LibreCMC: Free Software Router Distribution Beta

It's modern. LibreWRT is based on a very old code base that isn't compatible with routers out today. LibreCMC is.

Libreboot isn't a distro; it's a free version of Coreboot. You probably meant LibreWRT or LibreCMC.

The *key* thing here is this router ships with 100% free software, will be RYF certified (shortly), will be funding LibreCMC development, and is commercially available. No other commercially available router on the market ships with or runs a 100% free OS right now. If we get LibreCMC off the ground that'll change and right now this is the first router.

If you wanted to go out and buy another device LibreCMC may run on it still. It's a distribution for embedded devices after all.

The problem is that most routers and other devices are dependent on some non-free software. Even devices that you think are 100% free software friendly are really just less hostile. As a case in point there may be non-free code for a gigabit switch built-into an otherwise free software router. It's not obvious and LibreCMC may still run on it without problem. So you'll still be running the risk of running at least some non-free software if you buy a random router and try to run LibreCMC on it.

There is also the problem of model numbers not being sufficient to know what your actually getting. The community regularly bricks devices in the process of flashing them. People get upset because nobody will help them 'unbrick' them. Sometimes you really can't unbrick them if its a true brick (at least not without special tools, and even them I'm not 100% sure you can always unbrick either, etc). It's really quite unreasonable to get upset over such things when everybody has clear warnings about it, but none-the-less.

So far we have tested LibreCMC on several embedded devices and hopefully it'll eventually get used more widely. Right now it's just this one router and for us that's a good start. The other devices I believe we tested against: the Ben NanoNote, another was a higher end router with more flash, and a mini wifi adapter. There probably were a few others I'm forgetting about. The one router might have had some non-free component for the gigabit switch. Not sure what we concluded, but it wasn't a priority anyway given it wasn't a device class that would have much demand.

Yes- Like LibreWRT LibreCMC has taken OpenWRT and ripped out a ton of code. At the moment that's about all it is. However if there is demand and people contributing there is a lot of potential for the router and other devices or types of devices. Particularly ones that are freedom-oriented.

For example I'd love to put out a product that makes it easy to maintain your own personal mail server. Currently the laws in the United States make it easy for law enforcement to gain access to any emails stored on a third party server. While in transit they are protected, but after a certain number of days thats no longer the case. Using companies like Yahoo! and Gmail put your communications at unnecessary risk.

In an ideal world there would be a cheap device (under $100) that users could plug into a network and access remotely. Something that gives users all the benefits of web mail without the privacy head aches of using a third party.

Now this isn't the end-all solution to the problems governments present (think NSA). Such agencies are still going to be able to monitor unencrypted communications even if such communications can't be used in court (sadly, it is apparent it is being used to identify people of 'interest' though, commonly known as "Parallel Construction"). No. We need a solution that automates the use of encryption built into such a device as well. Ideally it would have the anonymity properties of Tor with the authentication of GPG. That is you can be relatively sure the person your talking to is the same person and nobody else each time you communicate and the other person (and governments) still can't identify you. At least not unless you identify yourself in some way intentionally.

The one thing I don't like about the FreedomBox is that it's dependent on hardware that is not free software friendly. We can't be sure what it's doing. It's a step in the right direction, but has quite a ways to go. I'm also not sure what the project is doing exactly. I think it's sort of got similar goals although its not entirely clear to me at its current stage who its target is, what its core features are, etc. Not to mention it has had major delays. It seems to me like we need an intermediary solution that solves some of the issues at least until it's further along. It may be a better solution, but not at all a solution for years to come (if ever).

I think it would be great if we all worked together more... but I'm not sure that those involved are on the same track/path. Everybody is on a different path and contributing for different reasons and with different ideas/development experience. As it is it seems to me the best approach to sustained development of a free software embedded OS is to fork OpenWRT and build a business model based on that to help maintain it. I think the lead developer of ProteinOS is sort of on the same track, but with unrealistic expectations and no clear business plan. It's unlikely somebody is going to just come along and pay up hundreds of thousands of dollars to your project to keep it going. He's eventually going to have to get a 'real job'. I already know RISCi doesn't have the time to work on LibreCMC or LibreWRT full time, but I think a business plan alongside it could provide limited funds to help maintain it if it were a more bleeding edge distro, but only if much of it is taken from upstream (as opposed to building everything from scratch).

Now there are two parts here I'll try and clarify. One is my perspective in at least how it relates to ThinkPenguins and one is RISCis. From RISCis perspective I think it is best summed up as things weren't going the way he wanted them to. Other people involved wanted to take things in a different direction (I forget exactly what it was, but it wasn't just a bleeding edge issue). From a practical perspective that direction didn't work for ThinkPenguin. RISCi's reasons for wanting to go in a different direction differ from mine as it related to ThinkPenguin though. The reason ThinkPenguin has not released a router based on LibreWRT is that the code base has become difficult to work on, dated, and become unworkable for all practical purposes. Before RISCi made an attempt to get LibreWRT working with a modern router there was another experienced developer who attempted to make LibreWRT work for ThinkPenguin on a modern router. Both he and RISCi failed. Two people with experience and working independently of each other failed. One of them being an original developer.

So to sum it up I tried working with LibreWRT in regards to getting a router in the ThinkPenguin catalog. We had no problem getting it to work on one particular supported router. That router wasn't an option. I don't believe we could get that router a year or two ago let alone today.

Long story short the code base appears to be a mess, dated, and problematic. Right now LibreCMC is essentially LibreWRT v2 for all intensive purposes. RISCi basically repeated what he did for LibreWRT, but based it off a more recent upstream code base (OpenWRT, just like LibreWRT is). The LibreCMC code base is on gitorious (see link at librecmc.org) and if the LibreWRT developers wanted to participate I doubt RISCi would have a problem with that. It's why its on gitorious. RISCi believes in public and community development as much as I do. The only thing I have to say about others participating is it's going to go in the direction RISCi and/or the money that helps fund it allows (and by that I basically mean that if people shell out for a $50 router we can't focus development on a $150 router). That might not be the direction its users like, but if your not willing to contribute/finance it there isn't much we can do. Developers have to eat too!

Next. We've been working with the ProteanOS lead developer on other projects and I've talked to him about his embedded distribution. From the sound of it the key difference between ProteanOS and LibreWRT/LibreCMC is that its built from the ground up. That's great if you have the resources to do that. My concern with ProteanOS is that it's too ambitious and there is no realistic long term maintenance plan. There is a reason small successful projects don't build from the ground up. It's more work. It can also be counter productive. It's better to fork and/or work with developers upstream than build everything from scratch particularly if you don't have the resources to do that. When lead developers get kicked out of their dwellings (ie parents home, government funded housing, or because they don't have the money to pay rent) they have to find real jobs and those jobs end development of projects like this. Just look at Trisquel's slow development this past year. It's a perfect example. The same can in part be said for LibreWRT (in my humble opinion).

If we're going to fix these issues we need to bring in money to help feed / house developers. I think we can do that with LibreCMC. I'm not convinced that's possible with LibreWRT or ProteanOS. Here is why I think this too. It didn't take a significant amount of time and energy to produce LibreCMC. Relatively speaking anyway. That means even a little bit of funding could make it feasible to continue its maintenance/upkeep. The reality is that there aren't that many people willing to contribute financially or able to do the work. The unwilling / unable (ie the masses) tend to buy cheaper routers though and if you can ship such a router you can start to fix this problem. That's why we're focusing on a modern low end router. Not the $150 router. Not exciting? Not interesting? Nothing I can do about that. I don't see how we could realistically come up with enough funding to keep an embedded distribution like ProteanOS going with this business model. It might be some large company wants to fund it. However I don't know which company would or why. Nor do I know where to look. What I do know is LibreCMC solves a problem ThinkPenguin has and in turn can fund development. LibreWRT didn't. ProteanOS maybe would too if not for the development cost.

This is the harsh reality as I see it.

That's exactly what I've been saying all along, because of all the fragmentation, we have very smart people all working on the same thing and in turn it results in less work done as what they're all doing redundant, we have people doing jobs that are already done elsewhere.

What's more, I'm not programmer, all i can do is use the software as it's made and donate to projects that I support, but when there are 10 projects all doing essentially the same thing, just slightly different, it's hard for me to support one project, as donating makes no difference is speeding up development. At best all it does is buys a programmer somewhere a coffee. I think it would be better to support one or two programmers who could devote a lot of time to a project than 100 programmers who only get enough money for a coffee from time to time. If the community could work together, under one project and everyone donated to that one there'd be more work done.

And if the FSF would back one good OS and pitch in and help that would be even better, the community could get a lot done. But as it is now there are just too many projects, too little time and in turn the free software communities software is always outdated and only supports "the old stuff" compared to other projects with commercial backing, like Ubuntu.

Diversity is good, but being too diverse is harmful, especially for a community as small as the free software movement.

Sorry. I added the model # to the title and it changed the URL on me. I try to avoid that. Here is the new URL:

https://www.thinkpenguin.com/gnu-linux/free-software-wireless-n-broadband-router-gnu-linux-tpe-nwifirouter

It'll probably be available, but not yet. We're still in development.