TurrisOS does not feel like a true open source project

First of all, don’t get me wrong. I have my Turris Omnia now since a couple of weeks and I’m really excited about the hardware. Sure, the software had some bugs in the out-of-box state, but nothing that couldn’t be resolved by searching these forums since most people had already encountered and solved the same problems. I also understand that the source/commits/etc are all available from github.

With that said, I am however slightly disappointed by the current development process. This forum is full of users that talk among themselves, with the occasional input from a developer - much like a product forum for a “normal” commercial product. There is no way to know the release status of the next version (which by now includes a large amount of fixes), what the roadmap is, what is being discussed within the devel team, etc.

To give a few examples:
A fix was quickly committed, but it’s not available for anyone not able to compile their own kernels. There’s also (AFAIK) no list of known bugs in the current version of TurrisOS maintained anywhere, so any new user trying out StrongSwan will encounter this bug.

(yes, that was my own pull request) Develop a patch, submit a pull request, wait a week, notice that the same functionality has already been implemented by the dev team meanwhile. And get told “We already had the plans (and some PoC code) for it when you submitted the PR”. I mean, it was a friendly reply, but seriously, anyone wishing to contribute to TurrisOS shouldn’t have to guess at the plans and current development work of the core dev team.

I guess what I’m trying to say is that while Turris meets all the formal requirements of being open source, it fails miserably in getting people involved, excited, keeping them up to date, luring new developers into the fold, etc.

A first step would be to have a proper mailing list and to make sure that the dev team actually uses it for communication, including internal communication rather than whatever means are currently used.

(Again, please take this for what it is - constructive criticism of how Turris development could be improved).

1 Like

Give them a ultimatum till februari 2017 and if things haven’t changed or so, then we have something to “complain” about. Right now i THINK everybody JUST got their Omnia.

I THINK till now they have been busy with one thing after another, while the software side has been neglected because of certain priorities. I mean fixing bugs is useless if people haven’t even got their Omnia right? So give them some time to organize things.

Ultimatum? No. Wait patiently? Sure.

And I do think that having the internal dev team dialogue in the open would be a minor “fix”, not a major undertaking.

That also takes time. We must keep in mind that this “thing” is also new for them.

The only “addition” i am having is that the forum is rather chaotic. We the users can help one another indeed as we have been doing for almost a year now, but important threads are being lost in the ocean of threads. Even this thread will be lost and somebody else will probably say exact the same thing over a couple of weeks :joy:

Even though I partially agree with some of your statements, I think you are mixing between open-source project (which it is) and community project (which it is not). Consider for instance the Chromium. Even though it’s open source, most of the development is done in-house without direct connection with the community. I don’t say it’s right or wrong.

Personally, I think most of the developer communication actually happen in spoken Czech language in the offices of CZ.NIC, so putting it down to an English mailing list would dramatically slow down the process.

To the speed of fixing bugs: There’s very big responsibility of the CZ.NIC not to introduce other bugs. For instance, today I switched one of my Turris routers to the nightly branch just to discover the update that should fix local DNS names actually breaks my DNS resolution completely. You certainly don’t want to distribute such update to 6K+ subscribers, do you?

The responsibility of upstream maintainers is, from this point of view, much lower.


In order for a product to be open source in any sense, it has to be the case that people who have the hardware can do a build from source and get the same software that is on offer from the vendor. While I think you have paid lip service to this idea, in practice I have no idea how to get that software, how to build it, or how to install it. What we actually have in practice is a router that is no more open than those DD-WRT joke routers that Linksys and Netgear sell. I call them joke routers because while you can get the source, it’s so ancient that there is no point in hacking on it.

If the problem is that your team is communicating in Czech, that’s perfectly fine. I have google translate, and actually wouldn’t mind practicing Czech before the Prague IETF. If there are things your team is communicating that can’t be public, that’s a problem, not in the sense that it’s wrong for there to be such things, but in the sense that it makes it impossible to communicate openly without hiring someone specifically to communicate, which presumably you can’t afford to do.

But the bottom line is that right now my Omnia is exactly as useful as a closed-source router, because I can’t do software development on it, and that is the entire reason that I bought it. I actually have a crappy old ASUS router on order from Amazon right now because I need something I can hack on, and that router is supported by LEDE and OpenWRT, whereas the Omnia is not. This really sucks. I am >$400 in the hole for a piece of hardware that is purely decorative.

it has to be the case that people who have the hardware can do a build from source and get the same software that is on offer from the vendor.

This is exactly the case.

in practice I have no idea how to get that software, how to build it, or how to install it.

The fact that you don’t know how doesn’t mean people can’t do it. It works like with OpenWRT / LEDE, the build process is the same. There is a wikipage in Czech about it: Kompilace Turris OS ze zdrojových kódů [Turris wiki] You can find details there and also mention, that there are some handy scripts available.

Thanks, this is quite helpful. I had gone to the english version of the page, which is blank, and it didn’t occur to me to click on “cz”. This still isn’t really what I want, but it’s a step in the right direction.

@mellon: I am not sure which search engine you use, but a simple search for me for “turris omnia build root” returns these as the top hits: