged into a Nokia .sis file and be very easy to install and run in the background of the phone. The other added advantage is it could be run without a computer, phone would consume much less power and could even be powered by small solar panels in some cases. Just a thought.