Merging hamish2014/assembly2 into v0.17, help wanted

[jmaustpc]

Jim, I am happy to change the licence to LGPL2.
How exactly do I change the licence,
do i simply update the licence in the hamish2014/assembly2 repo?

Hi Hamish,
that is a good question, I don't really know in detail. However my understanding is that the code you wrote, you own the copy right and can offer in any other licence you like. I think all you have to do is change the licence files/comments in the code and repo. This becomes more complex though because you have accepted pull requests from others. You can do what you want with your code but you have to ask them about their contributions. I think that you only have to ask then if their input was not "very minor or insignificant" but you would have to research that to see if I am right and if so to what extent or what details may apply.

Obviously you can choose whatever licence you want to and may have had good reasons to choose gpl3 but gpl3 was controversial when it was created, the Linux kernel and many projects refused to accept it. Choosing it can be problematic and doing so has been the death of some otherwise great projects.

I hope you find the time to come back and work on FreeCAD master code sometime in the near future. Life has a way getting in the way for a few of us at times. I have also been very busy and been forced to dramatically cut back on the time I can spend on FreeCAD over the last year or so.

Jim

[triplus]

If @DeepSOIC and potentially other developers will implement needed changes to make Assembly 2 compatible with FreeCAD 0.17 and going through the features to see if they are of the needed quality i would vote for merging it into master for FreeCAD 0.17.

Listening reasoning on how licence is a issue (after it was expressed complying with FreeCAD license policy isn't a problem). And how something is a hack and something else not. It's a Python module and it can be dropped from FreeCAD at any point in the future by simply removing it from the code base. And until C++ counterpart is made this is it. This is current FreeCAD assembly module whether somebody likes it or not.

[ickby]

It's a Python module and it can be dropped from FreeCAD at any point in the future by simply removing it from the code base. And until C++ counterpart is made this is it. This is current FreeCAD assembly module whether somebody likes it or not

No, you never can drop it easily once it was in master, it would break old files which is not acceptable.

[triplus]

It's a Python module and it can be dropped from FreeCAD at any point in the future by simply removing it from the code base. And until C++ counterpart is made this is it. This is current FreeCAD assembly module whether somebody likes it or not

No, you never can drop it easily once it was in master, it would break old files which is not acceptable.

PartDesign NEXT and TechDraw did just that. There was Python based spreadsheet expression support before the support we now have in production release. End users use Assembly2 and Drawing Dimensioning module regardless. I rememberer using a macro in the past that stop working as some administrative change was made in drawing workbench...

Therefore i feel that best we can do is to plan when we change something that brakes the support for old files. But every few years we will and do brake the support for old files. Having a compatibility matrix FreeCAD version <-> File Format is likely the best we can do. One of the main reasons why i haven’t migrated to FreeCAD 0.17 is i need to finish some projects first. And this projects will i guess always need to be opened in current production release (FreeCAD 0.16).

Merging hamish2014/assembly2 into v0.17 after adding FreeCAD 0.17 related changes to it and evaluating code quality seems like a sensible thing. Users get assembly capabilities by default and developers can exchange ideas on what is the proper way to implement something if it currently considered a (working) hack.

[sgrogan]

Thanks for the interest DeepSOIC.
Jim, I am happy to change the licence to LGPL2.

Just to make it clear: I don't plan merging assembly2 into FreeCAD. Just porting (taking advantage of containers, fixing bugs) and maybe improving.

Wouldn't this satisfy everyone's concerns going forward? Things stay pretty much like they are today, hamish's work doesn't die. I'm interested what DeepSOIC's fertile mind comes up with, unfettered by a merge to master.
@Yorik has pointed out the advantages (and drawbacks) of a 3rd party WB model in the past. I think here and in the related thread viewtopic.php?f=17&t=16590 we have both sides to the story.
Anyway just my 2 cents. Thanks to hamish for the work and DeepSOIC for the interest to continue it.

[triplus]

Wouldn't this satisfy everyone's concerns going forward?

It depends. If it turns out a proper solution i imagine it will be done by the time basic C++ assembly module will be available. Then we will have the whole drawing dimensioning dilemma all over again.

We are probably trying to skip steps. That is going from 0 to proper and the final solution in one giant step. Hamish started Assembly project in 2014 and i imagine it could take a year or two to get to basic C++ assembly module. That is i guess 4 years or more without any assembly support by default in FreeCAD compared to some assembly support in FreeCAD by default for all this years.

[yorik]

Wouldn't this satisfy everyone's concerns going forward? Things stay pretty much like they are today, hamish's work doesn't die. I'm interested what DeepSOIC's fertile mind comes up with, unfettered by a merge to master.

Me too!

[NormandC]

I'm interested what DeepSOIC's fertile mind comes up with, unfettered by a merge to master.

Uh-oh! The image that comes to mind...

[hamish]

Thanks everyone.
I am going to contact all the contributors to assembly2 about changing the licence to LGPL2.
This should help to keep all options open...

[hamish]

https://github.com/hamish2014/FreeCAD_a ... issues/120