Discussion: FreeCAD is not ready for 1.0

[carlopav]

I have to say that I kinda agree with some aspects of the @Zolko statement.
As a small code contributor I feel quite disoriented by the development direction and i'd like to see more discussion on that, or a clearer shared statement on the way we are going, and about which steps are necessary to get there... and so version numbering could come after that.
Of course no problem with "it's done when it's done" i'm a volunteer too and i'm fine with that, but I hope in a clearer shared direction.
In a certain sense, it's more about building a stronger community that recognizes and share some goals, than aiming for one feature or the other. And I believe that this community can be stronger if the decision making process is more transparent, even if in the end will be @wmayer to decide.
Happy coding to everyone

[Zolko]

I – and I think all professional engineers will feel the same – cannot invest a big project on a tool decided by unknown people on their whim

And do engineers know personally the people behind a commercial CAD system

no, and that's the goal: people come and go, organisations remain. I don't care what the names of the developers of LibreOffice, Gimp, KDE or Debian are. What I care to know is that "they" (whoever that is) will organise in a way that key people are replaced and the software is maintained.

Another point -- as several times discussed in the forum -- is the vendor lock-in

You're right, and that's also my reason to use FreeCAD, as I have told many times.

nobody can force them to implement certain features [...] they have the possibility to directly pay a developer to implement it.

This is mutually exclusive, and the exact problem I try to report. You cannot have it both ways: nobody will pay for a feature that might get stuck in limbos for years, even though it's ready in a development branch, without any reason, discussion and feedback. And without any person to address-to. If a company pays for a development, they want to be sure that it will be shipped when it's finished.

[wmayer]

This is mutually exclusive, and the exact problem I try to report. You cannot have it both ways: nobody will pay for a feature that might get stuck in limbos for years, even though it's ready in a development branch, without any reason, discussion and feedback. And without any person to address-to. If a company pays for a development, they want to be sure that it will be shipped when it's finished.

Well, you might be right if the company's interest is to have integrated this feature into the official FreeCAD project. But a possibility is that this feature will be implemented inside a new and independent module/workbench whose source code does not even have to be published.
And if the company uses the modified software in-house only (i.e. not to sell the product) then they are even allowed to modify the code of the core system without publishing the changes (see https://www.gnu.org/licenses/gpl-faq.ht ... stedPublic).

[Zolko]

And if the company uses the modified software in-house only (i.e. not to sell the product) then they are even allowed to modify the code of the core system without publishing the changes

but that would defeat the entire open-source argument then. Again, you cannot argue both ways.

[wmayer]

And if the company uses the modified software in-house only (i.e. not to sell the product) then they are even allowed to modify the code of the core system without publishing the changes

but that would defeat the entire open-source argument then. Again, you cannot argue both ways.

Why? It's not uncommon that a company who pays a developer will get the source code at the end. In this case from the company's point of view it doesn't matter if the code won't be published or not as long as it's handed over. Of course, the next question is if the code will be maintained as otherwise it will become unusable after a few years but that's a different story.

[carlopav]

And if the company uses the modified software in-house only (i.e. not to sell the product) then they are even allowed to modify the code of the core system without publishing the changes

but that would defeat the entire open-source argument then. Again, you cannot argue both ways.

Why? It's not uncommon that a company who pays a developer will get the source code at the end. In this case from the company's point of view it doesn't matter if the code won't be published or not as long as it's handed over. Of course, the next question is if the code will be maintained as otherwise it will become unusable after a few years but that's a different story.

I think the point rather is... are we pro or against a sort of FreeCAD organization like a FreeCAD foundation? (i guess we are pro @Zolko, since several steps have been already taken in that direction by kurt and yorik with "FreeCAD partnership with Software Freedom Conservancy", is it?) and that's a good argument to me for 1.0 for example.

[Joel_graff]

but that would defeat the entire open-source argument then. Again, you cannot argue both ways.

A couple points to consider here - first, "open source" and "free software" (GPL) are inherently different things. Open source exists because free software licensing wasn't broad enough to allow for commercial adoption in the mid-late 1990's. Second, I'd agree no company would pay for development of a feature if the community refuses to maintain it, presuming the company is not willing to maintain their own fork. It's probably common enough, but whenever I see companies paying developers to work on an open source project for their own benefit, they're usually maintaining their own fork, too, because the features are proprietary. Really, the argument fails *ony li* the company paying for development expects their work to be upstreamed without vetting and commitment from the community. That's just how it works.

Often times, I've seen commercial code drops under a free / open license just to meet the legal requirements of the licensing, without regard to whether or not the features are upstreamed or maintained separately. It's a loophole / weakness in the licensing and community model, I think, but that's another topic.

[drmacro]

Just me thinking out loud....

It might be worth having a conversation with Paul Davis. He is the original developer (and still lead, I believe) of the open source DAW software called Ardour.

It is a massive project and he has some how managed to balance having a commercial relationship with Harrison Consoles and having Ardour branded as their Mixbus product while having both Ardour and Mixbus benefit from the core development.

Not suggesting modeling FreeCAD in this way, just thinking another perspective from the open source world might be food for thought.

https://ardour.org/

https://harrisonconsoles.com/site/index.html

[Zolko]

A couple points to consider here - first, "open source" and "free software" (GPL) are inherently different things.

and what about LGPL ?

they're usually maintaining their own fork, too, because the features are proprietary.

looks like someone has listened to you:

Parametric CAD modeler – CAD Builder from Open Cascade
https://www.opencascade.com/content/op ... d-modeler

what do you bet how much FreeCAD code is in there ? Which is only fair, since they have invented (and open-sourced) OCC to begin with. They probably have taken a proprietary Qt licence, Python is free, whether they went with Coin3D or some other scene-graph would be funny to know.

They have spent some money for nice icons though.

[Joel_graff]

and what about LGPL ?

From what I understand, Stallman hates the LGPL because it leaves out a key feature of the GPL - requiring all software linking it to also be GPL. That's the 'viral' nature of the GPL that so many commercial vendors despise and it's the key difference between open source and free software.