Kafal Local Media Servers

Good morning Edgeryders :wink:

Future Makers Nepal (represented by me and @Matthias this time, surprise) decided to take part in the HackTheQuake call for projects - using one of the ideas we brainstormed about with our friends in RAN and Karkhana and making the Raspberries we brought over useful in a relevant way.

Here is the proposal – if you want to comment, contribute, or just read out of curiosity what the heck these guys are doing there in the middle of Asia (actually, for example, we are still swinging, as there was a major aftershock last night), you are invited to do so. If the project gets accepted, some input would be priceless - maybe you can help us tweak the idea to a higher level.

This week we will also publish the report we wrote for UNDP as part of the #FutureMakers Nepal project.

Ok, enough talking, voila, enjoy.

Brief description

We’d like to plant low-maintenance electronic Raspberries in Nepal’s villages. The fruits can be harvested by the communities – a Raspberry Pi is simply a low-power computer with storage space that we will fill with useful data for offline use: Wikipedia, open-source books on education and disaster-response, audiobooks, useful contacts, entertainment for kids and so on.

Data will be accessible for everyone with a basic smartphone, not requiring Internet connectivity. The content, along with the size of the storage, can be easily updated.

These digital fruits provide various improvements to well-being: instant and constant improved access to information, better knowledge of English, influx of innovative ideas around construction and farming, you name it.

Detailed project

In rural Nepal, many citizens own the tools for accessing and processing information and data: a smartphones, and enough electricity to recharge it. What’s missing is the actual information: mobile Internet can’t provide anything but the most essential pieces, since it’s slow, intermittent and spotty (or completely absent).

We think that the missing ingredient in this scenario are cheap, low-power, zero-maintenance media servers, providing huge offline datasets to wi-fi equipped smartphones, tablets and other computers. Such a server costs less than 80 USD and provides the equivalent to a library with thousands of books.

The exact content of the servers would be decided together with the villages who want such a server, and can always be extended later. However, here are our initial ideas of what to include:

  1. Basic instruction sheets and videos. Covering hygiene, nutrition, disaster preparedness, risk reduction, disaster response and basic medicine.
  2. Audiobooks. For people with limited literacy, audio is more appropriate content. Raspberry Pi servers allow audio streaming and downloads over wi-fi, and smartphones can be used to access both. for both. Apps for this would also be provided on the server.
  3. Offline Wikipedia. All Wikipedia articles in English and Nepali. The storage space need is around 10 GiB for a compressed archive without images, and more when including selected, downscaled images.
  4. E-Library. An open source library with e-books about agriculture, sustainable construction, medical help, handicrafts, technology maintenance and repair and various appropriate technology topics.
  5. Educational materials. Including e-books for school and “fun stuff” like novels, instructional books etc. made specially for children. Also including language courses with textual and audio content and ideally a special app to support learning.
  6. Yellow Pages. A list of important contacts and places – such as hospitals, schools, municipal offices, with phone numbers and addresses and a search to find the nearest ones.
  7. Maps of Nepal. Integrating an OpenStreetMap website incl. tile server, OpenStreetMap offline data for download and use in OSMAnd~, and several PDF maps.
  8. Android apps. Repository of free software Android applications. Basically an offline version of f-droid.org. Allows extended use of smartphones without Internet access.

The maintenance required for the local media servers is updating its datasets every once in a while. Developing an efficient system for this will require a bit of custom development effort. In total, this and other “special features” of the local media servers that need some development effort are:

  1. Internet-to-thumbdrive full downloads. To make obtaining the initial dataset very comfortable, all datasets should be available for combineddownload from a single place on the Internet. Legally this is not a problem, since only open source / open content datasets will be used.
  2. Internet-to-thumbdrive updates. To update a USB thumbdrive to the latest version of data (for taking it to villages and updating data there later), one would plug it into a computer and run Windows-based software from the thumbdrive that updates its data from Internet download sources. Even in places with (some) Internet access, updating gigabytes of data is slow and potentially expensive, so this process should be as bandwidth-efficient as possible. This is possible because most (>98%) of the data does not change between revisions. So instead of downloading everything, only so-called patches that contain the changes have to be downloaded.
  3. Thumbdrive-to-thumbdrive updates. To update the datasets, somebody would bring a USB thumbdrive with the latest datasets, plug it in and use a management function to let the server copy it to its own USB thumbdrive. Only newer datasets will be copied.
  4. Usage statistics. In a later version, each local media server would collect and store anonymous usage information. Usage information will show the degree of success of the village intranet project, and what content is the most useful, to inform future extension. To get this information relayed back to the project’s website, it will be automatically put on a thumbdrive when updates are installed from it, and when that thumbdrive is synced with new datasets over the Internet, the usage information will automatically be posted to the website.
  5. Edit and publishing options. In a later version, users would also be able to create and edit some own content. Typing is not that fast or comfortable on smartphones, but users can for example contribute recorded audio content with stories or instructions. Such content would be posted to the project website the same way that usage statistics are.

For the server hardware, we propose the following setup:


Raspberry Pi 2. Low-cost powerful single-board computer that can run Linux server software.

36 USD


Raspberry case. Lasercut plastic case.



Wi-Fi module. Normal wi-fi USB stick with access point mode.

14 USD


USB charger. WIth MicroUSB port. 220 V AC or 12 V DC input.

5 USD.


64 GiB USB thumbdrive. Or starting with a 32 GiB version for half the price. Up to two more storage devices can be added to the Raspberry Pi’s other USB ports later.

32 USD


(Photovoltaics power source incl. phone chargers.) Optional, but required in villages with very limited electricity. Smartphone charger cables should be long enough to work comfortably while the phone is charging. A 30-40 W panel, charge controller, small lead-acid battery is all that is needed.

(50 USD)


One set of archives that would be useful is the CD3WD project.

It was an archive of manuals in text and pdf format that was created by a professor at Harare university. While the agricultural texts are more relevant to the central African climate, the engineering texts will apply anywhere.

While the original website is offline, you can find back-ups on Appropedia, as well as elsewhere.

There’s a copy on the Internet Archive, here.

I’ll try to find a functioning torrent.

update: There’s a lot of dead links out there. one live set here.

CD3WD: Great (re-)find

Thank you! I once found this when researching free literature databases for EarthOS, and it’s really great (>40 GiB of material). Their original website was ec3wd.com (or as you found out cd3wd.com), but as you said, these are dead now. Great you found the stuff again!

Given relaxed copyright handling here in Nepal, I guess we could simply pass over the fact that not all of the e-books in there are technically open content :wink:

I spoke to Alex Weir about this project nearly 15 years ago.

It was after my suggestion that he started hosting it on Pirate Bay and a number of other torrent sites.

This was another example of a project like @hexayurt’s, where it was a simple technological solution, that the aid agencies refused to touch, as they felt it would be too effective. “It would mean they wouldn’t need us!” is a direct quote from an email that he forwarded to me.

This was also the reason that he was refused permission to distribute any of their documentation. So he stored them on a server that was “mistakenly” made publically accessible.

Since you’re over there under the UN umbrella, make them available, but don’t make much noise about it. :smiley:

1 Like

some links

Text only wikipedia http://en.wikipedia.org/wiki/WikiReader

The final edition http://imaginethecanaries.blogspot.com.es/2011/08/sporulation-early-draft.html?m=1 http://imaginethecanaries.blogspot.com.es/2012/04/sporulation-take-2.html?m=1

Wireless networks developing world http://wndw.net/

1 Like

70 GB of oldish homesteading literature:


Long Now foundation has a couple of additional good leads (not all copyleft):

For audiobooks:

Librivox.org !!!

There’s a chance the makers of instructables.com will give you some extra help if you frame it the right way. They originally wanted to center their effort on development aid. The sure won’t mind to give you privileged access to things that could help you. They often run themed competitions - perhaps that would be a way to leverage the community there?

I would very much like to give more input on building an audio-library, as I think that will be the most appropriate in many situations. Currently there is only a very limited amount of material available that could be used without adaptation. This can be seen as an opportunity as well though, as it is not very complicated to produce content - like you already mentioned. As a healthy system should have information flowing in various directions, one could seed** this library and listen to what they want/need more of. Then this could be supplied in a lean format (e.g. text + images) and slowly and collaboratively turned into a more accessible format. Basically producing podcasts*. I have made some back-of-envelope calcs on this already a while ago and my take is that this would be very promising in the long term.

Speaking of long term, I had envisioned a system that would work with extremely long lived electronics (caps/batteries/IO ports would eventually be separately replaceable). The packaging of the electronics would be critical if you want a lifetime of say 50-100 years out of them. I recommend casting them in high quality silicone and plugging extenders into the ports with very sturdy strain relief (e.g. paracord sheath along the outer 4/5th). You may want to consider using clean sand to help with heat conduction. Also it won’t hurt to put this into a wooden box - again with strain relief for any wires going out. I would strongly recommend to stock up on spare parts nonetheless.

For each each library server my idea would be to have 3 working and 1 dead used mp3 players, with about 20 partially working earplugs, some solder, perhaps acoustic membrane film, and a piece of metal that can be used as a soldering tip when heated in coals, and of course an audio guide to fixing earplugs and the most likely defects of audio players. The idea being that the audio players need very, very little electricity and they could be shared between three kids each - promoting social cohesion. I know this will probably go over budget, but If you think it might fly - one could try to expand the budget. It would be relatively simple to argue for this on a number of different grounds.

*The problem with producing podcasts “in the wild” as opposed to “in the basement” will likely be background noises that you can’t run away from, or shut out effectively. Also protecting the more sensitive microphone membrane (you may not be talking mobile phone level) can be a small issue in some environments. I’ve already researched this a little, and could give more leads.

**For seeding I think one should not “just dump in” but talk things through with the locals of course - keeping in mind that there may be a bit of a “blind man & elephant” issue playing into this. Personally I would tend to focus on things that have some entertainment value as well as content that addresses unknown unkowns, in a way that is really helpful in everyday life and perhaps also has a short payback time.

1 Like

Thanks @trythis

Some nice links.

For producing your own media, have a look at the templates created by StreetReporter.

Simple and effective templates which with a little practise, allows the creation of broadcast quality material.

Howtos always welcome!

I’ll sure check that out. However I am more interested in formats that are suited to dig a little deeper (lecture) or are more action oriented (from scratch instructions). I am not a friend of “regular news” format - even for news. Perhaps adaptations fusing Wikipedia articles with spoken instructables is what I’m trying to get. I had some nice solar concentrator material at some point - but I can’t find it just now.

Still, one does not exclude the other.

Another thing is libraries for the blind. The German one is pretty well established, but there must be equivalents from which to draw for Nepal. Those however will probably only be useful for a niche application again - they’ll not cover their “bread & butter” needs I’m afraid.

On audio recording in the wild and involving other communities

From my musician days: most devices like phones, videocameras and laptops use condenser microphones. If you use dynamic microphones (cheaper, tougher etc.) you get the added bonus of a much shorter range (explanation, based on video production rather than quake relief, but the engineering principles are the same).  Shorter range means a dramatic reduction in ambient noise. Sound sources that are not close to the microphone and directly facing its membrane will not be picked up by a dynamic microphone. This is why the (hi-res) 2000 EUR microphones are only ever used in soundproof recording studios, and even major stars sing into 100 EUR Shure SM58s.

@trythis: you seem to have an amazing network, and you are very generous with sharing your contacts, for which you have my personal thanks and, I am sure, those of the community at large. The reason why we may seem hesitant to follow up on your advice and reach out to the people you are pointing us in the direction of (just in this thread: the delay tolerant networks crowd and Europe Aid) is, as far as I am concerned, the following: these are high quality connections, and need a lot of maintenance. It’s not OK to shoot an email in their direction and then forget about it; if you engage, you have to be ready to follow up, again and again, and to show proper respect for the people involved and their project. If those people get really wired up and start writing you ten email a day, you have to reply to ten emails a day, until mutual trust has had time to harden. Doing otherwise would be irresponsible, as it would deplete your social capital. Therefore, engaging properly entails a significant coordination cost; it needs to be done wisely, in full awareness that a team like Nepal’s can probably handle only one newly opened such connection every month or so.

This is not to discourage you. On the contrary, I repeat, I think you are doing great work, and doing it very generously. I am simply managing expectations here: it appears that you can share great contacts much faster than we can take advantage of them!  smiley

1 Like


I am in general agreement with you here. But I also want to clarify that the Europe Aid connection is one more degree away (a foaf = friend of a friend). However I was invited to submit something by my friend (who has no stake in this) so I thought there is an above average chance that we’d at least get listened to. Also there was a fancy (and inappropriate) version of something similar that won some sort of prize in a contest for these sort of things (I’ll have to look that up again at some point).

The part that this is usually a substantial commitment*, and raising expectations that you cannot follow up on is not a good idea in networked collaboration (not marketing!) is very true. Sometimes there is an opportunity to really “ask a question for free” - but this is not often the case, and will usually depend on what status you have in the field. Most of the time these opportunities can’t really be forced - so you may just have to sit and wait for such an opportunity. Geres.eu was such a thing. The e-library probably won’t be. However, if I put my “potential contacts” on the table, the situation may change for another person - because those may have been the missing pieces in his puzzle. But this is something that usually takes some time and some very fit communicators to pull off. And, like you point out, you don’t want to do everything at the same time anyway for lack of capacity. One point though is that France as a whole usually takes the month of August off - I assume this is also true for Geres to some degree at least.

Your point about the microphone remind me that I had friends doing research for Sennheiser. I haven’t talked to them in a while but I think one was a member of Amnesty International as well. So that is on the table. I have another good friend who is very DIY, electronics, and very audio, but he is usually pretty busy and has a stronger Africa connection than Asia - but who knows.

*My usual rule of thumb is approx. 100 emails of 10 minutes, or 16 hours, or 2-3 full working/meeting days if you are working on eye level will be a realistic but lowish investment to bring a contact on board/up to speed. Of course that excludes all the prep work that goes into in having something to talk about. And you have to consider that the cost on the other side is often considerably higher, especially if you are talking about getting real money on the table. So if you’re fighting up-hill (you first have to convince each person of the value of the undertaking, not just the sustainability, or even your capability) I will generally be very skeptical about pursuing such a thing. If there is genuine shared interest it is something else.

Perhaps interesting for you

@pamela with the caveat of the post above, this thread might also be interesting for you. At least that is what I thought after watching the dadamac intro video.

We understand each other

Ok, glad that’s out of the way. :slight_smile:

I edited my previous comment to make the point about microphones clearer.

Honestly - I was all the way down to throat mics

They have many drawbacks - but you sure get rid of background noise. They would also work in an urban slum under a busy highway. Also they probably get rid of any pop-sound-issues you may have. On the other hand I am far from sure if all languages work equally well with them. Lastly, in the context of hacking for post-quake-Nepal specifically, the issue is pretty small.

More smaller memory cards?

Would perhaps reduce the risk of catastrophic failure modes?

Also if you’re talking resilience it would make sense to duplicate the data on a second drive. That could take the role of dissemination AND backup at the same time…

Academic research into delay tolerant networks

There is quite a bit of research on the general theme.

I could imagine they would be interested in working together to get real-life data for their theoretical architectures. This is something that won’t get around doing the empirical part at some point - and it’d be too expensive to run this in the expensive industrial nations in many cases. It also would not be terribly appropriate in the first place…

1 Like

No contacts

I don’t think I have any direct/1st degree contacts to the ones mentioned in the wiki article. Does someone? Perhaps @Pietro ?

I would probably spend more on the charger or buy several

depending on their typical failure mode (use time vs aging vs max load). I have some about that price range that don’t look terribly reliable based on the board layout alone. I have not gotten around to hooking them onto my PV panel a lot yet, so I don’t know for sure.

Also I would strongly recommend to go big on the “smart phone dock” and turn it into a mini-workstation. Otherwise you’ll quickly have a lot of wear&tear on wire and smartphone. You really want to avoid that!

1 Like

Very inspiring

Hey @trythis, @LucasG and @Billy_Smith, I must say this is one of my favourite discussions here on edgeryders.eu in a long time smiley

Should we win a prize in the hackathon, we will definitely (have to) move ahead with this project, with some support from friends here in Kathmandu. Otherwise it’s a bit much for the moment, though that is a pity because of all the content links and ideas in this thread. So of course if anyone wants to “steal” the idea and just do it, much welcome …

I just had an idea that would bring these e-library village Intranet servers closer to the current project, but I’m not sure how workable this is. Essentially we have still the problem of how to provide a virtual dialogue platform in villages, since there is no Internet to connect them. So maybe these servers can serve as a delay-tolerant network (as trythis calls it) for collecting and distributing long-form audio and video stories from villagers, recorded on smartphones? If that is not practical, then at least digital versions of the village newspapers that exist in many villages here.

Then, whenever somebody visits another village, they would charge their USB thumbdrive with the current village’s latest storyset, and plug the thumbdrive in at the target location, which will push the updates, and pull the stories from that village. Audio stories could then also make their way to FM radio stations in that and other villages, to be cut and broadcasted …


Draft proposal - with local input?

Honestly, I am tempted to try to hammer some of this into a draft proposal for Europe Aid or some such agency. I am relatively green in that field but I’d have a couple of contacts I may be able to bring in/lend an ear. And like I said, especially the audio part has been a pet peeve of mine.

One could perhaps ask for a little seed money to help the Nepal team support us in making generic architecture (which would of course need to be locally adapted). I would think a very basic regional cluster covering 3-6 villages, or a more focused approach supporting a more specialized infrastructure project (which needs skilled staff for e.g. maintenance), or a health focused version would be obvious applications. If there is enough agri/eco info available one can go deeper into nutrition and sustainable agriculture for different biomes/biogeosystems.

As one would focus on the hardware, maintenance & info flow architecture the cost would not be very high - especially once the wrinkles get worked out and the stuff is properly ruggedized. Then one would transition to the part where it becomes self accelerating - which is a real nice prospect. Combine this with language and culture conservation aims and it gets hard to turn down this offer. Still, the one thing that keeps me up at night is the risk that this could be subverted to less than humanitarian goals. This can be turned around though: as long as we don’t have long lifetime electronics - we have an automatic containment system (important if you start doing experiments you don’t really fully understand).

1 Like

Search red cross/crescent

If you search their site for .pdf you get a decent concentration of pretty useful documents as well. E.g. http://www.ifrc.org/PageFiles/95882/C.02.08.%20Field%20Operarions%20Guide%20for%20Disaster%20Assessment%20and%20Response_USAID.pdf from here. Oftentimes they’ll be too specialized to be of interest to the average subsistence farmer family. But there are times when some of them could become extremely useful to have available in a timely fashion.

And I guess it would not hurt if translations/audio versions could be done into large language groups. These could be submitted back for editing before they go public all the way. Here I am not so much thinking of the people who are busy rebuilding Nepal at the moment. I am more thinking of people sitting in refugee camps for the next 70 years with no opportunity of regular economic activity…