All internet discussions should be like this
(co-author: Jeff Waters, first published on LinkedIn, September 23, 2017)
Kim Ravn-Jensen (left) and Jeff Waters (chairman of the LinkedIn group "New Trends in CAE Simulation", right)
I have become somewhat tired of group postings which play safe and for that reason are simple, unilateral marketing messages. I therefore started the discussion putting myself and my message at stake. If somebody wanted to counter my views with good arguments, bad arguments or outright flaming, I opened for all those possibilities. Jeff Waters, chairman of the group in question, took my challenge in the spirit it was meant. Thanks to Jeff for that.
The result was an internet discussion aiming at uncovering how to possibly achieve progress for the world of mechanical engineering simulations which both parties cherish.
I think that Jeff and I did better than average. What do you think?
("OS" below means "open-source software")
Kim Ravn-Jensen: Would you agree that open-source software is on its way up? Last Wednesday, I went online with a company which aims at filling the gap between the headcount (http://www.simscale.com now has 100,000 members), enthusiasm and lack of experience associated with open-source software and the experience of the COFs to which group I myself belong. Concerning my website, you guys may have the necessary knowledge to go straight to this page: http://www.simxon.com/scurve.html
Do you agree with my analysis of the market situation? Do you agree that the big license-based vendors need to or will need to explain their pricing policy just as the case has been with established phone companies?
Jeff Waters: I agree that OS will continue to grow for certain parts of the market, and I agree that commercial vendors will need to adjust pricing models to account for easier access to expanding HPC capacity.
With our STAR-CCM+, for example, you may have unlimited cores for a set price with the POD "Power on Demand" license model.
How does technical support figure into your prediction? Customers aren't just buying software, they're buying experienced guidance and technical support. There is value in paying for support from the entity that creates the software in order to avoid the finger pointing game. For many use cases that kind of support is critical.
Let's back up and look at the overall value proposition of doing CAE at all: Let's say you spend $100k for software that solves a $1M problem annually. How important is it to cut money from that $100k software budget vs the risk around technical support and product reliability (from a known, responsible, commercial entity) with OS?
Kim Ravn-Jensen: I note that you expect open-source software to "continue to grow for certain parts of the market". I did not intend to be more insistent than that.
You should, however, be aware than many companies offer support for open-source products, so you can pick your own sweet spot for yourself.
If your sponsor is fully aware that the solution to a given problem is worth $1M, I would myself spend $100k on trouble-free software. But not all sponsors appreciate the benefits of simulation results that clearly. Today, many fine mechanical solutions emerge from organizations whose economies are much more starving.
Jeff Waters: Yes, that makes sense. My point (and I think you agree) is that commercial software pricing includes the value of technical support. Pricing models might need to shift as the market demands, but we're mostly not talking about a commodity market for CAE tools.
I definitely see value in the OS world. Had some great exposure to OpenFOAM, for example, while at ESI. Also a big fan of Simscale (and David Heiny, in particular).
But, I don't really see OS having value at the expense of commercial software.
On the tech support front, I think OS suffers a bit in comparison to commercial vendors. Yes, you can find independent OS service providers, but they are generally very small and not well funded. More importantly, they don't have immediate control over the quality or direction of the code in question. That will make large CAE customers uncomfortable, but may not be a problem for SMB.
Kim Ravn-Jensen: Jeff, you and I are definitely on the same page with only slightly different views on things. However, I want to comment on two of your last statements: 1) "we're mostly not talking about a commodity market for CAE tools" With 100,000 SimScale users, each of whom can use the platform for free, a commodity market is on its way, and this will be a game changer for us all. 2) "But, I don't really see OS having value at the expense of commercial software" The CAE business has plenty of room for different takes on the subject. But OpenFOAM should be taken as an example of how excellent an OS offering can be at its best. openfoam.com offers professional support, should be attributed a satisfactory staying power and they - as anyone else - are in complete control of the code (I believe that the best code parts drift between openfoam.com and openfoam.org in the interest of quality). Control of the code is a true benefit for every OS user, but entirely unavailable to a licensed software user.
Jeff Waters: Hmm. I mean a commodity in the sense that a web browser or gasoline is a commodity. Wildly different scales.
OpenFOAM might not be the best example due to the unkempt fragmentation/flavors.
Kim Ravn-Jensen: Jeff, 16 hours ago you talked about "wildly different scales". Sure, but it's still the same business, and the low-end market can do stuff that the high end could only dream of just a few years ago. Remember what Ford model T or the small Asian cars did to the car market. And if you consider OpenFOAM a flawed example, look at Code_Aster instead. French government backup makes a mark of consistency on that open-source product.
A pretty picture. For more pretty pictures, see here.
