[Discussion] Defining core workbenches

[vocx]

While looking at the FreeCAD wiki, I realize that it lists all workbenches in the main page. I don't see a point in listing things like Complete or Start or Web together with other workbenches. It adds clutter to the software itself and to the documentation.

Has it been discussed defining a set of "core" workbenches? I would choose Draft, Arch, Sketcher, Part, PartDesign, and TechDraw. These seem to me the most useful and feature rich, to start working on CAD modelling. A new user should be directed to learning from these workbenches about lines, 2D figures, 3D shapes, bodies, pads, extrusions, features, boolean operations, etc.

Maybe Arch shouldn't be considered essential, but I'm including it in this group because it builds on top of Draft and provides a feature rich set of tools that can be used quickly to work on buildings.

Then we'd have secondary workbenches, which are sufficiently usable, but focus on specific niches, like Path, Raytracing, Robots, Ship, Curves, FEM, etc. I'm not saying these aren't important, but I think a user should be familiar first with creating a solid object, with Part, PartDesign, or Arch, before trying to use FEM analysis or something like that.

And then there are support workbenches which really cannot be considered tools for modelling, like Start, Web, Test, Complete, Image, Plot. These should be in a third category.

I personally think the interface and documentation should be simplified to direct the user to the core workbenches first. It would help them progress faster in their discovery of FreeCAD.

[sliptonic]

Then we'd have secondary workbenches, which are sufficiently usable, but focus on specific niches, like Path, Raytracing, Robots, Ship, Curves, FEM, etc. I'm not saying these aren't important, but I think a user should be familiar first with creating a solid object, with Part, PartDesign, or Arch, before trying to use FEM analysis or something like that.

You're showing a bias toward your own workflow. A lot of users that show up to use Path, FEM, and other workbenches may not need or want the modeling aspects first. We see many users using other 2D and 3D tools to prepare a model but wanting FreeCAD's Path workbench because it's the best CAM available in the open source world.

[vocx]

You're showing a bias toward your own workflow. A lot of users that show up to use Path, FEM, and other workbenches may not need or want the modeling aspects first. We see many users using other 2D and 3D tools to prepare a model but wanting FreeCAD's Path workbench because it's the best CAM available in the open source world.

I am definitely biased on this, precisely that's why I think it's a matter of discussion. If Path is sufficiently stable, and sufficiently feature complete, and it's really very much used for manufacturing because it's the best alternative out there, then I think it makes sense to include it as a core workbench.

However, I ask you, what is what you think when you hear the word "CAD"? CAD to me implies creating a model from scratch, either extruding, padding, drawing, cutting, etc., something like that. I believe the core functionality of FreeCAD is more or less in this sense. Other functionalities can be added on top, which is great, but may be regarded as secondary objectives.

[sliptonic]

However, I ask you, what is what you think when you hear the word "CAD". CAD to me implies creating a model from scratch, either extruding, padding, drawing, cutting, etc., something like that. I believe the core functionality of FreeCAD is more or less in this sense. Other functionalities can be added on top, which is great, but may be regarded as secondary objectives.

The 'D' is certainly 'Design' but what's so special about the base model? Path uses the geometry kernel to design and display tool paths. Those tool paths are just lines and arcs - geometric elements created by extruding, offsetting etc.

But what's the value in defining 'core' workbenches? If the goal is to prioritize documentation or guide the refinement of the UI, then personal preferences shouldn't matter. We would need hard data about user behavior.

[DeepSOIC]

I don't see a point in listing things like Complete or Start or Web together with other workbenches. It adds clutter to the software itself and to the documentation.

Don't you think, if one came reading help, one wanted to know what the heck that strange "complete" workbench really is. IMO, list them all. Nearby, you may give suggestions, like "If you want to learn how to model a part, go learn PartDesign and Part; if you want to model a building, learn Draft and Arch; if you want to make a 2d-drawing by hand (autocad-style), learn Draft" and so on.

[sgrogan]

I don't see a point in listing things like Complete or Start...

Start is important. It's the first thing a user sees in a default installation.

IMO, list them all

I agree, if the Workbench is available by default, it is core IMHO. A separate discussion would be if some workbenches should be available by default/included in FreeCAD master. This is an ongoing discussion.

[chrisb]

For me the core is Sketcher, Part, PartDesign plus a few tools from Draft. And even these could be boiled down to Sketcher plus PartDesign or Part. These can be used to create 3D models.
Around these there are workbenches which use the models, like Path, FEM, TechDraw.

Others may use Draft and Arch instead of Sketcher and PartDesign.

To explain this is important but I don't think that it justifies complete restructuring of workbenches and documentation. FreeCAD is several specialized CAD systems in one and different users pick different pieces from it. So we may need different "user stories" for e.g. mechanical engineers, architects, etc.

And, of course, we should get rid of complete workbench, at least in its current state.

[vocx]

Don't you think, if one came reading help, one wanted to know what the heck that strange "complete" workbench really is. IMO, list them all. Nearby, you may give suggestions, like "If you want to learn how to model a part, go learn PartDesign and Part; if you want to model a building, learn Draft and Arch; if you want to make a 2d-drawing by hand (autocad-style), learn Draft" and so on.

Of course they should be listed all. I'm not saying we should remove the documentation. We should just put it apart from the other workbenches, because evidently Complete does nothing. It doesn't help you at all to design a part or a piece. It is just there for some historical reason. So, my idea is to have

• core workbenches,
• secondary workbenches, and
• additional workbenches (administrative, auxiliary?).

The way the FreeCAD program and documentation currently show the workbenches is like this, "Oh, wow 24 workbenches! That's a lot, where do I start? Oh, Start doesn't do anything; oh, Complete also doesn't do anything; oh Reverse Engineering also doesn't do anything; oh, Test doesn't do anything; oh, Spreadsheet is just a spreadsheet, it doesn't help me model a piece, I wonder why they included it there."

It'd be simpler to have a core of 5 to 7 workbenches which a user can say, "oh, this is definitely where I should start, with the Part (or PartDesign, or Sketcher) workbench."

[vocx]

I don't see a point in listing things like Complete or Start...

Start is important. It's the first thing a user sees in a default installation.

Start does not really contribute to the main objective of FreeCAD which is modelling 2D or 3D parts. That's my criterion to call it "important". It's basically a support workbench. You can configure FreeCAD to show another Workbench by default, so if a user does this, he or she won't see Start any more in their life.

IMO, list them all

I agree, if the Workbench is available by default, it is core IMHO. A separate discussion would be if some workbenches should be available by default/included in FreeCAD master. This is an ongoing discussion.

Again, it's not about removing the information of what a current workbench is about. It's about deciding which are core, which should be presented as such, and which aren't that important, so they could be shown in a separate tier.

Imagine I have a table, first row are the core workbench, and second row, the rest. It's just that.

[vocx]

...
But what's the value in defining 'core' workbenches? If the goal is to prioritize documentation or guide the refinement of the UI, then personal preferences shouldn't matter. We would need hard data about user behavior.

Yes, that's my idea, defining the core workbenches would help refine the documentation and the interface. It'd maybe also help the packaging, so you could install freecad-core, freecad-extras, freecad-experimental, or something like that.

Personal preferences always matter because every single one of us has a personal preference. Of course, it's better to have a consensus. I have the feeling many of us would consider Draft, Sketch, Part, PartDesign, to be pretty "core" to the mission of FreeCAD. Then the rest is up for debate. I have the impression that you'd nominate the Path workbench as a core functionality. I have no objection. Fine by me. If most people don't like Arch as a core workbench, I would oppose, but I'd be fine with whatever the majority would decide.