Posted by Ajohnw on 24/03/2016 23:55:35:
Your eyes will glaze over even more when you see where new code should go along with builds to go with it. It's here
**LINK**
I did watch the video and no amount of next move prediction would improve the g code. It would have to scan the lots and arrange things in a sensible order. That really would be a bit of a challenge. I also thought that the visuals were pretty naff really and could easily be better. Probably will be one day.
If I put my software hat on and was asked to produce something that predicted the next move I would immediately think bull 'hit as the next move is in the g code and do something entirely different if I did anything at all. I can understand why no one would want to pick up something like that. Not with the objective name suggests as an aim anyway. I'd know that I would be wasting my time.
The real problem is that for casual users it's never reached the desktop level. Probably down to little need for those that use it. They can write decent G code. They are going to be more interested in how it performs and it sound like Tormak should be very pleased that they have done it as it will have saved them a hell of a lot of work.
If they want to engage with the people that actually do it this is one place to start
**LINK**
As there is more interest these days I wouldn't bother using any of the others and anything posted on there might generate suggestions that it needs to go to the dev people. There probably is a route. Might just be a mailing list
Doc's are always a problem on Linux applications as it changes so quickly at times. Sometime radically. If you can't cope take up knitting instead.
Actually the gui doc's don't look to bad. Nor does the interface actually but as it covers rather a lot of options there may be rather a lot of variations.
**LINK**
My original comment about it being old hat really wasn't associated with either Neil's John S's comments. I just think that these days the whole approach is wrong but it does save rather a lot of work as it's there.
John
–
John,
In this case there can't be any prediction as it all separate conversational programs appended to form a code. The first program can't predict as it doesn't know what the second code will be.
You say If you were asked to produce it but this is where we differ, you are always looking under the bonnet, myself and 1,000's like me just want to drive it. I'm not bothered where to store code or the docs on Linux.
Why do I need these ? I just need the handbook on the machine / controller.
Mach 3 was started God knows how many years ago. Art has been out for years . When it came out W2000 was current. John Prentice wrote all the documentation for it and since then it's hardly been touched and we are on W10, 32 and 64 bit and in many cases it still works and the Mach 3 book is reasonable current. Even though it's no longer under development they are still selling licences for it.
Just checked, according to the copyright on the original manual it was 2003 so probably came out in 2001
Oldest email to Art I can find is November 1998 and it was still beta then.
Long short that program opened the door to a whole generation of CNC users.
This is the difference we are talking apples and oranges. I dare say you could write a program, I have no idea of your skills and wouldn't want to decry anything you could do but the development of Mach has show that although many rolled up to the plate very, very few stuck the course and if it was all for free / open source then what incentive would you have ?
Art proved it 4 times over, first with Master 5 which was really Mach 1, then Mach2, then Mach3 and now with Gearotic which has it's own controller called Auggie
What this has proved is there are more people out there who want to drive a machine than play with the bells and whistles.