I am not an optimist that a single person can handle and mantain such an app and make it comparable with for example FreeCad (I use it and it has significat problems).
I also use FreeCAD, the company that tried to commercialise it has just folded, and I feel that there are lessons to be learnt.
FreeCAD is not very intuitive. Of course, large programs of this type (Alibre Design, CATIA, Inventor, SolidWorks, SolidEdge, Space Claim) are also not very intuitive. Here, I am focusing mainly on the model (part, subassembly), and not on simulating the operation of the machine (because this is more advanced modeling). Preparing an element with a specific shape, with various holes, recesses, grooves, etc. requires performing a lot of auxiliary operations. And these in turn require setting various points, lines or auxiliary figures. That is why many people get lost in this thicket of operations - when to perform which, how to do it (i.e. what and where to click, in what order), why it is not possible to do something more easily (because, for example, the programmer did not foresee a different way of working), etc. The market for parametric 3D modeling programs is probably already quite saturated, hence it is difficult for another company to break into it. And Free CAD is very simple compared to well-known commercial applications.
The first is that very often "less is better": if somebody has something to be done it's easier to use a simple program that does an adequate job than a complex program that does a perfect job.
That's why large companies are constantly tempting with training courses to learn how to use their programs.
The second is that it's far too difficult to work out how to do... well frankly, just about anything. Just about every mature application package has that problem, and it applies "in spades" to open-source projects where there isn't a single whipper-in who can dictate that for the next six months all effort will be turned towards documentation. In principle AI- "spicy pattern matching"- should be able to help with that.
It would indeed be nice if AI helped create documentation.
IBM lost out to DEC, DEC lost out to CP/M and PC-DOS. ALGOL-68 lost out to Pascal and C. C's losing out to Rust, Pascal's losing out to Python.
In my opinion, the opposite is true for C and Pascal. C was, and still is, more confusing to use (its descendant, C++, even more so). In this case, other factors contributed to C's "popularity." C notation is messy and confusing (which is clearly visible in large projects). It uses a preprocessor (including macros), headers, but lacks modules. It has a terrible executable file building system (CMake is no better than previous "solutions"). Basically, it has one (big) advantage: its compilers generate machine code that is quite optimal when it comes to execution on the CPU. But this comes at a high cost at the program development stage.
However, you are absolutely right that simpler (but also worse) drives out better. However, in my opinion, Python is more confusing (not harder) and time-consuming to use than Pascal (Object Pascal). Unless someone writes a simple script consisting of a dozen or so lines. But I do not create "microprograms", so it is useless to me.
And I recently passed over a piece of antenna-design software (that needs a PhD to use) in favour of 1970s abandonware.
Not all people who can program can create a useful program. Many people are very knowledgeable about certain topics and are good at solving problems related to such topics. Many of them also happen to be good programmers. But they are not always able to design the UI of a program so that it is useful. Some completely disregard ergonomics (they think that their way of solving a problem is the best). Others, on the other hand, cannot design UI at all. As a result, such software is useless. Designing the UI and the way of working with a program is not an easy matter (unless it is Notepad, but even in such cases there are people who will screw up even such a simple project).
If users can't find out succinctly how to get the prevalent product to do a job, they'll use something simpler. It affects FreeCAD, and it affects us.
True.
Recent
