FreeCAD - is it just me, or bug city?

[max_torque]

Ok, i've been playing with freeCad.  It seems, considering it's free, to be pretty well featured, but there seem to be soooo many bugs it's pretty much un-useable. Have others found this, or is it something i'm doing wrong?

For example, i do a sketch, fully constrain that sketch, and then sometime later, after some further operations, it's reports an error with that sketch and it now has excess or insufficient constraints suddenly.  This is especially true if you try to make an even slightly parametric model, ie your constraints are based on other geometry!

Sometimes, if i save and restart, all is well again for a bit, but sometimes, say something pocketed out using the sketch geometry is now different.  It also seems just about impossible to go back into the lower items in the tree and change anything with messing everything up that's been done after that point.  What's the point in parametric design if you can't change anything??

 

[ebclr]

Do you know that you can have fusion360 for free is some circumstances, if you fit the criteria is a nice idea to take a look

https://www.autodesk.com/company/legal-notices-trademarks/terms-of-service-autodesk360-web-services/autodesk-web-services-entitlements

[free_electron]

fusion360. almost better than solidworks ! and zero cost.

For 2d work : Dassault Draftsight. Kicks autocad out the door, across the street, into the field where it lands in a pile of cow-poo (where it [autocad] belongs).

[NivagSwerdna]

Ok, i've been playing with freeCad.

I used QCad for 2D and Fusion 360 for 3D. Simples.

[max_torque]

Sorry, for the late reply!  (i've been away)


Well i'm giving up with FreeCAD.  It's fine for really simple stuff, but unless one can reliably go back and change dimensions in a sketch and have that then cascade down through the model correctly i can't see how it' s any use for real designing (where it's inevitable as the design progresses that initial dimensions will need to be changed)

I tried to do a simple tubular frame today, and when i changed a tiny detail on sketch, part of the frame dissipated completely and another bit suddenly appeared on a completely different plane to where it had been....

Fusion360 however looks excellent judging by the various tutorials i've been watching on UUUUUUUtube! 

[janoc]

Just beware of random breakages with Fusion360. Yesterday I had to fix some stuff because the last update broke some APIs at work. With the forced updates Autodesk treats Fusion360 users as unpaid alpha/beta testers for things that may appear in Inventor or their other CAD products. It is not something I would rely on for anything serious.

It is also cloud-only - no internet, no play, unless you have the foresight to cache your files locally. Oh and it is slow as molasses with anything just a little larger. We are using it at work to open CAD files from clients and it can easily bring even a fairly decent machine to its knees with a small model ... Heck, even the completely in-browser Onshape is not that slow and doesn't need hundreds of megs of updates every few days.

Re FreeCAD - I feel your pain. I have found the more recent versions a bit more usable, but I have still managed to make it go nuts with some changes in the sketches. It is usually possible to find a way to draw the object I want that avoids the problem, though. Of course, one shouldn't have to do that.

The current stable version also lacks any support for assemblies (joints and such), which make it difficult to use. Assemblies are in the development version, but that is not for the faint of the heart ...

If anyone knows a good hobbyist-level CAD that doesn't cost arm and leg and can run in Linux too, I am all ears ...

[Mechatrommer]

fusion360. almost better than solidworks ! and zero cost.
For 2d work : Dassault Draftsight. Kicks autocad out the door, across the street, into the field where it lands in a pile of cow-poo (where it [autocad] belongs).

you dont compare 2d only drafter with something that is 3d capable (albeit nonparametric) like autocad, heck any descent programmer can build his own 2d drafter easy in the home, not so easy when it comes to 3d modelling and operations... being a autocad user for many years, i've been having a hard time to learn new notions of modern parametric 3d modelling like solidworks, inventors and i guess the same as fusion360... any 3d objects pretty much can be modelled in autocad except anything involving fancy extrudes, revolves and lofts... btw, i can see your love to the free tools, since in one angle you are supporting autodesk (fusion360), and from another angle are not (autocad)

anyway, back to OP's comment about being useless, i dont know about freecad 3d operations maybe its full of bugs, but its been very usefull for me to convert autocad iges model to altium step model (with coloring capability), i dont know what is better for the purpose... ymmv...

[JPortici]

fusion360. almost better than solidworks ! and zero cost.

For 2d work : Dassault Draftsight. Kicks autocad out the door, across the street, into the field where it lands in a pile of cow-poo (where it [autocad] belongs).

I like and use QCad/Librecad Reminds me of the simpler days when they taught us R14 in high school... and there was none of the bullshit introduced since autocad 2008 or so

[sleemanj]

I've used FreeCad for the odd job where scripting was necessary, but  other than that it's just horrid compared to OnShape or Fusion360, not very stable, not very intuitive, not very easy.

OnShape is my preference, partly because as a Linux user I don't have to go start a Windows instance to use it, but also I just prefer it's way of doing things over f360.   

Fusion360's  not-always-quite-a-feature model (some things are not proper features in the timeline and can't be edited or rolled back to properly), and it's clumsy  (and slow) not-cloud-but-not-local-either file management makes it a pain on occasion.

Of course, OnShape's pricing model does mean all my designs are public, that's not something I care about, often enough I want them to be anyway.

[b_force]

You can also still use an older version of Sketchup (version .

It's to unfortunate though.
A lot of free/open source programs have heaps of potential, but it's just simply not usable for anything serious.
When will they learn, first make the interface rock and rock solid, than work on your fancy stuff??
(this is especially true fro KiCad as well)

I was considering FreeCad for some simple 3D drawings, but now I am hesitating.
(I used to work with CoCreate a lot)

[janoc]

You can also still use an older version of Sketchup (version .

It's to unfortunate though.
A lot of free/open source programs have heaps of potential, but it's just simply not usable for anything serious.
When will they learn, first make the interface rock and rock solid, than work on your fancy stuff??
(this is especially true fro KiCad as well)

I was considering FreeCad for some simple 3D drawings, but now I am hesitating.
(I used to work with CoCreate a lot)

The problem with free/open source programs is the manpower. When people do unpaid work on these projects on their own dime, they obviously want to work on things that interest them. UIs are both unsexy to work on and an opinion minefield where you can never satisfy everyone.

Cut them some slack. I much more bugged by a horrendous user interface in Fusion360 or something like Visual Studio where there is a large team and a ton of money behind it than something like FreeCAD which is essentially developed by 2 people in their free time only or Kicad which is over 20 years old project and was developed by a single person until 2008 or so.

It is not apples to apples comparison. OTOH, there are plenty of very polished free/open source projects - but they have also a lot more resources behind them - e.g. Firefox, LibreOffice, Blender etc.

[free_electron]

and from another angle are not (autocad)

I have never liked autocad. it is a horror contraption that talks with a lisp (pun intended) that produces data in a crappy fileformat ( DXF/DWG the file format form hell incompatible with itself ) it was crap on a 5 1/4 inch floppy on a 8086 and it is still crap. They kept building on crap.

Fusion360 works really well. I use Solidworks and Catia daily and like Fusion better than both.

[b_force]

The problem with free/open source programs is the manpower. When people do unpaid work on these projects on their own dime, they obviously want to work on things that interest them. UIs are both unsexy to work on and an opinion minefield where you can never satisfy everyone.

Cut them some slack.

In my opinion it's about priorities. So get the basics done first and make sure it's (as good as) bug free.
To make sure that most people are satisfied, make an interface that can be customized. This is what you typically see in most professional programs. Were the user can practically change everything.

In fact I am working on some software myself with some people (for acoustics). All the way on the top in the priority list are; GUI first, gui first and gui first
Only after you nailed the gui, work on the fancy graphics.

Ones again, it's not blaming developers (cutting slack), but like I said, it's very frustrating to see things with potential failing on relative simple things.
See this more as a piece of advice.

[janoc]

and from another angle are not (autocad)

I have never liked autocad. it is a horror contraption that talks with a lisp (pun intended) that produces data in a crappy fileformat ( DXF/DWG the file format form hell incompatible with itself ) it was crap on a 5 1/4 inch floppy on a 8086 and it is still crap. They kept building on crap.

Welcome to the world of AutoDesk

The only software from them that somewhat works is the one that wasn't developed by them but they have acquired it from elsewhere (3ds MAX, Maya, Eagle, Fusion was also an external project, AFAIK). Then you just keep praying that they won't bastardize/"autodesk-ize" it too much - Max and Eagle are good examples of that happening

I have also written some add-ins for Fusion360 for CAD data preprocessing at work and I can tell you, the API and the bugs are pretty insane ...

And re Lisp - actually the Lisp scripting in AutoCAD was a groundbreaking idea at the time, same as writing extensions for the software. Yes, the Lisp is strange for people used to more "pedestrian" languages like C but at the time they have introduced AutoLisp in 1986, there weren't really scripting options like we have today. It was also the peak of the AI and Lisp popularity, so getting Lisp developers was relatively easy. And once the thing got entrenched, it became next to impossible to remove it and replace it with something else - it would break all the backwards compatibility for extensions and CAD documents.

E.g. 3ds Max which is much newer than AutoCAD uses something called MaxScript - an absolute abomination when compared to AutoLisp.

[b_force]

What autocad is for 2D, I have it with Solidworks.
Just can't get used to the interface.
The most annoying thing is it's almost the de facto standard in mechanics.
If you want to get anything done or supplied, people ask for Solidworks

[janoc]

In my opinion it's about priorities. So get the basics done first and make sure it's (as good as) bug free.
To make sure that most people are satisfied, make an interface that can be customized. This is what you typically see in most professional programs. Were the user can practically change everything.

That's the problem. Tools like KICAD or FreeCAD were not started/written with the intention to become professional tools or to satify the needs of more users than the author of the software. So once the projects got large enough and useful enough for being of interest to the larger community, there was a technical debt to pay.

That's not an excuse but people tend to ignore that large projects like this weren't developed overnight and that there are years of technical decisions and reasoning behind them which are the reasons why they work (or don't) in certain ways.

If you are designing something from scratch and you know that your objective is to sell it or to be used by a large community of people, you design things differently. However, when you are working on a 25 year old codebase that you have "inherited" from someone else (KICAD), it is a different story.

Of course, that also assumes that the developer is actually an experienced software developer - e.g. the KICAD main author was a French university professor in an unrelated field (electrical engineering & image processing), not a software engineer or UI designer. He made it work well enough for what he needed to do - capture schematics and design boards.

In fact I am working on some software myself with some people (for acoustics). All the way on the top in the priority list are; GUI first, gui first and gui first
Only after you nailed the gui, work on the fancy graphics.

I thought one would first work on getting the basic functionality right, not the UI. That doesn't mean the UI isn't important (and for some applications the UI is the application!), but beautiful UI is of little use when the code behind isn't doing its job ...

Ones again, it's not blaming developers (cutting slack), but like I said, it's very frustrating to see things with potential failing on relative simple things.
See this more as a piece of advice.

Yeah, that's true. It irks me too. On the other hand, it is not the worst issue one can have  The programs are usable for what they were designed to do, working and free.

[Mechatrommer]

whats frustrating is that when the developer ask for $1 for the software license, it will go to the bottom of recommendation list. 1st in the list will be free tools, no matter how crappy they are, and then come the people who never satisfy mourning and complaining. 2nd in the list is multi millions softwares, but then only elites can afford those... so the ecosystem is sooo "encouraging" for the very little developers who want to make their small way in the business, i guess knitting will give better revenue... as of fusion360, i dont see it as purely a free tool, sooner or later you are going to be hit by something, when you are in love and tied too much to the tool, such as when you are not a student anymore.

[janoc]

whats frustrating is that when the developer ask for $1 for the software license, it will go to the bottom of recommendation list.

$1 for a license is just dumb, the hassle with collecting the money (taxes & all) is likely going to cost more than what such developer will make. Either make the software fully commercial or leave it free (as in beer - and get something else to pay your bills). Such half-assed approach doesn't help anyone.

Also, do people actually buy CAD tools based on some recommendation lists made by download sites and such? That's perhaps true for small utilities and tools that are dime a dozen but not really much else.

However, the developers have only themselves to blame for this mess - with the mass jump on the "freemium" bandwagon especially on mobile, where unless you have a free tool and then try to milk and dime the user on other things (ads, microtransactions, etc) you don't exist, the users learned to expect "free" (as in beer). Why would you pay when you can get (supposedly) the same thing for free?

Also many of those supposedly free programs in those lists are not really free but basically time limited demo versions that you must buy after e.g. 30 days.

However free (as in freedom)/open source software (Kicad, FreeCAD, etc) is not really in the same category - they neither rely on such recommendation lists (+ those rarely cover anything but Windows/Mac software anyway) nor rely on such business model to survive.

as of fusion360, i dont see it as purely a free tool, sooner or later you are going to be hit by something, when you are in love and tied too much to the tool, such as when you are not a student anymore.

Fusion360 is not free. It is a subscription based software (and quite expensive at that). It is free only for hobbyist/small startup (making <$100k income) use with a free 1 year license (which you can renew).

E.g. our company (a 7 person startup) doesn't qualify for the free license anymore and we had to take a subscription on it ($300/year single seat, the Ultimate version with better simulation and CAM is $1500/year single seat). Granted, it is cheaper than buying SolidWorks or Inventor outright but hardly free. The only free part are the free bugs whenever it updates itself.

[Mechatrommer]

i meant $1 is analogy of 'cheap', it can be $15, $100 etc cheap,and also of course including free 30 days or 1 year or whatever demo version...recommendation list i meant is not something properly written somewhere by pro, but words that lingering around the net esp forum like this and some 'unqualified in the area' content provider... if a developer can get $1500 per anum per user, then there is no excuse/tolerance for bugs, thats where the mourner and complainer should go if anything... as for free tools, if the maker want to see their invention to go, they need to rectify, otherwise its a no go, but rectification is not an obligation since he get nothing from users (except complains and mourns), we cant blame him much if he go fishing while we in the middle of battle trying to save the world.

[b_force]

Of course, that also assumes that the developer is actually an experienced software developer - e.g. the KICAD main author was a French university professor in an unrelated field (electrical engineering & image processing), not a software engineer or UI designer. He made it work well enough for what he needed to do - capture schematics and design boards.

I don't think that has to be true.
You can have wonderful ideas, but not doing the actual work.
So you will be the team leader and the experts do the bits and pieces.

In fact, that's exactly what I am doing as well.
I am not a full time programmer (far from it), but I know the basic concepts.
So I use other people to translate these concepts to something that works.

(in fact we are still looking for programmers)

[janoc]

I don't think that has to be true.

No, but it certainly was the case for KICAD. A lot of the UI and workflow strangeness that KICAD was known for was due to the original design decisions of its (then) sole author. When you are alone on the project, you can't afford to "have wonderful ideas and have others implement it" because you are the only person who can actually do the work. So you both prioritize and some things will be less than ideal by sheer necessity - nobody is expert in everything.

[Karel]

Just beware of random breakages with Fusion360. Yesterday I had to fix some stuff because the last update broke some APIs at work. With the forced updates Autodesk treats Fusion360 users as unpaid alpha/beta testers for things that may appear in Inventor or their other CAD products. It is not something I would rely on for anything serious.

Like Eagle 8 users...

https://forums.autodesk.com/t5/eagle-forum/eagle-8-3-2-eagle-exe-has-stopped-working/td-p/7368957

[free_electron]

That's the problem. Tools like KICAD or FreeCAD were not started/written with the intention to become professional tools or to satify the needs of more users than the author of the software. So once the projects got large enough and useful enough for being of interest to the larger community, there was a technical debt to pay.

Add to that: they were built by programmers that don't understand the root problem of making a board. It is still considered as a polygon pusher. give 'em some drawing routines , store vectors and make some and/or/xor operations and off we go. -FAIL-
Everything else is an afterthought. "but it's all connected with according to schematic" . yup , and we all have seen the kind of schematics they produce ... it's amazing the router technology is not merely converting the flylines into copper. then again ..

[b_force]

I don't think that has to be true.

No, but it certainly was the case for KICAD. A lot of the UI and workflow strangeness that KICAD was known for was due to the original design decisions of its (then) sole author. When you are alone on the project, you can't afford to "have wonderful ideas and have others implement it" because you are the only person who can actually do the work. So you both prioritize and some things will be less than ideal by sheer necessity - nobody is expert in everything.

I guess you're right.

It wouldn't be my approach, but it all depends.
If someone is thinking longterm you can better have a proper looking interface, simply because it's much easier to show for potential investors etc.
It makes or brakes a program.
You can have wonderful things, but if people don't use it because it's to buggy or simply doesn't work well, it's still worthles.
Than it's just a great idea idea sitting on a shelf.

It's to bad that most developers (also in electronics) don't realize this concept.

[Mr. Scram]

Do you know that you can have fusion360 for free is some circumstances, if you fit the criteria is a nice idea to take a look

https://www.autodesk.com/company/legal-notices-trademarks/terms-of-service-autodesk360-web-services/autodesk-web-services-entitlements

Putting all your data in someone else's basket and potentially losing your IP and infrastructure overnight doesn't sound attractive.