I don't think it is fair to have this discussion using that topic.

"However, you'll have to provide your own mesher and post-processing tools. Support is not free. There also isn't any documentation. Or turbulence models. Don't try and move the mesh. Your statutory rights are not effected."

Support is not free? and in commercial codes it is, right? That's why you buy support licenses.

I think there is a fallacy attached to "free" open source codes. If you want complex geometry handling you need to pay for a meshing tool, if you want to be able to effectively post-process and visualize large data sets you have to pay for a post-processor. If you want support you have to pay for it, if you want training you have to pay for it, if you commercially developed models you have to pay for them.

No one is argue that commercial codes are free (that is why they are commercial) but there is a lot of benefit to be gained using them (to long to list here) and to say that using an open source code (and lets face it we are talking openFoam here) is free in anything beyond an academic environment is completely wrong. If it was, the companies like openCFD and ICON would be registered charities (i.e. not for profit)!

There is a clear bias towards open-source CFD in general and open-foam specifically on these forums and if people from commercial codes aren't allowed to help line their pockets on these boards then neither should employees of openFoam and associated companies.....

I agree that commercial software provides great pre-post- processors and solvers. The tools are in general more robust and they are very accurate.

But if I am not a CFD company and I want to study CFD as a hobbie I don't want to pay for a CFD software. I want to be able to get something free from the net. I want to look inside and see "what makes it tick".

The open source model is here to stay. Google has realized that, why can't you?

You have not done a good search. Salome, a pre and post processor (www.salome-platform.org) is open source and free. Try it, you will change your opinion, period

Well I would love to give it a go but I don't have Debian or Mandriva....Either way I have an extremely large (1+ GB) assembly with a few hundred parts (currently in native CAD format so exporting to IGES will be a pain), most of the parts are overlapping/have gaps in (in common with pretty much any industrial CAD geometry) will this software be able to assemble all of this then generate a 20+ million cell model (with a decent quality prism mesh near the walls and volumetric refinements)? Before running it then post processing accordingly...Oh and I don't want to use nasty tets either.

My current process takes about 1 day, if you can suggest a truly open source alternative I am willing to listen.

You do not need to install anything right now, just download the caelinux file, burn it to a disk and use it as a live installer. (www.caelinux.com) The salome mecca version on that file, is much better , it has code-aster (a finite element solver) and future versions will include code-saturne ( a cfd solver). All these programmes are being developed and used by the French Authority for Electric Power (EDF) and given as open source, free software. There is a wide difference between open source codes developed on the other side of the Atlantic ocean than on this side, but this is another issue. FireFox, Linux OS are just two examples of open source soft ware. On the source forge web site, there are almost 100 000 open source programmes, not all of them can be classified as trash. The best symbolic algebra programme is maxima and it is open source. By the way, I am not a design engineer, I am an analyst, I get the geometry files done, but if you are familiar with ProE or work bench, this salome is a real competitor for both. Good Luck

Paraview is not bad neither. In my experience, paid support is often OVERRATED! I am an active CFD commercial user but i still consult this website because more often than not i can get useful advice from this website, better than the crap i get from the paid commercial code tech support persons. That aside, free codes simply cannot beat commercial codes due to lack of robustness in the applications. I think it is a moot point to compare commercial codes robustness, geometry handling capabilities and so on with free codes.

The free codes are mostly for academicians to use. In their research geometries dont have to be complex but the complex science behind is what of more interest. In industry, not much time to dwell on the sciences, just use it as it is deployed. Yet, ironically, we still find plethora of turbulence models and numerical options to choose from in commercial codes. Commercial codes should have less "academic-oriented" optional buttons because 80% of the time, industry people dont use them! We are paying and indirectly funding academic curiosities of the commercial developers!

"Yet, ironically, we still find plethora of turbulence models and numerical options to choose from in commercial codes."

Maybe you should learn something about fluid mechanics, turbulence modelling (especially their inadequacies) and numerical analysis and not just use the commercial code as a "black box". There are reasons for all these options and there not all academic!

Note that this is not a defence of open source CFD in anyway (I don't use them).

"free codes simply cannot beat commercial codes due to lack of robustness in the applications. I think it is a moot point to compare commercial codes robustness, geometry handling capabilities and so on with free codes."

That reminds me of the negative campaigns launched from time to time by the developers of Windows against Linux and other open source projects.

People who write such negative opinions, Please tell us what code have you used to do what...

Be Objective or stop spreading your negative waves

I think that commercial companies and open source community can work together. The idea that open source codes are a threat is just because some of them are just too damn good. They have a large community around them.

In my modest opinion, that is what is going to prevail in all of this. It doesn't matter how good you are, if you don't have a large community around using, developing and testing your software you will not survive (commercial or open source). Today, software is driven by the community. Commercial codes can also take advantage of this. For example, providing an API and/or programming environment where people can develop plug-ins and interfaces.