(1). What you are saying is true. That is definitely related to the operation of a company. (2). If a company treats the employee as member of the big family, the employee probably will work on the company problem longer hours on his own. If a company also treat the employee as s partner, he knows that his year-end bonus and stock shares will depand on how he treat the company and his work. (3). On the other hand, if a company treat the employee as a computer slave, the company will have problems for a long time to go. This is especially the case when the company is having problems. (4). The same situation can happen to the commercial CFD vendors when the company change hand. But that is fairly easy to solve, as long as the programming work is a team work there will be several engineers involved at the same time. And through the daily communication, they should all become familiar with the code. So, there is a minimum critical mass for this to happen. If there is only one person familiar with the code, it is then up to the manager's intelligent decision as to whether a salary raise is more practical or the code is cheaper to keep. (5). In CFD field , one is dealing with the intellectual properties. So, there are different rules to handle it. There is no way to delete the program and logic from the employee's brain when he decides to join other company. Those are governed by the copyright law, patent laws and contract between the two parties. (That's why sometimes I said real working schemes are unlikely to end up in commercial codes. Why should he? Unless he is part of the company.)

i agree with you. from my experience using commercial codes(no all CFD some FEA) most companies are extremely willing to help even do the analysis for customers. after all if you're paying $25000 or more for liscences the companies should not only wipe your ass but take a shit for you as well.

uh oh i sense a tangent ; Danger Will Robinson tangent, tangent

are youtelling me that you're paying big dollars for you code and then youhave to pay more for help desk support. i guess that's the new scam. they charge large $ for the code, make it hard to use, and fleece you for the help desk fees. tell them to bug off and get another code that'll give you help for free

i disagree with you Tapio i think it's good the way it is. else it becomes like the journals in the engineering library. first you had only a few journals, say journal of fluid mechanics. now every little area of fluids has it's own journal so you have to subscribe to all so you don't miss anything and you have to publish in all or you may not get read by who you want to read it. i think the discussion is good for everybody. the commercial users get to read the theoretical discussion and the theoretical people get to see what can be done on the commercial side and everybody learns from and discusses with each other. if you don't want to read about large eddy simulation or flow of ketchup in gas turbines skip the thread.

No I'm actually paying medium dollars for my CFD program(s). The thing is that I am working for a university and most CFD companies offer special discounts when the program is for non-commercial use. This is fair for several reasons:

1) We do not earn money on their product.

2) Their programs are mentioned in scientific journals, where it is customary to describe what program is used in order for other researchers to test the hypothesis presented in the article.

3) We train students through their education to use and understand CFD and thereby expand the market.

4) We often come in to contact with industry and are able to advice on the use of CFD in their company (also good and bad experiences with different CFD programs)

5) Through our research, new models are developed and published. The results can be used by the CFD companies to improve their own code (if they like).

So the companies get their money back through different channels. However, as we have academic licenses one CFD producer has limited our access to support, so we have to pay pr. hour. If this is fair is doubtful when taking the 5 point above into account but I will leave that to other to comment on. The consequence is that we have 3 CFD packages in use at the moment, so when one CFD producer starts changing the conditions to our disadvantage - we simply change CFD packages. It is not cheap to keep 3 CFD packages in the air - so all in all - we pay for it !!!.

Regards

Michael

mcihael: it would seem to me that your providing a service to these CFD companies whose value is well in excess of the lower price they give you. you are promoting their code, expanding their market and help with code development and validation. in my opinion this is worth a lot. i don't know how much you're saving but take this example into account. i've heard of two impressive aerodynamics calculations using StarCD (which from my understanding is widely used in europe). both were published and reasonably well received. in these two cases the results were obtained by corporations not schools so you can assume they get economic benefit so they should pay full price for the code. however if it had been researchers from you school who'd done this there'd be no direct economic benefit (ie the school isn't making a profit) and StarCD gets good exposure for free to their target audience. if because of this StarCD sells even one full commercial liscence you would have made them more money than you saved by getting academic liscence. what i'm trying to say is that if your department is providing a cfd company with code development, training, and promotional services which help it to save money and sell more product you should at least be getting free help desk support. if you weren't doing this they would either have to provide these services themselves or pay other to do it.

Such benefits to the commercial CFD vendor would be seen as 'intangible' by the Board or Sales and Marketing Depts. Becasue of this, and the fear of setting a precedence, the academic license will continue to be viewed as the 'backwards' cousin of the commercial license (and treated accordingly).

The benefit that is sometimes provided by good quality CFD studies in academia is always going to be outweighed by the mass of poor investigations conducted by novice users. Going back somewhat in this thread ..... 'The worst enemy of a commercial CFD vendor is a bad user' and boy, are there some bad users in academia (probably because they can't afford to pay for either good training or help desk support!)

(1). When you have three commercial CFD codes, do you and your students have access to the source codes? (2). Do you get the same results from all three codes? How do you handle the situation when students obtained different results using different code? (3). Do you also advice industry to use three CFD codes at the same time? (4). When you developed a new model, do you offer the vendor a paper, a set of equations or a plug-in code? (5). When your model is incorporated into the commercial code, are you responsible for the user's results using your model? (6). How do you know that the model is incorporated in the code correctly? (7). After graduation, who is responsible for the training of your students, when a brand new version of the code is released (that is his knowledge about the old version becomes obsolete.)? (8). I am getting the feeling that the school is becoming a training center of the commercial codes.

in response i have two points. (1)for the sales people to say this is lazy (although i agree that they'd say that since it cuts cost.) because undoubtedly they have a few MBAs in there who can do cost analyses and come up with good estimates/predictions of the benefit to their company in hard dollars. indeed the academaic users of these codes ought to do this analysis and present to the companies so they can get a fairer deal. after all engineers should be able to figure out how to do cost analyses as well. (2)you made the valid point that bad users are bad for the commercial cfd company as they make their code look bad. as such it'd seem to me that the cfd companies should do as much as possible to make their users (particularly the academic users who'll (hopefully) be their commercial users in the future better users. this is what the better FEA companies have done even to the point of giving free in house training to university users because they know that the well trained university user will be more satisfied with their code and well tend to use it when they get into the working world. the FEA companies should be the model for the whole commercial computational engineering community because they have been able to expand their market and be successful over the years. not more than fifteen years ago the FEA market was small only a few companies offered expensive hard to use codes which were only used by people (with advanced degrees) for critical applications which required high levels of precision (aerospace etc). now ther are more commercial fea codes which are orders of magnitude cheaper easier to use while not sacrificing accuracy and applicability (ie they're not cut rate) and even perform better. this is where the commercial cfd vendors must go if they want to get big and make piles of money (that's the purpose of business after all)

(1). Sometime in the future, somewhere on the package, there is a line of fine print. (2). It says, "WARNING: users under age 18 has been determined bad users, thus the use of this commercial CFD code can be hazardous to your health." (3).It is an irony, that on the front of the package, there is also a line of print, it says " Our policy is: the customer is always right."

(1). After a week or so, thinking this commercial code support issue, I will state my personal feeling briefly here. (2). I have a strong belief that CFD is neither a code nor a computer. I think, CFD is a human activity which involves in the numerical solution of fluid dynamics problems in general, at some stage in the process, it requires the help of the computer. (3). In the past, the computer hardware companies used to involve in CFD heavily. And In many cases, the super-computer, the mini-supercomputer were all designed to solve CFD problems. And hundreds, and thousands of CFD related computer codes also have been developed and written by researchers and engineers in CFD field. (4). Since the issue of the support of the current commercial CFD codes are related to some specific commercial CFD codes, I think, it is a good idea to have a separate forum on Commercial CFD codes. (5). In this way, the questions unique to the specific commercial CFD code can be addressed. I think, this is a better approach because those questions asked specific to the operations of a commercial CFD code, in gereral is not CFD questions. CFD is not about selecting an option, setting a boundary condition, or reading a format file in a commercial CFD code. Those questions are purely related to the operation of a code. (6). I think, this is an important issue. Answering commercial CFD code questions, promoting commercial CFD codes should be treated in a special category. A forum on the applications of commercial CFD codes would be more appropriate in this case. (7). In this way, it also will take the pressure off from the readers who are interested in answering the posted questions, without worrying about being attack by a pro-CFD code user. (8). It is not healthy to continue on the current format.

I will try to answer your questions:

1) For two of the codes there is access to the code through Fortran routines. It does not mean direct access to core code, but it is pretty much possible to re-program both CFD programs from the lowest levels - if you know what you are doing (and read the manual!). I use it for example to alter the rheology in multiphase flow, as the standard CFD programs do not have the facilities that I need.

2) There can be differences between codes due to different implementation and solution techniques. When I get at new software I always test it on standard hydraulic problems, where I know the answers. Students use the same software though their projects, so we do not - on a student level - spend much time on comparing different packages. (I don't think it is important).

3) I do not advice industry to use different packages. They usually have the money to buy the program, the training and support they need. I only try to analyse what need - the few companies I have been involved with - have and point in the direction where I think they will get the results they want. It is very informal and most times they have never thought of using CFD - so we get an interesting discussion. For my part I get to see different industry related problems, which I can use in the education.

4) No what I meant was that the result coming from the academic area is public domain. Every vendor can use it or not. If you look at the models implemented in the CFD programs, many of them heavily sited in the literature ( and thats why all CFD programs have many of the same models implemented). The CFD programs are rather conservative, when selecting what model should be implemented. There need to be a general acceptance of for example a turbulence model before it is offered as an option in a CFD program. In that sense you pay for the implementation of the model - not the model itself. I don't think there would be any commercial CFD software if it was not for the work done for example at the universities. BTW I think that in some academic licenses, the user are obliged to send a copy of an article if it is based on CFD program.

5) I do not take credit for any model and therefor no responsibility (see 4).

6) As a good engineer I test the model under different conditions, where I know the answer. If I cannot develop an analytical solution to my model, then I compare it to measurements.

7) I can only speak for myself. I do not train students in a specific program - I find it uninteresting. It is more interesting to teach the students to be critical to the code they have in their hands and develop skills for analysing and interpreting results. If they use a specific program the it is just a matter of reading the manual and learning how to get things set-up correctly. In my experience - if you know one program very well you can fairly easily change to another.

generally speaking I find that commercial CFD programs are launching pads for development. I don't want to spend 3 years getting a unstructured grid generator to work, when I know that this has been done many times by others. But if I can use a "open" CFD program to study the physics of what I'm working with, without spending lots of time programming, then it is great !!

Regards

Michael