I would like to disagree with that. Most people who learn ROS2 don't concern themselves much with QoS and DDS, and it still works. What's more, ROS2 is definitely used in industrial applications. Take a look at the contributions from industry participants at ROS CON. I personally know robotics engineers who use both Ros and ROS2 in industry. ROS2 is not nearly as complicated for beginners as you describe. If that were the case, the mechanical engineers who work with Ros would be the exception. I think that's wrong, considering how many Ros-enabled robots you can buy. It is actually much more difficult to teach beginners Rust than ROS2. Mainly because you can easily learn it in Python. You will have to learn Rust if you use Horus.
HORUS has python binding and we will likely bind C++ after the archtecture is stable. But it is fair to say that there is a tradeoff to learn Rust, giving the current adoption of Rust, it is quite a bold bet, its whether ecosystem of Rust will become more powerful, enough to adopt embedded aspects of C++. However, Rust is promising in robotics applications, giving people are adopting Iceoryx2 and Zenoh. So the remaining concern is whether students think they should learn Rust to adopt more promising libraries and codebase or learn just ROS to know enough what they think they need. If you dig deeper to the original papers for developing ROS, the initial idea for the architecture of ROS is meant for prototyping, especially the kind of wheels robots that serve in house. As components are swappable, fault-tolerant and increasingly develop( plug in components and layers without breaking the system), ROS is ideal for this. But come to production and industry standard, companies are drawn to a static and firm flow, yielding more performance, predictable behaviors and safty guarantees. ROS2 is meant to escape from the philosophy of being for pure prototype of ROS1, thats why I said DDS is industry standard, Rovers are famous for using it. The only reasons behind keeping ROS for release-version of robots might be cost of adoption, rebuilding, ecosystem or debugging, developed, repaired in production. If you stick to where everyone is, there is certainly solutions to every problems you face, this is where ROS is ideal to stick to. This can be easily proven by how many companies use ROS for prototyping stage compared to ready-for-sell stage.
Seriously, though: are your texts from an LLM? You're explaining ros2 to me for the umpteenth time. Stop this nonsense. I know how ros2 works. I actively develop with it.
The joke is, no one who learns ros is trapped in it. I don't have to choose between ros and rust. Nowadays, both are possible. Because there are rust bindings for ros -> I mean without Horus. No one who is actively involved with ros needs horus to write rust nodes. Besides, the vision statements of ros no longer have much to do with the current state of affairs. In the end, I have the best of both worlds: I can write cool Rust code and integrate it into ROS in the form of ROS nodes, publishers, subscribers, services, or actions. Horus is not necessary for this.
Horus may seem promising when you read through the readme. But when I see that you only have a few developers and two of them seem to be AI, at least at first glance – one of them is simply called Claude (lol) – it makes me extremely sceptical. When I then look at how much repetition there is in your texts – I mean, you're basically ignoring what I'm saying and repeatedly reciting the ROS proposal to me – (I mean seriously. As if I need you to read the ROS proposals to me...) WTF dude?
What makes you think I text from using LLM?
I'm talking about ROS2, as I sense from you the hatred toward ROS being attacked, that's why I'm giving it a ground up build in understanding, and working from that to convey the idea of HORUS or how ROS is being pushed away. To be honest, I don't feel like I need to explain people having problems with ROS, you can explore on your own and if you don't feel like people have to stick with it, then it is your personal perspective and conclusion. And since when your words sound more spicy, the post of this is literally about HORUS and how I'm trying to create what I wish I can personally work with instead of ROS, that is why I'm actively conveyed about what it is capable of. And attacking on the contributors being AI assitants and little contributors is questionable. You expect there should be a big tech, orgnization backing up for HORUS? A free, open source project? Telling ROS proposals means conveying the philosophy of building, which is like a vision draft board. If you are in a creator perspective, you are not only trying build something architecturally, but how is it differ from current system like ROS or conceptual systems as in the original design of ROS. It's also your viewpoint that ROS current design is not related to original conceptual philosophy, it is on my end that still personally seeing that ROS is actively used on maintream labs, more than industry. ROS inherently has the architecture for prototype, and as of now, it still holds that position, if you don't agree with that, then, consider it as my personal judgement then. I don't get the idea of accusing here with the LLM text and AI-assistants? If HORUS in your opinions will likely not neccesary, then just pray it die and move on then. While there's people who is unsatisfied with what you are actively develop and want to build something on their own instead.
Let's just break it down. You don't like ROS because you can't deal with it. I do. That's not hard to understand. What bothers me is your reasoning. To be honest, I find it completely far-fetched. The arguments you put forward are not wrong. However, the narrative you build on that basis is. You're only looking at a small part of the whole picture. And you can tell that because you obviously can't come up with any new arguments. Seriously. Read through what you're writing. It's exactly the same content every time. And honestly, that's not enough. The world isn't as small as you describe it.
If your only real argument for Horus is that you like it, that's fair. Everyone likes the tools they've built themselves. But then just save yourself all your ros hate and focus on that. I never said that I would only accept Horus once it had more than five contributors. At the same time, I understood your announcement to mean that Horus is being promoted as a competitor to Ros, and that simply does not fit with the current situation.
Firstly, you should know that this is a post, we don't read it like text, and after 2 days, how do I remember what I responded to other comments, so not to repeat myself. Secondly, what I repeat is what HORUS is visioned in my head to improve, meaning architecture and designing components. The only parts I'm telling about ROS, is what I'm trying to make HORUS different, put in my perspective, you can see I'm actually telling what your good old toy have, and how my new toy is different, it is easier to tell you what my toy is, from scratch, then just pulling ROS out and pointing out for reference. It is not enough arguments, but beside ecosystem, and architecture, what does ROS have left in theory. My reasoning is " I'm telling you about what I know about A, if you proceed and agree with that, we can move on to why I build B" this is normal narrative, because in this way, I get to understand you better if you disagree with what "I'm telling I know about A" and now we know, to put our foot on better place. Rather than just me telling compelely about B or saying all bad things about A, what I'm trying to do is give a common understanding and agreement on both sides and then we proceed. You feel more attacked when I'm just trying to understand all sides better.