Elting Mechanical Enterprises

SolidWorks 2010 showed some of the future plans that Dassault Systemes has for cloud computing. The thought being that CAD software would no longer be platform dependent and therefore the designer would no longer have to deal with performance problems related to drivers, OS issues, etc. Okay, I'm all for more stability and better performance. What concerns me is something far more sinister may be in the works.

Dassault Systemes is talking about how with their PLM software, CATIA users would not even have files stored on their workstation but rather, the CAD data would be stored on Dassault's remote servers. Instead of "open", there would be "search", no saving only "publish". This concept of having Dassault being in charge of maintaining my CAD data is frankly, more than a bit unnerving. But a more serious issue is related to the defense work that I do for some of my clients. I know this will "not fly" with the DoD.

But aside from security issues, the main issue I have is since we will no longer have CAD software to install on our workstations, what keeps Dassault from blocking a user from logging onto their servers? This would also force users to maintain subscriptions or jeopardize their access and use of their CAD data. I see a storm coming in the form of cloud computing.

At the recent SolidWorks world conference, the presenters showed SolidWorks running inside the Mac OS. The main point of this charade was really about computing "in the cloud". In really, the Mac was running a dual OS system.

I recently sent an email to Steve Jobs asking them to seriously look at opening discussions with the SolidWorks folks about porting to the Mac OS. The implications would be a watershed moment for the CAD community. Not that there are no other CAD systems that can run on Mac OS, its just that those few are Mickey Mouse.

Having a powerhouse like SolidWorks, running on a Mac OS would force the major players (Autodesk, PTC, Siemens) to re-think the Windows only platform for their CAD products. This could also cause Windows to start thinking about the crap they dish out and maybe, just maybe, they'll wake up and start getting back to solid, tested, and stable OS.

A number of mechanical software developers have been working on new/old method of creating solid models that us not history based. Meaning that new features (holes, cuts, etc) are not part of a parent/child relationship. Now this is nothing new. For many years, we could create solid models in Autocad which had no dependant feature history. The real difference now is that we can create them using parametric sketches. This is an enormous improvement over past methods.

Why Explicit Modeling is Important

For years the only way to work with CAD files from different developers was to export the native file (i.e. Pro/E, SolidWorks, etc) to a neutral format like IGES or STEP. Once one has imported the file into your software, editing the solid can be a chore since most modeling software don't have the tools to deal with these "dumb" solids. This is where the explicit modelers really shine. The next few years will show these types of modellers to start gaining market share. Some of the larger software developers have seen the writing on the wall, namely PTC (Pro/e - CoCreate), Autodesk Inventor Fusion, and Seimens Synchronous Technology. So far, SolidWorks and Alibre Design have not expresses interest in developing explicit tools. I truely believe this is a mistake. For those of use that need to modify customer files for stress analysis, having tools to easily de-feature a model is a must.

The "Holy Grail" of CAD

Many have thought that CAD software developers would come together to accept a unifying file format that could be read by different CAD software. For years it was IGES and more recently STEP (STandard for the Exchange of Product model data) formats. Where different CAD software can import the geometry, their really isn't much in way of "intelligence" left in the part. Meaning that all the parametric data that developed the features of the solid are lost in translation. The "Holy Grail" was thought of if a version of say STEP could have the parametric feature data inside the code of the file, to be read by whatever CAD software and the geometry would then be read in as if it were a native file. The closest any of the CAD developers have come this has been with Alibre (www.alibre.com) submitting its STEP schema to the International Organization for Standardization (ISO) for review. That's great, for everyone else. Until all 3D CAD software exports using this schema, all the receiver will get is a dumb solid. Its been eight (8) years since Alibre submitted this to ISO so I don't expect much came from it.

In conclusion, as I see it the best alternative to the "Holy Grail" is having really good tools for "carving up" dumb solids. That means, Explicit modeling.

Finding the right solution can be difficult but, even worse if one is looking for it in the wrong place.

Often clients will tell me that they have a production problem and it is all because of our "widget" machine and they want me to have a look at and see what can be done to fix it. Once I've gotten to see the widget machine in action I ask, "so, tell me what's the problem? It appears to working fine. I see that its pumping out gadgets at a pretty good rate so, what gives?" The usual reply is something like, "its working fine now, but it breaks down all the time" or, "its doesn't produce the gadget to spec but, today its doing okay" or something to that effect. Sound familiar?

At this point, I start my "investigation". I round up the usual suspects (line operators, maintenance techs, engineers) and start in with the questions. It never fails to amaze me how no matter what kind of gadget a company produces, they all suffer from the same problem when it comes to solving production issues, perception is their reality. Facts don't lie but, people do. Or at least they lie to themselves and they believe their own lies and then the lies become the "truth".

So, how do we fix their production problem. First, we start with the end in mind. Meaning, what is the target throughput of there operation. How many gadgets need to be shipped. Working "backwards" through the operation, allows me to get down to what is truly important, finding the real bottlenecks in the processes. One of the best ways to find bottlenecks is looking at the data that has already been collected. Once its been correctly analyzed, patterns can be found that usually lead to causes in lost production. The trick to this is the "correctly analyzed" part of this investigation. People have a "funny" tendency to see what they want to see or they just don't think to look at the data in a way that can help them see the real issues.

Once a client sees that the widget machine is not the real problem and its actually the gizmo machine that is truly slowing the process down, its like an "ah ha" moment. Sometimes, it just takes the a bit of courage to press the client and find the real truth.

Lately my wife and I have been watching a lot of Gordon Ramsay's Kitchen Nightmares. Even if you are not a foody or into reality TV, it's still a good watch. Well, one episode Ramsay goes into this one restaurant that is dying a slow death. The menu is overly complicated, the kitchen is filthy, the order process is completely screwed up, and to top it all off, there's to much middle management and the top brass hasn't the balls (or bollox like Gordon says) to keep them in line. So, does any of this sound familiar? Ramsay has his head in his hands and is saying, "where do I begin?".

At this point my wife turns to me and says, "My god Steve, isn't that what you do?". In all honesty, it's exactly what I do. So, now my morbid curiosity has me glued to this TV show picking up "pointers" on how he handles leaders who aren't leading, managers who run around like chickens sans their heads, and workers who are complacent, unmotivated, or completely incompetent.

Tip #1 - Cleanliness

Gordan rails about this topic on almost every episode. Any one who works for a company that has embraced Lean Six Sigma practices, knows what 5S is all about. If things aren't clean and tidy, how on earth our you going to get things done efficiently. The five S's are:
Sort: Going through all the tools, etc., in a work area and keeping only essential items.

Set In Order (or Simplify): Keep the needed things in the correct place. (I can still hear my Dad saying, PUT IT BACK WHERE YOU FOUND IT! Ah, I miss him....)

Shine: Clean it until it shines....

Standardize: This is the hard one for most companies. Procedures need to be put in place so the first three get done.

Sustain: Make it a habit to get it done.