As the market sits today, unfortunately, I don't see an immediate fit for free open source software (FOSS) having a huge impact on the CAE market. It could have some, but not huge. I think the reasons are due to the shear size of the market, complexity of the software, size of the developer community and the accountability of the developed code. Big claims, so let me provide a bit more.
The size of the CAE market is on the size of 10s of thousands to 100s of thousands at the absolute most. The size of those that actually have the time/skills to actively develop is a few % of that. Plus, the culture and expectations of the user community is one of dependence on the commercial vendors to handle everything. There hasn't been an outcry by the user community to provide "the source" or allow access to the internals. In fact, the outcry is usually the exact opposite. The expectations from the user community are immensely high in that they want to adopt a new way of product development (ie simulation), but.. they expect it...
The size of the CAE market is on the size of 10s of thousands to 100s of thousands at the absolute most. The size of those that actually have the time/skills to actively develop is a few % of that. Plus, the culture and expectations of the user community is one of dependence on the commercial vendors to handle everything. There hasn't been an outcry by the user community to provide "the source" or allow access to the internals. In fact, the outcry is usually the exact opposite. The expectations from the user community are immensely high in that they want to adopt a new way of product development (ie simulation), but.. they expect it...
- To be super simple to use and implement
- Extremely accurate with minimal user input
- Reasonably priced
- Fantastic support
Surely not claiming that the shear size and complexity of CAE software is the barrier to a FOSS model. I am claiming that the size of the market and the early stages of mass adoption are the limiting factors right now. There is simply not enough demand and resources out there to really make it work across the board.
The CAE space is very competitive and many people still view it as "nice to have". Until it is mainstream and on the critical path, it will be difficult to see a vendor making the big leap into the wide open world of FOSS.
I am a big fan of the FOSS model in mass adopted applications - firefox, thunderbird, ubuntu, banshee etc... Would love to hear an argument how to take a very competitive, "niche" product like FEA or CFD and make a business model that can survive as FOSS.
Would love to hear alternative arguments to this.
5 comments:
An interesting point about the size group who could actually develop FOSS CAE tools. It is certainly tiny compared to a project like a web-browser or word processor.
I think it's possible though, considering that so many of the people who would be tapped to do it are in Academia... they are all writing code anyway at the MA and PhD level, and they are working towards something other than a paycheck at the moment.
The challenge there is the diverse interests of those people. They are SUPPOSED to be writing solvers for arcane, niche needs like blood crystalization on Mars. So, how do you reign them in to focus on basic CFD/FEA first? And, the bigger challenge, how do you get them to put usability and elegant design first?
I would suppose you'd need to search for a seed team at various universities in the CAE areas... but, also it would be interesting if you could get some human factors & GUI design folks involved from the more colorful side of the school.
Interesting thoughts:
"As the market sits today, unfortunately, I don't see an immediate fit for free open source software (FOSS) having a huge impact on the CAE market."
I guess if we narrow the focus to the CFD market then the current 1000 lb Gorilla in the CFD FOSS space is OpenFOAM. OpenFOAM is developed and supported by a number of highly respected and profitable companies (e.g., OpenCFD, Wikki, TotalSim). I would go so far as to say these companies are thriving. It's not a hobby for them developing OpenFOAM - they are getting paid often by big-name CFD users, just check out their client lists.
Up until now most OpenFOAM development has focused on niche applications with ease-of-use not a primary concern. But who's to say that is fixed in stone? Who'd have predicted that Unix would evolve (through Linux) into an easy-to-use desktop system such as Ubuntu?
Here at Symscape we are developing a GUI environment around OpenFOAM that offers the standard benefits expected of a desktop CFD application. While our software isn't FOSS (to see my reasoning try Open or Closed Source Software), it's modestly priced thanks to the FOSS components it uses.
Maybe FOSS hasn't caused a revolution in the CAE market yet, but in terms of the evolution of the CAE market, we are only just at the end of the beginning.
Richard.. valid comments, for sure. OpenFoam is clearly not a hobby effort, but still geared to "big name CFD users". This is a great way to establish the market and push it forward.
My sentiment is that it is far from mainstream and as the CAE space becomes "mainstream" and not targeted to high end dedicated analysts, the need for open source or "high end" capabilities will lag as well.
In summary, I am all for open source, I'm a daily ubuntu user. But, I still think there is an early adopter mentality when it comes to these things. Hence, I lump it into a niche, "high end" market.
I have downloaded symscape, very smooth interface. Looks like there is some potential for a tool like this to help modernize the niche market.
Post a Comment