Better direct editing

caradon

As much as I like the behavior of direct editing in Onshape, I can't help but feel direct editing is underplayed.
I think it should get a more prominent role. Especially with regard to working on imported geometry.

Things I would like to see:

• A history-free direct editing environment. Note that history-free doesn't mean you can't parametrize models. If I can save a version after initial upload of a 'dumb model' and save a version after some edits, then why would I need a history (of many times arbitrary features)? I think the history is a hindrance when doing early concepting work and design reviews. Especially when non-CAD experts are involved and using Onshape too.
• More intelligent direct editing tools. Not just offsets, not just absolute translations/rotations. I want to (re)specify diameters. I want to translate faces upto or with an offset relative to other faces. I want to rotate faces at an angle relative to other faces.
• More robust direct editing behavior. Being able to cope with geometrically more challenging conditions. E.g. Onshape direct editing can't reapply conic fillets. In Fusion 360, direct editing (I tested Draft and Replace face) is able to handle curvature continuous fillets.

Many of these points resemble a typical SpaceClaim workflow. I'm not saying 'direct mode' Onshape should mimic SC 1:1, but the fact is SC is a great tool for some tasks.
What do you think?

Dries

pete_yodis

@DriesV I would agree. I think this would be especially important for working with all the imported data that Onshape might eventually take over from existing data management and CAD platforms. I could see the real draw with Onshape initially being the platform for managing design data. Having a CAD tool with this platform that is able to make quick edits to data that came from other tools, would be really nice. For example, I have a casting model that I need to tweak a wall location or fillets or draft, etc.. and it was built in another tool - Creo, SolidWorks, SolidEdge.. it wouldn't matter. You could make the edit to that geometry and keep moving - very SpaceClaim like. There might not be a need to keep the old programs around anymore if this could be accommodated.

caradon

I just think there is a viable use scenario for a history-based, a well as a non-history-based approach.
I think Onshape needs to support both modes. Fusion 360 has both modes. Solid Edge and NX too.
For me it's not SpaceClaim or SolidWorks. It's SpaceClaim AND SolidWorks. Both approaches have value in different situations.
One of the great appeals of Onshape (for me) is to get many stakeholders involved in the early stages of design. Let's be clear: give a history-based model to non-CAD experts to make changes and you will probably get a cr*p feature tree.

Dries

jakeramsley

DriesV said:

• I want to translate faces upto or with an offset relative to other faces.

Is the behavior you want different than using an offset off of a replace face?

caradon

JakeRamsley said:

DriesV said:

• I want to translate faces upto or with an offset relative to other faces.

Is the behavior you want different than using an offset off of a replace face?

As far as direct editing goes, what I want is direct manipulation on the model. No menus.
Replace face with/without offset is a great feature. However, its application and usefulness is not intuitive or clear to the novice/occasional user. Even in SolidWorks, Replace face is awkwardly misunderstood/unknown. So far, Onshape doesn't do a much better job in making it more accessible. So replace face, as an alternative to 'translate upto', doesn't fly from a usability point of view.

My ideal direct editing workflow:
Just a single badass graphical direct manipulation tool (read: triad).

• Translation, rotation, offsetting... faces
• Pick references for translating, rotating, offseting relative to other faces
• ...

I think the current direct editing features in Onshape are great, especially in a history-based parametric model.
For pure direct editing -especially on imported geometry- I would love a more graphical manipulation tool. And not necessarily a recorded history.

IMHO.

Dries

billy2

Dries I think you are absolutely right about direct modeling and how it needs to be brought forward in the design process/ manufacturing process. Parametric models are impossible for any one except the originator to understand and therefore down stream departments that have to use this engineering data are lost and rendered non-functional due to it's complexity.

I was hoping to release (branch) a version of the engineering design for manufacturing and hopefully strip out all the parametrics and leave them with only geometry. Just give them the geometry and a limited set of change tools (direct modeling). Hopefully this would be adequate for them to control their processes and get things made by controlling a very simple model using simple tools vs. understanding design intent and a parametric model.

On the flip side, when version 2 of the product comes down the path, it would be nice to (merge) these changes back into the engineering branch for further design improvements.

This would truly be one dataset that represents a product throughout it's life. 

Seems like engineering wants robust models that change easily and manufacturing wants models that are stable and predictable. These are 2 different things and is the reason why the current system isn't working very well.

Looking for the next thing,
  

caradon

@bill‌
Exactly.

I truly believe the greatest enabler for Onshape's success is the ability to draw in people who currently aren't using CAD at all and get them engaged in manipulating 3D design data.
For those users, the Onshape (creation) experience should be like sketching, doodling. It should be completely self-explanatory, simple, quick, no frills, usable.
Onshape should be a delight to use, even for occasional users. If it scares them off, Onshape will be dead in the water for them. In my company the proof of the pudding is exactly this: getting traditionally CAD-agnostic people to use Onshape to get better/faster/clearer results.
And then you also need the hardcore CAD functionality for power users.

Not an easy exercise for Onshape, I imagine...

This is just my opinion. I'm interested to know more people's view on this...

Dries

julian_leland

@DriesV‌ - absolutely, that crystallizes a lot of what I've been thinking about re: OS's interaction experience. I think there's a huge market opportunity if you can make yourself the default CAD platform in industry by getting nontraditional/first-time/younger users through 1) intuitive, friendly, fast basic tools (like you said, with hard-core functionality hidden behind that for folks like us), and 2) free access to the software. If every kid who has even a passing interest in making and designing things picks up OnShape independently because it's free, easy and functional, then the purchasing decisions of K-12 schools, universities, and eventually employers will follow.

caradon

JulianLeland said:

...
If every kid who has even a passing interest in making and designing things picks up OnShape independently because it's free, easy and functional, then the purchasing decisions of K-12 schools, universities, and eventually employers will follow.

I think that's one of the main reasons why SolidWorks has been so successful.
Get into education and have those students push the tool into design departments, workshops...

Dries

billy2

I remember the Soliworks EDU program. Did either of you learn SW in school?

caradon

bill said:

I remember the Soliworks EDU program. Did either of you learn SW in school?

I learned SW in college, classes around 2007-2008. It used to be pretty much the default CAD package for edu. At least in Belgium, it is slowly (but steadily) being replaced by NX and Solid Edge. I think NX is gaining ground (in edu) because of its highly modular structure. Many technical high schools here are using NX for CAM.

Dries

julian_leland

I started out in CAD in high school (in the US) using Google SketchUp, mostly because it was the most visible free CAD package available. I then shifted to SW in college, but used the school's license. I'm almost completely self-taught, though. I went to a very small college with a vanishingly small engineering department, and the cost of SW licenses combined with the number of students who would be using them regularly lead to very few licenses available --> low visibility --> low use of SW --> even less interest, and minimal educational support.

kevin_quigley

I truly believe the greatest enabler for Onshape's success is the ability to draw in people who currently aren't using CAD at all and get them engaged in manipulating 3D design data.

For those users, the Onshape (creation) experience should be like sketching, doodling. It should be completely self-explanatory, simple, quick, no frills, usable.
Onshape should be a delight to use, even for occasional users. If it scares them off, Onshape will be dead in the water for them. In my company the proof of the pudding is exactly this: getting traditionally CAD-agnostic people to use Onshape to get better/faster/clearer results.
And then you also need the hardcore CAD functionality for power users.

Not an easy exercise for Onshape, I imagine...

It is a great ideal Dries, but wasn't this exactly the gameplan Spaceclaim had? There are already 3D doodling products around like SketchUp, and to some extent Spaceclaim was alike a SketchUp on steroids (both functionality and cost wise). But unlike SketchUp Spaceclaim didn't actually deliver the package most users needed. Pure direct modelling (as from Spaceclaim) simply does not work in many cases. I know, I used it for 12 months. The Spaceclaim guys gave me an extended trial on the basis that I showed it to some of my CATIA using customers. Yes it (and direct modelling) has some neat tricks but as an original SketchUp user I was left cold a lot of the time. Had it cost £1000 rather than £3000 they might have succeeded (I know I would have bought a copy at that price).

No, here is the reality for 3D CAD doodling. Every system on the market can do this. Every single one. We have a new start last month and he has been modelling up some concepts in SolidWorks. It is interesting for me to see his approach - extrude/sweep/blend...then a lot of direct face or direct body moves. This is conceptual modelling. Of course some do it better than others but once you have even a basic understanding or a CAD system you can do "lego block" conceptualising. Therein lies the issue.

I have worked with a lot of designers, engineers, marketing, CEO level people over the years and I can count on the fingers of one hand the number who actually want to use 3D CAD for actually making edits or creating. I think it is one of the great myths generated by tech savvy media that everyone wants to be a maker/creator/designer etc. What they want is to contribute to the process but they do not want to actually sit down and use this stuff.

There will always be people who want to learn a new skill, or to tinker with a new piece of software. Everyone here comes into that category. But out there in the everyday world most people just don't care. Ask yourself this, if you like music or a particular tune do you learn an instrument to play it or download a track? Every industry has its consumers and its creators, and I think that is a good thing. We all have different skills and often the best ideas result from conversations or simply watching something. How many times have you as a designer been astonished at the insight a totally non techie person might have on your new product project? I know I have, many many times.

So I just dont see the market for a low cost or free "anyone can use it system. There is a market for a system that opens any file format and lets you take off dimensions, section, save out different file types etc.

I probably come over as negative on this but I'm not. Nobody is more enthusiastic about 3D CAD than me (my first 3D CAD was ROMULUS 30 yrs ago) but I talk to businesses every day and the feedback I get is "why would I want to learn all that when I can get it done?". 

Some people want to do this, but most don't. In the same way some like growing vegetables and some prefer to just buy in a supermarket. 3D CAD and 3D Printing are full of vendors trying to push this fiction that everyone wants to use it. They don't.

kevin_quigley

One more thing about education. Free software makes no difference to take up. What makes the difference is incentivising the student and the teacher to learn it and use it. The fact is (and vendors know it) any student can get free software. They download a cracked version...

so if Onshape wants to get students and faculty using the product they need to promote it and nothing does that better than hard cash or paid opportunities.