Why real time kernel is needed with board like MESA?

It seems we have another case of Educated Far Beyond Their Capacity to Undertake Analytical Thought or at the other end of the scale are we seeing the Dunning–Kruger effect in full swing ?


I'm happy just being a dumb ass labourer. Actually whatever role I took on I'd still be a dumb ass.

You could have been humble.
Proposed some ideas.

Linuxcnc is good in itself.
You can start a new project.


you can fork linuxcnc and make some changes.
port the whole system to a FPGA, call it EMC3 or whatever.

Because what you seem to ask is not Linuxcnc really meant to do, that is basically an Arduino with CNC sheild and being controlled over usb.
I know there is nothing wrong with that. We need projects like that. Many new Open-source browsers evolved even after firefox. And there are more than enough linux distros. So, scope of your suggestion is pretty big, but you cant judge what other people have done.
But changing an existing project fundamentally is not good.

Projects you showed in examples are good and can be refined for daily use, so support projects like those, pay some programmers to work in those.
When it comes to community, you cannot force people. (though I like the way Linus Torwalds work )

Pulling the motion control out of Linuxcnc to stick it in hardware outside, is kind of defeating the whole point of Linuxcnc (controlling a cnc with a PC). It's like someone going to a railroad enthusiast and saying you can do it all better with cars and trucks, but the enthusiast will reply, but then it isn't a railroad.

The cnczone thread he links to has been dead for 10 years or so.
Hackaday meh.
One of the other links didn’t have a great deal of substance.
Two academic papers, one with spelling errors.
I think our friend might be an enthusiast of technology rather than one who has interest in machining. The type that weighs in and proclaims “current thinking is this way”.
Mayhaps a Technology Fashionista ?
Weight till he sees that it is possible to emulate a FPGA on a GPU. The vid I saw on it confused me greatly.

Its quite simple really. Every system has a paradigm that defines its use case.

Linuxcnc has a paradigm where the PC is the core of the system. This has a lot of advantages but it requires a real time operating system.

Other systems have different paradigm. Maybe they put the core of the system in a FGPA because they are not running on a real time operating system.

What you are proposing is to change the Linuxcnc paradigm. Take the core out of Linuxcnc and there is nothing left. So you are proposing the impossible. Far better to play with the correct paradigm on another platform.

I love that this platform is an open one.
So he can feel open to present here his ready better solution.
m5c
Mike

I can't understand why such a simple and logical proposal scandalizes some of you like that. Obviously, some of the participants in this topic are smart enough men who can judge that an idea is not a detonator that can blow up what they have today !!!
Why should you be afraid of the future? Let's think realistically, in terms of the paradigm that rodw mentioned.
Do you think that subtracting real-time from PC will change the LinuxCNC paradigm to such an extent?
Is it really not understood that the change is not dramatic at all? You just have to believe that the performance of the task by FPGA will be as expected and even better than the current situation - the impulses to the motors to be generated by the PC.
My opinion is that this will not be a weakness and a deviation from the paradigm, but will give a decisive advantage for solving tasks in which the PC is stronger.
Especially to BeagleBrainz who does not want to read a Chinese document, just because he had spelling mistakes, I will give a link to a document in his language, by an author who I hope is reputable (for him). I think there are no spelling mistakes.

Thats the thing, you cannot simply extract the real time requirement from Linuxcnc. That is at its core. It can't be extracted!

The Mesa FGPA cards generate the step pulses but it is Linuxcnc which tells them what frequency to run at. Soit is clear you still don't understand the environment with linuxcnc.

You really need to stop this nonsense and research the Linuxcnc architecture. If you did that, you might actually begin to understand the amazing benefits real time bring to linuxcnc..

If I didn't read the document, how would I be able to say there were spelling mistakes ? I'm not psychic.

Anyway enough of that.

I think the major problem here, as you stated clearly yourself, you can't understand why your proposal is not accepted. This would appear to be your major failing. Not a failing of others or the project.

It is fine wanting change, but those that want change must first consider whether that change is needed rather than just wanted by a small section of the community. You need to evaluate what your proposal means in the way of change and how much it would completely change the goals of the project.
Actually this is what some "isms" do, a small but to begin very vocal group try to push their ideas, with the idea that they are right and anyone who disagrees with them are wrong. They believe a minority has the right to bend the majority to their will. One of the most evil in doing this was Che Guevara. After a while the idea loses all focus and Revolution becomes the focus.

What you are suggesting is a complete change of goals, objectives & direction. Goals, objectives & directions that maybe not wanted by a majority. Yes I accept that there has not been a large enough representation of the community active in this thread. I think too that being an active member, either as a developer, tester or member that puts time in to help others can help.

I think the general idea is that your proposal does have merit in the "greater scheme of things", but your proposal does not fit within the goals of the project. Some of the current modularity would be removed.

Maybe the idea might be to, you may already be actively doing this, promote your idea if other forums/places of interest and create a group of people that are willing and able to join in your endeavours. I can't for sure estimate what your hit to miss ratio will be. I honestly wish you luck, but I don't think it will come by proposing major changes to existing projects. At the very least you would have to present a detailed proposal, rather than quoting what others are doing. I wish you well, but your goals would seem to be at odds with the project as it stands.

Whilst you have the right to express your opinion & not be silenced. You do not have the right not to be disagreed with.

rodw wrote:

Thats the thing, you cannot simply extract the real time requirement from Linuxcnc. That is at its core. It can't be extracted!

The Mesa FGPA cards generate the step pulses but it is Linuxcnc which tells them what frequency to run at. Soit is clear you still don't understand the environment with linuxcnc.

You really need to stop this nonsense and research the Linuxcnc architecture. If you did that, you might actually begin to understand the amazing benefits real time bring to linuxcnc..

The gent is entitled to his opinion Rod, but researching the project is very sage advice.