DriveWorks

Design Automation goes Mainstream

When SolidWorks Corp first approached us to incorporate DriveWorksXpress into core SolidWorks for their 2008 release, we were obviously pleased. We have spent a long time preaching the good word for Design Automation, always believing it to be a mainstream market.

The problem in getting there, I believe, has been in the words.

Have a look at our partner category for SolidWorks (Knowledge Based Engineering). I have to admit that we didn’t take too much notice of the category back in 2002; it seemed like a good idea at the time although now I understand how it held us back.

Knowledge Based Engineering sounds hard. It sounds like the type of activity that you need to be really really smart to do, and have a lot of time to do. It sounds like it will run complex algorithms to work out the size and shape of an aircraft wing based on flight time, weight, aerodynamics and material analysis. It sounds as though it will run for 10 days and come up with a series of answers that require interpolation rather than arriving at a perfect solution.

That’s not what DriveWorks does, at all. DriveWorks does Design Automation. Design Automation is where you already know the outcome; you are just too busy to do it. Yes it will calculate sizes, select features and requirements so that you don’t have to, and yes it will create new files in all the right places, create rescaled drawings, documents and send data throughout your company. These are all tasks that you could do yourselves, and have been doing for a long time. Design Automation through DriveWorks just does it a whole lot quicker and more accurately. It also allows others in the company to fill out the forms, and create all of that data.

You will also see from the partner listing that there are now 8 companies in our category. A sure sign that it is a mature market. Another sign is the research that SolidWorks commissioned before talking to us about DriveWorksXpress. The research showed 2 things.

1. 65% of engineers say that 50% or more of their products are based on variations on a theme.
2. Requests for custom products are on the increase, while requests for standard products are decreasing.

The final indicator of mainstream applicability comes from other CAD vendors. Autodesk bought Engineering Intent a while back, Dassault bought KTI (ICAD), PTC are partnering with RuleStream, Unigraphics have Knowledge Fusion and of course SolidWorks now have DriveWorksXpress in their core product.

Long may the design automation revolution continue! 

Use a fish to catch a fish

A company in the mid west made some basic assumptions

The assumptions

They have 4 engineers

They have SolidWorks

They have DriveWorksXpress

They could afford to free up one engineer for 1 ½ hours per week.

For every 4 minutes that one engineer spent with DriveWorksXpress, he could save each engineer (including himself) 1 minute per week by the next week.

Any saving they made in time would be re-invested in using DriveWorksXpress

If they freed up one guy for a week, he would be trained on DriveWorksEngineer

They would put this plan into action.

The Result

After 5 weeks they were saving enough time in design to send one engineer for a week long DriveWorksEngineer training course WITHOUT AFFECTING PRODUCTION

Here’s the math

	Time Spent On DriveWorksXpress	Time Saved per Engineer	Total Time Saved
Week 1	1.5 hours	0 Minutes	0 hours
Week 2	3 hours	22.5 Minutes	1.5 hours
Week 3	6 hours	1 hour 7.5 minutes	4.5 hours
Week 4	12 hours	2 hours 37.5 minutes	10.5 hours
Week 5	24 hours	5 hours 37.5 minutes	22.5 hours
Week 6	DriveWorksEngineer Training Course	11 hours 37.5 minutes	46.5 hours

new features -v- reliability

Its an age old question, a question that has been around since man first decided to to make things better.  Do people prefer new and exciting advancements in technology, or would they rather have rapid change at the expense of robustness and reliability.

From a software developers point of view, we hate the idea that our competition may have a feature better than ours, even if we know that our overall product is better.  We also take revenue from upgrade contracts, so we know that we need to provide value for money through enhancements.

The question then is this:-  Is it possible to deliver software on time, that is both feature rich AND robust?

We worked out the other day - mainly because we were curious - that to test our software on every permutation of operating systems, Microsoft office, SolidWorks and SQL that we currently support would require 1620 testing environments.  Testing on DriveWorks takes around 5 days. 

Clearly 40.5 years would be a long time to get out a new release.

Obviously we don't test DriveWorks completely on every permutation that we support.  That would be commercial suicide.  Releasing flaky software would also be commercial suicide.

So - what can be done about it?

Testing automation provides the main key to unlock the problem.  We use TestComplete from AutomatedQA.  Using TestComplete we have automated the testing of our new version of DriveWorksXpress (The one that is being shipped inside SolidWorks2008) which now takes just 5 minutes to run a complete test on a single environment.

The initial use of TestComplete has also shown significant benefits for the testing of our main products, reducing the testing of DriveWorks Administrator down to around 40 minutes on a single environment.  More work needs to be done to include some of the richer features of DriveWorks that still require manual testing, but its a great start.