Errors on mesa-card selection
But it is a phantastic tool which improofed over the years. In the moment i work a bit with seeedstudio to get parport running in epp mode on their x86 mobo. And for that i base on your tool to have a simple reproducable 7i43 config. Maybe i can convince them to support lcnc as this little beast has nice diy features.
In my eyes the best way of working is to setup as extensive as possible with your tool and then swap to hand modification. But then the learning curve into hal is demanding for a beginner.
Hope you lonely guy do not give up.
If I got it right, pncconf is intended for setup of new machines. No matter whether this is your first machine or not. It helps a lot in writing tedious stuff.
So beginners may expect, that a generated config runs out of the box.
Experienced users of linuxcnc might know enuf from hal and all the internals, so that they could fix unproper generated files.
In my world a beginner is someone who does not know how to solve his problem. An experienced user can help others solve their problem.
So I can't see, why using pncconf "demands" for a beginner.
I believe, that the best way to say thankyou to developers is to give feedback about their work. There's no improvement without feedback.
Hope you lonely guy do not give up.
So using something without feedback helps just yourself.
Linuxcnc is a platform that does not restrict you. I allows you to adapt on lots of places and has a lot of described interfaces (here i do not talk of hw).
It's center point is the hal and to configure this is a bit outdated. Once you understand the moduls concept and its "spelling" you are on the right track. (fx. I would love to see the -> and <- beeing imperative, and also a cross-reference)
From beginners point i agree that the two generators should produce running applications. Problem is in my eyes the wast variety of possible hw and "machine" configurations. So Chris is in my eyes doing a super job, as what i read here, his generated solutions are often near to the end solution. I would not dare to try writing such a tool.
The question is how much effort (of a not so big develover comunity) should be in making it easy for beginners that only expect a first-time-right solution versus giving beginners a good starting point for their self-made solution. I myself love to be proud on my self-improved solution.
I totally agree with you: Yes, development only can become better with feedback. Most, as you, give it in a nice way. Problem starts with peoples that try to command the developers how they have to spend their unpaid time... Luckily you see that kind very rarely on this forum and they dissappear soon
And they do for most use cases. The stepconf will produce a working machine in 5 minutes, granted using parallel port with no BOB, always. The issues come into play when using the wide variety of cheap BOB's that do not use standard pinouts, require several power supplies, do not have pull-ups on inputs, etc.
Mike_Eitel wrote: From beginners point i agree that the two generators should produce running applications. Problem is in my eyes the wast variety of possible hw and "machine" configurations.
Same with Pncconf, it will generate working configs for 80% of use cases, hell it produces a working config for stepper systems with encoder feedback, needs only a single line omitted for each axis. In my book that is magnificent, given all the possible options.
I can NOT expect Pncconf to generate a working config for a 5 ton machine with 6.5 meter working envelope, with several pneumatic and hydraulic valves/pistons/clamps/sensors attached to them, brakes on all axis and two of them on Z, double X axis, 8 tool carousel, 3 side tools one of them on both sides, serial absolute encoders, etc and all the logic required to run such a machine.
But, it will generate a config that will work in general and give a nice starting point for further editing.
Chris deserves his picture painted on the moon so we can stare at it at night after successfully running a machine that would require weeks or even months to get working without this tool. And he keeps adding features to it and to plenty of other LinuxCNC stuff every day.
Mike_Eitel wrote: So Chris is in my eyes doing a super job, as what i read here, his generated solutions are often near to the end solution.
It's true I appreciate and frankly need feedback.
As the programmer I'm biased to use it as I originally thought - others use it as they require or discover and that helps to uncover corner cases and any lack of utility.
I would love for it to work 99% -It's a tough problem - but I'll keep tinkering on small bugs.