Poll: Interest in a usable free software friendly phone?
I've been thinking about what projects and products we could put out that might be both worthy to take on and draw in new money.
Right now I'm looking at mobile phones and trying to decide if this might be a worthwhile effort. There are free software efforts to create a mobile OS that isn't dependent on non-free software. They aren't 100% free although the main non-free pieces have been removed. There is clearly a lot of work to do yet in this area and I'm sure they could use some funding.
The one project working on a free version of Android is Replicant. There is one older model smart phone in particular the project supports (there aren't many to begin with) that I think would probably make a good candidate between the price and design. It appears the phone has the least issues from a privacy/free software perspective (would need to verify).
If we looked into it further would you spend $200 USD on a phone (excluding VAT) that excluded as many unnecessary non-free bits as currently possible? This would be a usable phone unlike developer targeted phones such as the freerunner.
Normally I would not even consider adding such a device to our catalog due to the non-free bits. However the equivalent in the PC would probably be the BIOS for the components that are non-free. These are all things that need funding to fix and will not occur without a product and demand from users.
There would be no significant work on our part beyond the typical work required to get it to a usable point. Researching freedom friendliness issues, flashing, and making it generally ready for purchase. The main advantage here is it would work out of the box and help fund a free software project.
We would have to talk to the lead developers of the Replicant project before we moved forward on anything. They would get a portion of any profits made. Probably somewhere in the equivalent ballpark Trisquel currently gets.
Here is a comparison of free phones:
According to this page, the Freerunner, N900 and Nexus S have a modem that cannot access CPU, however the last two require non-free firmware for some of the basic functions (wifi, bluetooth). The new generation Freerunner (GTA04) also requires proprietary software for wireless and bluetooth.
It would probably be a better idea to supply a free digital audio player. Something like a 5.5 generation iPod loaded with Rockbox, with Rockbox developers getting a cut. There should be far less freedom issues in this case as any non-free firmware needed in this case (e.g. to charge USB) should not be malicious unlike the GSM modem.
Since the Neo Freerunner supports booting from microSD cards, you could also sell supported cards preinstalled with popular distros like QtMoko (regardless of whether or not you distribute the phone).
I have a HTC Dream running replicant. It has some issues just so you know so you would have to be prepared to support those. I deal with them because I know that is what comes with using free software but someone who bought a phone with replicant pre-loaded may not feel the same way.
What kind of issues?
The real question is does it work well enough to utilise. Little issues may not be that big a deal. Particularly if we can document issues and clarify what features are supported, what is being worked on, and what might never get supported.
Just because every feature isn't available doesn't make it a bad move.
I think as long as the critical features work this is an incentive to buy such a phone. You know where your money is going. It's going to support the developers working on fixing your phone and develop a free software friendly ecosystem. And the main feature after all is not the normal smart phone features. It's the freedom.
I may not be developing for a smart phone, but I would certainly prefer a phone that was catering to freedom, privacy, and at least basic support.
I think the fact some/many people have these already is an indicator that there is demand. Well, we will see. I'm going to get one and see how things go.
Umm here are some of the issues I have
* Battery life (may be because I got the phone used)
* No wifi (requires non-free software)
* If I lose coverage I have to restart the phone to re-enable coverage
* Sometimes phone hangs on reboots
* over the air internet connection drops out often requiring reboot.
Don't get me wrong the phone is usable if you fight with it but some of the issues can become tedious (constantly having to reboot) and not having stable internet.
In a smart phone, any of these issues would be a deal breaker for me.
Looks like I'd have to sacrifice a little freedom if I were to start
using one of these things.
> * Battery life (may be because I got the phone used)
> * No wifi (requires non-free software)
> * If I lose coverage I have to restart the phone to re-enable coverage
> * Sometimes phone hangs on reboots
> * over the air internet connection drops out often requiring reboot.
The phones comming after are more usable(nexus one, nexus S, galaxy S, etc...), and paulk probably has improved the htcdream with a new RIL(the library that talks to the modem).
Thanks for the link. It looks like a more dire situation than I thought. I still think the idea is good though. If we didn't all already have these I might think otherwise. The problem is we do already have these and people do need to replace broken phones. The question is are they going to be supporting free software developers / a free software ecosystem or are we just going to sit and depressed about it.
I don't need a new mobile phone. So the answer is no for me.
I recently wondered about buying a Freerunner because I wanted a portable audio player. But there is a problem playing audio on the freerunner. And it also seems too weak to play videos. So, I stuck with the Unix way (one old mobile phone and a portable audio player).
But if I would change my mobile phone for a smartphone I would be glad to find a free software-friendly one (if not too dear).
I think there are probably a lot of people like yourself.
The last I used the freerunner it made a very very very poor phone. It was mainly because you couldn't hear people well enough and/or they couldn't hear you. I would not advise buying one unless you were seriously going to work on fixing issues and otherwise developing for it.
I like the idea of the freerunner. I think we need to step back though and recognise not everything is going to be 100% free. It wasn't 100% free and with more customers there could be more significant resources going into development (freeing a phone).
The Freerunner should work well as a phone if you perform all the hardware fixes (e.g. Buzz fix, Echo fix, Deep Sleep, Bass Rework). Then the phone is at least somewhat attractive (you should get at least 4-5 days of battery life for making and receiving calls). Audio configuration files that work well for calls, SIP and bluetooth should also be configured for the user (and sent to him or her, e.g. via CD or email).
That is interesting. I wish I had a working phone to try. I never did attempt the buzz fix and it was an issue at the time. The other problem with the freerunner is price/availability. As far as I'm aware you can't get this any more. There is a newer version although it's only the motherboard and the cost is astronomical. It's not a good model to generate funds for such a project. It might make more sense to do a Replicant phone and then fund (or partially fund) a project that is trying to move forward with the freerunner.
It's pretty easy to get a Freerunner. I got a non buzz-fixed one from OpenMobil.eu (see the Openmoko distributors page: http://wiki.openmoko.org/wiki/Distributors) and you should also be able to get a second hand one if you post on the mailing list or visit the irc room (I also got a second hand buzz-fixed one via irc). The phone is still being sold by the distributors, although if you want one with all the hardware fixes it will probably be very costly (if you intend to become a reseller yourself it will probably be much cheaper to simply buy non hardware fixed versions and hire someone to perform the fixes).
Even with all the fixes I don't think it would make a good solution for the fund raising objective. It is still a very developer oriented phone at a price that doesn't really work. Most people end up with $500-$1000 phones as the result of subsidies made possible through more expensive contractual services. At least this is how it is in the United States. I'm not sure of the details on the pricing or the average sales price in other markets for phones. I would bet the price most people pay for phones without a contract though is significantly less than $400-500 USD.
I don't know too many people who would spend $500-1000 on a phone if they had to pay that up front. When people are presented with a monthly price it becomes psychologically more bearable. It is also more bearable when someone starts low and goes higher.
We don't have the resources to offer a $400 USD phone for $150 with a payment plan of $30 / month.
Otherwise it might work although even in that situation it is still going to reduce the amount going toward development. There is ultimately a budget every consumers has to manage. It's very difficult to continually spend more than the amount you take in. Credit eventually gets cut off.
I was thinking about designing such a portable Free Software device, but from scratch. I'm not really sure about the OS, though Maemo looks like a nice choice, better than Replicant IMHO. The target board to use is http://rhombus-tech.net/allwinner_a10/ but it hasn't started shipping yet. I believe that such a phone can be targeted towards students, Free Software Supporters, and in my case, visually impaired people. I strongly think that a cheap computing device such as this is really needed in a world where there are people who need to do basic tasks and dont need "Apps" so much.
P.S. My goal is to keep it under $150.
The one issue with the newer phones is the PowerVR. There are many others. If we could design a phone in the $150 price range I think it would work. If you move forward with a project to design a free software friendly phone let me know.
I think some of the ideals for such a project should be:
Processor isolation; I believe this means the modem can't access the memory or control other parts of the phone. no spying in other words.
Graphics, mic, audio, wireless, and bluetooth that are not dependant on proprietary drivers/firmware. Any any other component.
The modem may have to remain closed although if we can isolate it I think we can partially work around the non-free software issue. I believe this has been discussed once before. It can be treated as a circuit or something to this effect.
The one thing that I would like to see if at all possible is a one way receiver. I'm doubtful this could be done although at one time I believe there were chips that received but didn't send over the cellular networks. The reason you would want this is because it could impede tracking. If the modem can be turned off and the one way receiver chip kept on then a system could be designed in such a way that when a call came in it would first 'page' (like old pagers) a device (a broadcast goes out to every device in a given area) and if the device detected the page it would turn the mode (once the user accepted the call) in order to receive the call. The tracking would be limited to knowing the time/place a call was received/sent as opposed to having a tracking device that could pin point you at any given time and keep tabs on you at all times.
The alternative to this might be to simply find a way to turn the modem into a receive only mode. It's probably not possible to do this without the source code and even then it may not be feasible. I'm not sure how the old one way pagers worked exactly or if that technology is still around. I believe they used the cellular networks though so I would think in theory you could turn a modern modem into a one way pager with the right sources if the one way pager technology on the other end still exists.
В 23:59 +0200 на 03.07.2012 (вт), chris [at] thinkpenguin [dot] com написа:
> or control other parts of the phone. no spying in other words.
In other words architecture like in the FreeRunner - two CPUs, the main
one running the OS and controlling the baseband/modem.
> Graphics, mic, audio, wireless, and bluetooth that are not dependant on
> proprietary drivers/firmware. Any any other component.
The Openmoko mailing lists have some recent information on that topic. I
think Golden Delicious, the company that is developing the GTA04 board
for the FreeRunner casing, had released some information about their
findings and success with chip manufacturers and suppliers.
A work-around for proprietary firmware for such chips (AFAIK supported
by FSF too) is to load the firmware only from an EEPROM/Flash IC by
"hardware" and treat it as a circuit. For example, the FreeRunner's
wireless chip firmware is loaded in such a way. I think there was
similar effort for some of the GTA04 board's chips. It is hard
(impossible?) to find integrated circuits for mobile devices with free
firmware or without firmware at all. Sad but true.
> can partially work around the non-free software issue. I believe this has
> been discussed once before. It can be treated as a circuit or something to
> this effect.
> The one thing that I would like to see if at all possible is a one way
> receiver. [...]
> The alternative to this might be to simply find a way to turn the modem into
> a receive only mode. It's probably not possible to do this without the source
> code and even then it may not be feasible. [...]
The Osmocomm project  is working on variety of free software projects
related to GSM. They have a sub-project called OsmocommBB  which aim
is to make free software firmware for baseband processors. I believe
they are still in the early stages, but some of the supported baseband
chips might be the best way to go if available on the market. The
Calypso, one of the supported chips is even used in the FreeRunner, so
theoretically such modifications can be tested before development, if
they are at all technically possible. After all the GSM towers have to
find where a device is to send it calls and data (assumption). Other
than that, AFAIK it is still not legal to use custom baseband firmware
in a GSM network not operated by you. At least that is what I've heard
the situation was almost a year ago. Maybe lawyers in the free software
world are already working on fixing/changing that.
Good to hear. New info is always good.
I'm interested in such a phone. I use an old mobile phone right now which I'll give to my parents eventually and it would be awesome to have a free-software friendly smartphone. If it will really cost below $150 and work with German telephony providers I'd buy that.
I don't think there would be an issue with it working in Germany. However at $150 USD it is unlikely to generate much profits. It might be possible to get the cost down to around $100 USD. The problem is even that really wouldn't generate that terribly much. I think the prices keep dropping for the one phone which would work so it may be that a large quantity purchase + a little more time could see it turn into $80-$90 USD cost and $160 USD sale. At that price it would be a good solution.
How about "Boot 2 Gecko"? Since Mozilla is behind this OS, I hope it will be as free as possible... but I could not find any information regarding the drivers/firmware.
Maybe ThinkPenguin would like to be among the first companies to ship those phones at an international scale (the first B2G phones will be made by TCL Communication Technology and ZTE and sold in Brazil by Telefónica at the beginning of 2013).
Maybe grvrulz would like to consider this OS for his project.
I just read the article about Firefox OS phones being launched in Brazil. This seems like a nice thing, although I doubt any of those phones will be as Freedom friendly as we need. Most common issues might be regarding drivers for the GSM modem and WiFi chip. Though I believ it would be possible to run it on an A10 CPU with a reverse-engineered MALI400 GPU and some trickery.
The main thing is, we need a working screen reader and a daisy book reader for the project, and if those are available for Firefox OS, it'll be awesome :D
> might be both worthy to take on and draw in new money.
> If we looked into it further would you spend $200 USD on a phone (excluding
> VAT) that excluded as many unnecessary non-free bits as currently possible?
> This would be a usable phone unlike developer targeted phones such as the
I think that such a possibility would be probably the only reason I'd
switch to a "smart" phone.
The price range looks reasonable for me.
This is what I mean by demand. I don't think it is a good idea to encourage people to buy phones with non-free software or any phone if ones doesn't have one already. Although for those of us who are already stuck getting such phones it makes sense to shoot for furthering the projects working on such a device.
Again- I have a phone coming that is unrelated to this. I'll probably see how it works out with Replicant and maybe we will go forward from there. Maybe not. We will have to see.
If somebody can propose a good argument not to do this from a free software perspective I'd like to hear it.
I don't wanna rain on anyone’s parade here but this isn't really a runner is it?
I can't talk for other territories but here in Ireland the mobile market is mature and saturated with low cost Android devices. So much so that people are actually upgrading now and selling their old Androids and iPhones. I suspect the situation is similar elsewhere.
.... and thats just the new market, there's also a thriving second-hand market were you can pick up an Android phone for €30 and throw on Cyanogen Mod yourself. Plus the very group of tech savvy people who might be interested in this are the ones who are well aware of these alternatives. In short people have far too much choice already when it comes to smartphones.
While a fully free smartphone might work, if you keep the costs low and build on a "build/buy to order" basis Dell style, its never gonna be mass market.
You have to start somewhere. I don't think the goal here would be a mass market product. It's just more of a token gesture to raise money for free software projects and help foster a free software ecosystem that furthers the goal of a free software phone. Right now we have people working on free software for little to no pay and free software advocates using non-free software. People who would otherwise buy a free phone although don't necessarily have the time to flash a phone themselves. Even if it trivial and takes less than 30 minutes our time is more valuable than the little bit extra it would cost to get it preloaded (from a financial perspective, not a free software one).
I'm kind of getting the feeling this is a worthwhile effort. At least worthwhile to investigate further.
Well uh, https://www.mozilla.org/en-US/b2g/ is meant to be free as in freedom when released I believe.
It most likely won't 100% meet the FSF guidelines BUT if I'm not mistaken, a free software spin off could be created from it.
We won't be releasing a phone with replicant. There are too many critical issues in connecting to wireless and 3g networks (no Internet connectivity at all).
Dose that mean that the free software friendly phone is abandoned or you're still looking for options ?
For the moment the idea is abandoned. It was never a priority project. We have priority projects (replacement for the AR9170 USB N chipset for instance) although had a free phone been feasible (enough functionality to support basic features at a price point which would sell) I was thinking we could roll with that and make something more readily available (something you could actually buy) and in the process help foster future development. It isn't there yet.
It might be something we would still contribute to financially. We are working with other projects and have plans to setup a free software fund of sorts. It would most likely work like this. 10% of the proceeds help the project itself and another 15% would go into a general free software fund to help less prominent projects. What I mean by less prominent projects are those which don't have a web site that users visit. An example of a project with such a web site would be Trisquel. A project where there isn't such a web site would be GNash, Lightspark, or OpenShot. These projects are either components used in distributions like Trisquel or the programs are obtained through a repository. There is little to no opportunity to ask the majority of the user base for donations or raise money through means such as advertising.
The exact details aren't set in stone yet. There is at least one agreement with a general understanding as to the plan although no specifics.
I'm sorry to hear that.
You mention some people don't have the time to flash and configure the phone themselves. Here's where I think there might be an opportunity, flash the phone for people if they don't have time or expertise.
I haven’t worked it out yet maybe something like this .... you charge to flash per phone plus delivery, or people can order the phone through your web site and get it delivered ready flashed with free software. Something like that.
For the right price, I'd consider buying this service.
> phone themselves. Here's where I think there might be an opportunity,
> flash the phone for people if they don't have time or expertise.
That is basically what we would be doing. The problem is that a phone shipping with Replicant right now wouldn't work terribly well. Even the G1 (the oldest model) does not have wireless (802.11G) or 3G support. You can make calls and that is the extent of it.
The best thing you could probably do right now if you wanted to support that effort is to make a direct financial contribution. As until we have something else to offer that at least half works there isn't going to be much of a product.
You have to remember that we have to support it, flash it, and so on. It takes a considerable amount of time to flash these phones. It took maybe an hour to flash one phone. It is possible that this could be reduced significantly although there is still the issue of half the basic functionality not working. If the 3G or 802.11 chip worked with Replicant it would probably be "good enough". The actual phone seems to work OK and much better than the freerunner did.
I think even the freerunner though had wireless and/or 3G a couple years back. The thing was slow and unusable and had voice issues at the time. The main issue with it though is the cost. If you could get the freerunner down to $100-120 price per unit in quantity and then a consumer price of $200 USD with a bit faster CPU, working code (browser, calendar, contacts, etc), and no voice issues then it might be a workable solution. That isn't the case today. The replacement board is several hundred dollars and that doesn't include a case. By the time you had a final product you would probably find it selling to consumers for around $1,000-1,200 USD.
In a time when network connectivity via cell phone is becoming ubiquitous, I see the benefits of having a completely free cell phone. It is important to realize that a free cell phone doesn't free you from the grasp of wireless carriers (pun intended). You are still at the mercy of "Big TeleCom." A project like OpenBTS is working to resolve that issue though.
I have been working on making an Arduino based cell phone for a few months now. The device I have made is ugly looking (and about the size of an original Nintendo Gameboy) but is completely free (both hardware and software) and can place/receive phone calls and SMS messages.
I am now in the process of re-working my singleton code base into a few disjoint libraries for ease-of-use in other projects. The code is hosted on Github at https://github.com/ZachWick/ArduPhone and is/will be dual-licensed under the MIT and GPLv3 licenses.
I found this initiative inspiring http://freenetworkfoundation.org/
A dual licensing MIT/GPL does not make much sense to me. The MIT license basically let anybody do anything with the code, including sublicensing it under the GPLv3 (or under any other license... which can be proprietary).
If you can come up with a plan for hardware that is mostly free (things like wireless, 3G, etc), financing, and a production price point around $100-120 USD (in small quantities) we would definitely be interested in working with you.
Sounds great - for this first iteration, I have spent about $300USD on parts. This surely will come down because many of the parts that I have purchased have been Arduino shields which are not needed in a "final" version; They just help the initial prototyping go quickly.
I will keep the community updated on the progress.
:) Me in particular please. Send me an email when things are further along.
I know the upcoming Firefox OS is trying to be the poster child for a "true open source" phone but I am guessing they may have their fair share of non-free parts. If the Mozilla team is sticking to hardware that only uses free components and drivers, I would be surprised and Chris may even leave a nugget in his pants.
I sent some emails although from the sounds of it what they mean by free and what I would call free are two different things. Neither of which are really completely free.
I should also point out that what they are releasing is 100% free (most likely from the appearance of things). It's what they are not releasing that isn't free. It is the portion which they didn't write. It what the hardware is dependent on from the companies manufacturing the hardware (drivers/firmware) that will ultimately make the device a not even close to "free phone".