+1
Review of FreeCAD
Forum rules
Be nice to others! Read the FreeCAD code of conduct!
Be nice to others! Read the FreeCAD code of conduct!
- thomas-neemann
- Veteran
- Posts: 11939
- Joined: Wed Jan 22, 2020 6:03 pm
- Location: Osnabrück DE 🇩🇪
- Contact:
-
- Posts: 52
- Joined: Sat Nov 30, 2019 7:45 am
Re: FreeCAD on the web
It may be obvious for you, but then don't be surprised with videos like this calling the software too immature to use.adrianinsaval wrote: ↑Tue Mar 21, 2023 8:15 pmisn't that obvious? fillets fail many times (and yes due to occt limitations, and yes commercial kernels work better for this) but I will riot if someone tried to remove that from the software.racemaniac wrote: ↑Tue Mar 21, 2023 6:19 pm because apparently having things that sometimes work is better than not having them.
And a button that very rarely fails, but has some small issues, sure. Once topo naming is fixed fillets will become fairly usable. But tools that have a high failure rate...
- thomas-neemann
- Veteran
- Posts: 11939
- Joined: Wed Jan 22, 2020 6:03 pm
- Location: Osnabrück DE 🇩🇪
- Contact:
Re: FreeCAD on the web
I don't know how the rate is with other programs, but here in the forum it seems that about 90% of the members give up after about 10 posts. I think there is great potential for a second workflow (simple, fast and stable)
I suspect that this will improve the rate
I suspect that this will improve the rate
Re: FreeCAD on the web
that's the way I do think also, but apparently not those in charge of FreeCAD. I dunno why.racemaniac wrote: ↑Tue Mar 21, 2023 5:57 pm You in no way think that if your changes make FreeCAD a lot more usable, it should come that way when you download it?
- thomas-neemann
- Veteran
- Posts: 11939
- Joined: Wed Jan 22, 2020 6:03 pm
- Location: Osnabrück DE 🇩🇪
- Contact:
Re: FreeCAD on the web
Linux has many "flavors". why not also with freecad. anyone who wants can create one
Re: FreeCAD on the web
Not sure if that's the culprit. Your are definitely right regarding the history, but on the other hand, the remaining issues with OCCT (at least that's my personal gut feeling) are a bit different: both problems mentioned above (fillets and thickness) have documented weaknesses. To me, hat seems to imply these aren't just "simple" bugs, but serious shortcomings in the algorithms that are well known to the developers. No easy fix.keithsloan52 wrote: ↑Tue Mar 21, 2023 10:47 pm Re what we need in relation to OCCT is better way ofBack in the day around 2013 when we were developing the OpenSCAD importer we were finding a lot of OCCT problems.
- * Capturing OCCT problems
* Being able to create simple test case to recreate the problem
* Good interface with OCCT developers
In some ways diverse but as the problems occurred when parsing csg files and making FC requests, things were quite repeatable.
Where as sometime problems with OCCT occur and people and people are not sure how or what they did and hence how to recreate.
shoogen ( Sebastian Hoogen ) wrote a utility i.e a program that at the time was able to perform what was required to recreate a OCCT problem. Doubt that is still around or up to date. The alternative would be to get OCCT development to accept recreating OCCT problems by FreeCAD, as opposed to an equivalent C++ request to OCCT.
Just me two penneth worth.
Cheers,
Markus
Markus
Re: FreeCAD on the web
Because it partitions also the development power.
A Sketcher Lecture with in-depth information is available in English, auf Deutsch, en français, en español.
- thomas-neemann
- Veteran
- Posts: 11939
- Joined: Wed Jan 22, 2020 6:03 pm
- Location: Osnabrück DE 🇩🇪
- Contact:
Re: FreeCAD on the web
I just mean different configuration files that create customized user interfaces.
edit
this can be done by some users, the programmers don't have any more jobs
Re: FreeCAD on the web
I think since we are using OCCT's api, we should be able to report to OCCT the exact circumstances that can cause a problem. Maybe add a build switch that logs all interfaces to OCCT. Next time we get someone asking why some operation doesn't work, (probably in about 2 minutes) we can recreate it, and ask OCCT about it. Once we know the circumstance THEN we can know if it's algorithms/interface/order of operations or whatever. Sometimes it is an easy fix but we don't know until we know.mfro wrote: ↑Wed Mar 22, 2023 9:24 amNot sure if that's the culprit. Your are definitely right regarding the history, but on the other hand, the remaining issues with OCCT (at least that's my personal gut feeling) are a bit different: both problems mentioned above (fillets and thickness) have documented weaknesses. To me, hat seems to imply these aren't just "simple" bugs, but serious shortcomings in the algorithms that are well known to the developers. No easy fix.keithsloan52 wrote: ↑Tue Mar 21, 2023 10:47 pm Re what we need in relation to OCCT is better way of...
- * Capturing OCCT problems
* Being able to create simple test case to recreate the problem
* Good interface with OCCT developers
Working on...
https://github.com/iplayfast/CycloidGearBox
https://github.com/iplayfast/CycloidGearBox