"Start writing code" is the exact WRONG way to start the software project. In fact, it is a classically wrong way that is common with beginners. Every software management book written in the last few decades starts out showing all the common mistakes and this is #1 on the list.
The best analogy is to building a house. The beginners say "why not just buy some lumber and start work, everything knows what house looks like. He gets halfway done and someone tells him the foundation is wrong. So he fixes that. Then even worse after the house is done they tell him the floor plan is wrong and it needs to be a single floor because the elderly owners can't walk upstairs. So much software is done this way that it is the #1 reason why projects fail. The correct way is to spend considerable time upfront looking at what the customer really needs, not what he says he needs. Then you might do some prototypes to mock-ups and show them around to get feedback and more ideas. You can do this a few times. Finally, you can get into high-level design, figure out what the major parts will be and how the parts interact. Next, I'd do prototype work and verify some of the harder parts can even be done. Finally, you get to detailed design stage, do that, show it around and get more input. Now you are ready to start writing code. But why bother when most LinuxCNC users are happy with what they have? Most DOS users were happy with the DOS command line. But there were few DOS users. It is kind os self fulfilling, those not happy with the way LinuxCNC works never become LinuxCNC users. Look what Tormach did. There target customer base was NOT LinuxCNC users. They went after a larger group of potential costomers So this email list is actually a very poor place to get ideas, A better place is to find where the hordes of machinests who looked at LinuxCNC and said "not for me" and moved on. What were THOSE guys looking for? On Fri, Sep 4, 2020 at 11:57 AM Mark Wendt <wendt.m...@gmail.com> wrote: > On Fri, Sep 4, 2020 at 1:05 PM Chris Albertson <albertson.ch...@gmail.com> > wrote: > > > On Fri, Sep 4, 2020 at 9:50 AM Mark Wendt <wendt.m...@gmail.com> wrote: > > > > > On Fri, Sep 4, 2020 at 12:41 PM Chris Albertson < > > albertson.ch...@gmail.com > > > > > > > wrote > > > > ...So, to make LinuxCNC nearly universal, hide it inside a product > that > > > is > > > > slick and easy to install and use. No one should have to look at HAL > > > files > > > > or know it runs on Linux. They can learn, but if learning is > required > > it > > > > will always remain a niche product. > > > > > > > > > > Great stuff! So, when are you going to get started on this project? > > > > > > > I wrote that to show why it will never happen. But also to make a point > > that there does exist a pattern in the way complex niche products become > > mainstream. Usually, another layer is built around it. PCs were kind > > of rare until Windows covered over the DOS command line. > > > > > And I wrote that because of all the pie in the sky whining about what this > set of programs needs to be, and no one stepping up to make the effort to > do something like that. The vast majority of LinuxCNC users are quite > happy with what our developers write for us to use. It's open source > software. If you don't like the direction LinuxCNC is going, fork off, > write your own version so that it does what you perceive to be the next > golden goose, thump your chest and tell the world you now have the latest > and best CNC machine controller. > > Otherwise, lets stop flogging the deceased equine. Unless all you guys > that want this kind of stuff take the bull by the horns and start writing > the code for it, it ain't gonna happen. > > Mark > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > -- Chris Albertson Redondo Beach, California _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
