I got to see Stallman speak the other week in Illinois. His idealism about Free and Open Source is pretty hardcore. The man refuses to buy an AmTrak ticket because you can’t pay in cash and he doesn’t want to be tracked. He pays for his transit cards in cash instead of using a card at the machine.
During the lecture he said he decided early on, if he couldn’t work and earn a living entirely using free and open source software, he would rather wait tables (emphasizing that waiting tables is a very respectable profession).
Looking at this background in this article though, he grew up in a different world, got in early on some ground floors and made some early strides that are difficult to achieve today. I recently lost my job and am very reluctant to go back on the market and go through the entire interview process again.
I would love to work entirely in open source software, and have been working on getting back into a PhD program (I currently hold a masters). I work with a professor a knew from an old startup and he uses some of the software I wrote (BigSense.io) in his classes and we’re currently working on some new tutorials, docker containers and lab manuals. He’s still struggling to get funding though, and even if any of it comes in, I’d be lucky if I got a stipend to that would cover my rent.
The big question is, now, today, in 2018, how does one live entirely off FOSS development? I feel like the big developers for things like the Linux kernel; over 50%~60% of them probably work in the OSS divisions of Redhat, Microsoft, AMD or Intel (I have several friends I graduated with in Intel’s division out in Hillsboro/Portland).
What types of grants and fellowships should I be applying for if I want to be able to work on FOSS full-time?
If you want to apply for a job writing entirely FOSS (or doing other work for a fully-FOSS organization), there are a number of job boards for that:
You and @zxtx are right about some of the other options: if you start working on a big FOSS project that lots of companies use, if you’re good enough one of those company will eventually offer you a job out of the blue.
Personally, I would recommend starting your own company that sells FOSS in some way. There are a number of great business models for this (and some not great ones, which are unfortunately used a lot) - I gave a talk about these at LibrePlanet this year:
My own example is https://jmp.chat/ (and to a lesser extent https://ossguy.com/ss_usb/ ). I was fortunate enough to have a few months of runway to try it out, and it ended up working out. In general, you do need some runway for any of the options listed above that are not the job boards.
I think it depends on the kind of work you do. The popular approach is to have an open-source library important to a company so your work is effectively maintaining this library. Many developers use patreon, and for some niches like scientific computing there are non-profits like NumFocus which fund open source projects. It’s not a straightforward path, but there are openings.
Denver from https://jmp.chat/ here (JMP is a free/libre and open source SMS and MMS gateway for XMPP, and a Soprani.ca sub-project). Happy to answer any questions people might have about JMP, WOM, or Soprani.ca!
Prosody + a couple of modules with Conversations.im and Gajim can be serious alternative to commercial offerings. End to end encrypted chats, history sync and seamless image / file sharing works as good or even better than Hangouts. I’ve been using this setup with a friend for several months and it works so surprisingly well that I’m seriously considering migrating my entire family to XMPP.
There is also an option for running connections via Tor (both hidden service for the server and client connections) for the truly paranoid.
Yeah, I’ve got a big chunk of my family on XMPP and it’s been nice.
The stragglers I SMS using JMP.chat so that I never need to leave my XMPP client :)
Wow, JMP.chat looks really useful. Shame it’s US/Canada only (I’m in the UK) :(
Feel free to join the low-volume notification list to find out when UK support is added: https://jmp.chat/sp1a/notify_signup/ The UK is likely to be the next country that https://jmp.chat/ supports.
JMP.chat looks very cool. Unfortunately I already have an unlimited everything plan and a number but the idea is really clever.
I had an unlimited everything plan, but I really love the extra freedom I get now to move across devices/plans sims/wifi when I travel or when I’m at my laptop. Ported my number to JMP and got a “tablet” data-only plan for my mobile. Not for everyone, but my wife and I love it.
Data-only plan… that sounds nice. I didn’t think about porting the number to JMP - that’d be convenient.
Do you also use SIP for regular calls in JMP?
I do. Works ok over LTE and fine on wifi. If you need more reliability on lower-bandwidth connections callwithus.com supports opus codec for outbound calls and you can set callerid to whatever you want.
This sounds like it would be a great fit for https://soprani.ca/ , particularly the long-range communication part of the project: https://wiki.soprani.ca/ThePlan#Long-range_radios . If they were suitable for use in most countries, the 433 MHz bands could be a great way to deliver text and even picture messages in the absence of a cellular network.
Does anyone know off-hand if these bands are generally license-exempt in Canada or the US? It seems that’s probably not the case in the US, per https://en.wikipedia.org/wiki/LPD433 , but I had difficulty finding info for Canada.
I think you could use 900mhz for the same purpose?
Yes, after some more research it sounds like we could probably get similar range with 900 MHz. I had assumed that range decreased significantly with higher frequencies (using 150 MHz (MURS) as the benchmark, where I’ve seen 35+ km range in my testing), but I’m not a radio engineer, and my brief search suggested you could indeed get 40km line-of-sight with 900 MHz, too. 900 MHz also tends to have fewer regulatory restrictions than 150 MHz, which would be good for the project, where we may wish to do mesh and similar things not allowed on 150 MHz.
For sites that still require you use a phone number as part of 2FA, it’s probably best to not use one of the big carriers (which are quite susceptible to social engineering, per the article). Instead, get a number from https://jmp.chat/ or similar, and give the sites that number rather than your cell. This also lets you hide your real number from them.
I mention https://jmp.chat/ in part because it lets you “login” to your phone number using whatever authentication method you like (since it uses Jabber on the back-end). So you’re free to make your phone number as secure as you want.
Wow, I have been looking for something like this for ages for authentication.
Glad I could help!
Also, I’ve noticed a few people use https://jmp.chat/ for their primary phone number - they just get a cheap data-only (or “tablet”) cell plan, and then use JMP for SMS/MMS/calls on cellular data (or wifi). A nice way to make your phone number more flexible, and secure.
I’m the speaker in this talk; happy to answer any questions anyone might have!
The main project that I discuss (and demo) in the talk is https://jmp.chat/ . Since the talk I’ve also made an outline of the overall plan (a more concise and complete version of the slides): https://wiki.soprani.ca/ThePlan . Comments and/or suggestions on that are also welcome.
I find it’s a bit strange you started at the end you did….
To me the main problem to freeing the cell phone is the baseband… ie. This step…
Once we have the above done, there is still the problem of how to communicate with people when outside of wifi coverage, which is something that a cell carrier normally provides.
So I have been doing a fair bit of digesting the standards in this area and poking at cellular modems and the like.
And I would say most of the complexity and code in the whole cellular stack is around revenue collection. ie. The carriers will fight you tooth and nail to preserve their revenue stream. ie. The problem isn’t the simcard.
ie. They will never let you on their infrastructure if they can’t extract the full amount of revenue.
One alternative in this space are these guys…. https://villagetelco.org/mesh-potato/
Even then, wifi itself is a vast bundle of incredibly complex proprietary magic..
The only guys even starting to step outside of that are https://gnuradio.org/ and on top of them is http://openbts.org/
I think you’re right in that it should be IP first, voice second… (BTW, it looks like LTE is doing that) but I sure range beyond wifi is the main problem.
It sounds like we mostly agree on the solution here (but correct me if I’m wrong). In particular, the long-term goal is to not need a cellular baseband at all, or any connection to the cellular carriers.
The reason I mention some cellular-dependent options in the talk is because we need some interim solutions (ie. ways for people unwilling to switch devices to use non-cellular more easily). But in the long run we can hopefully migrate people to devices that don’t need such workarounds, especially when we have better and longer-range community-run radio networks.
especially when we have better and longer-range community-run radio networks.
I guess that’s the bit I’m interested in….. Partly at a professional level since I work with LMR radio. (Hey, I get my pay check from number of handsets and base stations we make… not per call made.)
It would be great if we could chat more about this - we’re always interested to hear from people who work on radio. Feel free to join our XMPP group chat at email@example.com (you can join via the web if you like) or send me an email using my contact form.