How To Be As Lazy As Possible

It’s Christmas time again, and that means I’m kind of searching for new topics. I keep an eye on the SolidWorks forums, and there is a wealth of great topics there to choose from. The unfortunate thing about the forum is that it’s hard to get one message through clearly. A lot of people have something to say, and if you’re there looking for opinions, that’s great, but if you’re there looking for real answers, it’s great less often. I’ve always participated in the SW forums (except when I worked for Siemens), but my involvement now is a little less. If someone gives what I consider bad advice, I let it slide now in a way I didn’t do before. Except that thing about zero thickness. It just annoys me when lazy people are also stupid. Lazy smart people is a different thing, I mean, that’s what it’s all about, right? And smart people who work hard is almost an oxymoron, because why else would you want to be smart if it didn’t allow you to be lazy? I just don’t get it.

Which brings me to today’s topic, which is How To Be As Lazy As Possible, especially as it relates to CAD, design, and other things that sound a lot like hard work. People on the forum never ask this question directly, but it is what they really want to know regardless of what question they actually ask. How do I look like a genius without breaking a sweat?

Hardware

If you are going to be lazy, someone has to do the work. So get a good computer. You can’t be lazy with a slacker computer. Standards change over time, so unfortunately you’ll have to replace your computer from time to time (says the guy on a dusty 8 year old desktop), which I know is a lot of work, so if you can get someone to do that for you, better yet.

Software

WordPress is free, and just gave me a free upgrade to WP 5.0. It helps formatting text, and gives kind of a WYSIWYG interface. I’ve used WordPress for at least a decade for blogging. It’s pretty good, works cross-platform, cloud tool, which is great because I only use it for stuff I want to give away.

SolidWorks is a tool I bought years ago, and I also have a partner arrangement with the company for writing books and things. SolidWorks is a pretty good tool, although there are more modern things out there with more advanced ways of doing things. But here’s the catch. I’ve spent ~22 years developing expertise on using this one tool, and I can do things pretty efficiently, and solve most any problem I can get myself into. Which is to say, that I can look pretty lazy while using this tool. If I changed to something else, I might have to learn something, and of course that slows you down. But if I stay with SW, I can be lazy.

So if you’re young and you have to learn everything from scratch anyway, you can benefit a lot from those with more experience, which a truly smart lazy person would understand, but a lot of kids really aren’t that smart, and have to learn everything for themselves.

Efficient Modeling

I really started writing about what things slow down rebuild on history-based models. I should point out that rebuild is a concept that doesn’t effect direct edit type modeling, because in direct edit, you just make a selection, and make some sort of change to it. But that’s a little bit off topic for right now. I’m still on that direct-vs-history warpath, just so you know.

There was a conversation on the forum about what sort of things in an assembly are “costly”. To me that means cost you rebuild time. I don’t think we have a real good definitive list of how long it takes to rebuild various things in  SW, it would be at least partially dependent on your local set up – your hardware, network, other software, settings, etc. But I think you could get a good idea of what it takes to put together the model with the least rebuild time possible by measuring what kind of information is needed for that particular feature. For example:

• point in 3d space: X, Y, Z
  
• plane parallel to standard plane: x=5
• sketch line: plane plus line mx+b plus two end points (limits)
  
• solid cube: four planes, trimmed surfaces, and an inside/outside vector
• NURBS curve: Taylor series type equation
• NURBS surface: series of Taylor series curves plus interpolation algorithm

…and the list goes on. Not just for geometrical data, but also for mates/relations such as colinear, tangent, parallel, all these things require calculations. (This is why it is hogwash when people say that a fully defined sketch is faster than an un-defined sketch – the geometry itself has to be calculated, then the relations take additional time. The lines of an un-defined  sketch must only be calculated by themselves. This is something you can experiment with yourself. Make a big sketch with a lot of sketch relations and use performance evaluation to see how long it takes to rebuild it, then delete all the relations and evaluate the result.)

Beyond that, you’ve got things like relations between sketches, between features, between edges that would be created between two features (so these are dependent on both features), and even dependencies between parts (Insert Part) or in-context (where a specific part in an assembly is referenced, and must be open to rebuild).

Even further, you’ve got assembly features that are dependent on a part (and some of its features) plus, the position of a part in that assembly (so you have to solve some subset of mates to get that), and possibly further dependent on other assembly features like planes that may be dependent on other assembly planes or other parts in the assembly (again have to rebuild mates to figure out positions).

So you can see this ranges from the computationally very simple to combinations of computation and referenced data, and sometimes several iterations of calculations to get everything updated.

Just be sure that you’re not both lazy and stupid, because if you’re stupid, you’re not taking some of these very costly techniques into account. Anything that is a reference is going to be fairly costly. Straight math is pretty easy for a computer. I’d rather have a big complex surface than a complex multi-layered system of external references for a straight line.

Being both smart and lazy will mean that you develop a good intuition for what kinds of things are going to cost you rebuild time (or however you measure cost – possibly hair loss, if you’re like me).