Twice As Smart
Things don't look too good for manufacturers these days.
One theme I keep hearing is that the most critical thing manufacturers can do in this economy is to offer products that their customers want. But, how can they do it?
Scott Adams, creator of the Dilbert comic strip, riffed on this theme recently: The "pointy haired boss" told his his engineers "we need to be twice as smart to survive in this economy." Dilbert, the engineer/nerd, replied, saying "Good Plan. I look forward to spontaneously developing an I.Q. Of 400."
I'm not so sure how realistic it is for manufacturers to count on their engineers spontaneously doubling their intelligence. It might actually be more realistic for them to count on losing some of their most intelligent engineers to the economic crisis. After all, it seems to be the case when reducing headcount (for whatever reason) that companies lose some of their most talented people.
Here's something that some corporate executives don't seem to get: design and engineering is where value is created in their products. You can do all the marketing and sales you want, but if you don't have products that people really want to buy, you're fighting a losing battle.
CAD is the essential tool that engineers and designers use to create value. While some engineers and designers spend a lot of time with their email and word processing programs, it's the time they spend with their CAD programs when they transform ideas into real value.
Sadly, the CAD industry has done a terrible job of giving engineers and designers the tools they really need. The vast majority of mainstream CAD programs are not even designed to be used by engineers and designers. They're designed to be used by technical specialists – CAD gurus.
So far as I know, there is no formal research into the overall productivity of CAD users – at least, none that anyone is willing to admit to. My estimate is that only a minority of engineers and designers can use a modern (sophisticated) CAD program to start with, and, of those, less than 10% to 20% are really proficient. I've asked a number of people who I consider to be CAD experts, and their estimates are generally even more pessimistic than mine.
Imagine that you ran a company, and invested millions of dollars in software and training, and, in the end, got such dismal results? I doubt you'd be very happy.
Yet, it doesn't have to be this way. At least, not now. There are new generation CAD programs that can be used – and used effectively – by "average" engineers and designers who are not CAD gurus.
The common denominator of new generation CAD programs is that they are optimized to enable direct editing of CAD data that comes from any source, unlike older generation CAD programs, which are optimized to edit native (proprietary) data. As almost an unintended side effect, the new generation programs are significantly easier for average engineers and designers to use than old generation programs. (The technical reason for this – at least the shorthand version of it – is that the new generation programs use feature inference in editing operations, and ignore parametric history. As a result they don't have to contend with hidden history-based dependencies that can cause modeling failures, and confound users.)
While CAD experts can rightly argue that the new generation programs don't do everything that the old generation programs do, that's not really the point is it? Imagine that a company has 100 engineers and designers, of which only a small percentage (choose any number you want) are truly proficient using existing CAD tools. How might it affect the company's overall product development productivity to give the rest of the engineers and designers new generation CAD tools?
I've spoken to a number of people who use new-generation CAD programs, and my offhand estimate, based on these conversations, is that an average engineer or designer (one who is not a CAD guru) will be from two to ten times as productive with a new generation CAD program than with even a best-in-class old generation CAD program.
I'd argue that the best way for a company to make their engineering and design staff "twice as smart" is to give average engineers and designers tools that they can actually use productively, on a day to day basis.
Evan Yares
12 Comments
Reader Comments (12)
While improvements in the latest direct editing tools are useful, they will not eliminate the benefits of parametric design intent.
The real issue is most companies dont invest in adequate training. I also found many users are too lazy or apathetic to master the tools they use.
Having been involved in several CAD users groups, I am shocked at the lack of personal motivation to become better users. Attending a meeting to network and snack is not going to make you a better CAD user. I've also found most companies are reluctant to let their users attend technical conferences and users groups on company time.
Parametric CAD is not rocket science!!
I believe this is more in the culture of the society we live in. That culture tends to be a control one. Leading to peoples who ask what to do and follow.
On the other hand we value simplicity that had the effect of not valuing intellectual effort.
Best example of that culture is the economic crisis. Where peoples bought house that they cannot paid and on the other end those who offer those mortgage had found a simple way to make money.
So we have people that just follow with people who find simple way of making money, both did not make any intellectual effort to see if these make sense.
I see reflect of that culture in our industry...CAD/CAM. And I believe that CAM has been touch more than the CAD by this...
For years now the CAM industry relied on operator who’s job is to operate automate machine the same trend is spreading in the CAD. We are slowly shifting toward CAD operator instead of CAD designers/engineers.
In the CAM few years ago programming a CNC was demanding, but slowly CAD programs has made this easier to a point now that we have feature machining where the CNC programs build toolpath automatically. It is so simple that is must be good. I left the CAM market few years ago, but I would wonder how many CNC programmers use manual command? How many retouch program after being process by the CNC program? I believe that they are more CAM programmer operator then we are willing to admit.
Same in the CAD few years ago we switch from paper to 2d to 3d in less than 30 years quite fast when we compare to CNC.
Same trend in the CAD, we demand for automation so we push a button and poufffff design is complete. Next thing we will see is peoples not verify their design; we take what the machine throw. It is simple so it must be good...
We will have the impression that we lead because we operate a CAD program, but in fact we will follow. Because the impression that we lead we will think we make intellectual effort. But if we take result has it is what effort do we do?
That culture will touch the last member of our trilogy soon FEA, I predict five year from now we should see the first symptom.
In conclusion
I am agreeing with John, 30% can be attribute to corporation not investing enough to motivate their employees. This is what it takes to make sure that 50% (the employee) get motivate the last 2o% fall in the side of software developers.
John refers to latest development in direct editing; I assume he talk about synchronous technology. It took roughly two to three years to develop. If corporations and users wait for the perfect tool we might wait for long.
What we need to do is to maximize/optimize what we have at our disposal now. To achieve this we have to make sure that the most efficient tool is put to work, the brain.
I often make the comparison between Olympic athlete and a CAD department. To maintain optimum performance we have to train them on a regular base. In a more precise way I will think about decathlete that has to master ten disciples.
Automated tools like "Live Rules" in NX and SE Synchronous Technology will require practice and experience to use effectively.
Perhaps we should go back to boolean solids. Doesn't get much simpler that "hack and whack" modeling.
The area you seem to be concerned with is based on parent child relationships in parametric modelers. Even as a casual user, I just don't find them difficult to deal with.
I've been directly involved in supporting parametric CAD for over 10 years. The biggest stumbling block in the users and managers.
A few minutes ago I got off the phone with a relatively new user. He is working on a 30 sheet drawing (184Mb!) and running into out of memory problems. I have repeatedly told him, his peers and his managers that he either needs more RAM, or change the way he constructs drawings. In our conversation he said "we haven't added the RAM you suggested yet".
When users and managers don't listen and learn, it doesn't matter what CAD software they use.
Additionally, in some organizations, for an ‘engineer’ to be required to use any CADD system would be a huge waste of resource and yet in another it is more than appropriate; the reasons for this difference is the fact that one may work in a manufacturing plant, overseeing production issues and the other in a manufacturing facility making complex automotive components. There are too many combinations, and that is part of the problem; ‘engineers’ do what?
“Sadly, the CAD industry has done a terrible job of giving engineers and designers the tools they really need.”
I’m not sure that I totally agree with this comment Evan. An engineer is very much in control – as s/he should be – of what tools s/he uses; if s/he has the wrong tools I would argue it is a problem s/he should rectify.
“ The vast majority of mainstream CAD programs are not even designed to be used by engineers and designers. They're designed to be used by technical specialists – CAD gurus.”
Yes but, wasn’t CAD originally developed as a tool to be used to document designs and products and used by draughtspersons? Evan the first system I ‘used’ in the early 70’s was to be driven by drafties. Also, I have known some so called ‘CAD gurus’ and quite frankly knowing how to drive a CAD system goes only a very small way to making a person a good detail or design draughtsperson (and no way to being a good engineer). I don’t know a single person that could not use a pencil, rule and compass but few can draft an engineering drawing and even fewer could use these same tools to design and draft the required documents at the same time.
There was a period of time where there was such a beast as a ‘CAD operator’, and very highly paid was s/he compared to a draughtsperson on a board. PC-CAD fixed that - for a period - but verticalization of CAD products, into various industry streams, is returning this situation and I do not believe that is better.
So, is it possible to say CADD is a tool for an engineer or not and then try to find a solution in what an engineers’ CAD system should be like and or do? I personally think it is a bit like a dog chasing its tail!
“So far as I know, there is no formal research into the overall productivity of CAD users – at least, none that anyone is willing to admit to.”
By formal research I take it you also mean independently conducted research that can be validated? This is of particular interest to me and yes I have done my own’ research’ into the actual and possible productivity improvements CADD systems achieve and or claim. I have not published any of what I have done as it has been done for my own benefit and or customers but it is fair to say that what has thought to have been achieved has not; what people had hoped for was never going to happen and that absolutely none of the ‘white papers’ published by vendors can be relied on or validated.
It is also correct to say that sooner or later ‘bean counters’, in some companies, are going to take a very close look at the cost of CAD and I believe it will be difficult to sustain an argument that will allow the waste that has occurred in spending on CAD, 3D in particular.
What I have done has shown considerable gains in productivity early on (implementation and application of 2D CADD) with 3D changing very little if not reversing, in some situations, earlier gains.
Lots of reasons and lots of arguments here, as to why, but in talking with a few managers who have also taken the time to track this type of information, similar results have been experienced or recorded. Many of the, 3D, productivity claims from vendors and users, when examined, are based on ‘what they believed has happened’.
The comments about direct editing and parametric etc. also show a form of polarization and again one is not better than another in many aspects and what can be considered great by many just simply does not jell or suit many others. As someone who spends a great deal of time introducing and training users in CAD, professionals and in several institutions, I know only too well how one person sees something totally differently to another. This important fact has never been recognized by CADD vendors and this is the area in need of research!
The first vendor to get on top of what I see and has engineering tools that fit is the vendor who will move to the front and stay there for a long time. Just look at Autodesk’s ‘success’ from a completely different angle, through many users eyes, and you may see what I mean.
Another comment I will make is simply; if an engineering product fits it will be used and successful! 2D and 3D CADD tools have been around for essentially the same period of time: one has dominated and the other has been’ flogged’ like a race horse and still struggles to gain acceptance even tho’ it ‘appears’ and is said to be the ‘right way to go’. I would suggest those pushing the ownership of the race horse need to understand most punters watch horse racing but own draught horses that earn their keep!
“I'd argue that the best way for a company to make their engineering and design staff "twice as smart" is to give average engineers and designers tools that they can actually use productively, on a day to day basis.”
I agree, but would the best choice of tools, for many engineers, include CADD tools?
After being out of the design game for a while, I took a position in research and development, and "re-learned" a 3D CAD package I had not touched in a few years. It took a few months to get up to speed, and I faced constant issues with fillets and other limitations of the software (vendor agreed they were bugs). Frankly, this kind of basic stuff should have been well developed by now. The sad thing is I could have gone to the tool room and cut the darn mold by hand in less time that it took to model it.
As long as I have been a designer, I had to translate my mental image into 2D board drafting, 2D CAD, and 3D solids. The vendors think these are different things, but I'd contend they are the exact same thing - a method to _document_ my design. I work out drafts, fillets and relationships in my mind, not my computer. CAD simply gives me a way to document and share this with others. It really doesn't do much to answer questions because 3D solids are an artificial model, in which everything is too "perfect" - perfectly flat, perfectly round. Ask your CAD system if a 0.99999" peg will fit a 1.00000" hole. Now array the pegs and holes. CAD says everything fits fine. But in the real world it is a lie.
Designers are busy and their time and attention is being pulled in multiple directions at once. They get far enough along to document the design, then move on to other things. Hard to blame them, really. I was always good with CAD because I LIKED it. I suspect everyone reading Evan's blog is a CAD fan. But for non-fans, it is really hard to justify upgrades and training beyond a certain point, once you think about employee turnover and look for real productivity figures in an ROI analysis.
CAD adds to a designer's workload and promises a payback that may never arrive. It is a monster that wants constant feeding. Engineering students study CAD at the expense of studying manufacturing processes.
So what we really need (or what I need/want) is a system that is, in effect, a manufacturing process simulator - one that builds upon manufacturing knowledge. I say drill a hole there. Now tell me what I really get (min/max, position, cost, etc). Warn me that it is unlikely that the process that I specified will allow those pegs to fit the holes. The more I know of manufacturing, the more I know about the CAD package. The more I use the CAD package, the more I learn about manufacturing.
We are unlikely to get this from any CAD company. They are too stuck in their ways to see this. When it does happen, it will likely become the biggest revolution in engineering. It is really not that hard and can be built on the existing foundation we already have. Much like the idea of using the Web browser as a platform, it becomes the tail that wags the dog.
Sorry for the long post. Evan has hear this from me before.
However, this is for land development - so I'm not sure it would be of great interest to you, other than we DO represent examples of shifts in the CAD software industry in general to easier, iterative design, cloud computing, etc. Others will follow.