Over the last few weeks there has been a lot of questions on this forum related specific commercial codes. These questions are important and interesting, but they are often not of general interest to the CFD community. I've also noticed that several readers find them a bit irritating. For this reason I will soon open a few new forums, where each forum will be dedicated to one specific code vendor. This will hopefully reduce the number of code-specific technical questions on the main forum. It will also provide a more undisturbed channel for these questions. Today they tend to dissapear in the high volume of the main forum.

To begin with I will only open a few new forums and we'll see how it works out. If you have any suggestions about which forums to open please drop me an email. Let me also emphasize that these forums will be independant, although of course all vendors are welcome to answer questions on them.

I'd appreciate any comment or suggestions on this. We've discussed this a few times before and I think it's time to do something.

(1). I think, the commercial code specific questions should be answered by the vendors professional support engineers. It is the user-vendor relationship about a specific piece of software. (2). Unless the code is a public domain code, it is really hard for a non-user to verify his answer to the code specific questions. And with the newer versions of the code, the old experience and answers are likely to become obsolete quickly. (3). Ideally, all the questions should be answered before the user starts using the code. But in the difficult field such as CFD, there will be questions all the time. (4). It is a good idea to have a separate forum dedicated to commercial codes. As long as these questions are answered by the support engineers, the non-user can always learn something from it. Otherwise, it is hard to make it useful to other readers. Thus reduces the effectiveness of the forum. (5). I do have suggestions for the vendor, that is, since the vendor always has more than one experienced user, and a vendor also sometimes has a users group meeting or conference, the vendor can select some experienced users to answer the code related questions. (6). This is similar to the technical journal reviewers selection. (7). I don't know whether it is going to work or not. It is likely that everyone is trying to keep the answer to himself in this highly competitive world. And there is nothing wrong to it either.

Jonas,

Good idea about commercial codes.

Have you ever thought about organizing a www-based seminar where people could give presentations followed by a discussion? Could be very interesting, and, certainly, there is already a community here which can form a core for such a thing. (Not that I belong to it, unfortunately, I am often too busy )

The purpose of every seminar session should be to help the speaker in his research as well as disseminate the knowledge. Therefore, an extensive discussion is implied.

The difference from this message board should be in facilities for graphics and formulae for the speaker, on-line drawing facilities for those giving comments, and for simple mathematical symbols too. Also, it would be good to be able to point out something in the picture presented by the speaker and ask 'what is this?' instead of describing it in words. And, of course, certain selection procedure for the speakers might be necessary, depending on the number of those who wish to give a presentation.

I want myself to create such a seminar on fluid dynamics, not CFD, the idea is old, but CFD one would be good too, and I would be glad to share more details I thought of about such a thing. You are in a much better position to make it, and I could learn from you, then

Also, we can discuss all this here first. Any comments are welcomed.

Sergei

Hi,

First of all I would like to say that I think it is a brilliant idea to have a forum for Commercial codes. I do not agree with you John, that you will get the best answer from the vendors. If it is details about code implementation, then I think that the vendors are the right forum. But for other question as "how hard is it to set up a radiation case", then the users of the code are the ones to answer the question. Also, discussions about problems and things that do not work as the vendor says, could be discussed in such a forum.

Regards Anders

(1). The nice thing about the forum on Internet is that both sides are shielded against the direct contact. So, the question and answer is not really directed toward individuals, but rather to the person surfing the Internet. (2). With this in mind, one can look at the issue more clearly. In other words, forum on Internet is the opposite side of e-mail. The e-mail is one-to-one, and is highly emotional. On the other hand, the forum is not. (3). This difference must be observed by all parties, in order to share the information. It is a library environment, and the person who answer the question is like the librarian try to help finding the information in the library. The same is true for Yahoo, AOL, Netzero,...... They help the readers to find the information. (4). For commercial CFD codes, the situation is different. By definition, the contents of the code and the instructions must be hidden from the users, unless the users have certain purchased right to it. The same is true, that one does not go to the library and ask the librarian to help him writing a windows program using VC++. (5). My feeling is that Internet forum is like an extension of a library, a distributed library, where the librarians are not located at one place. (6). So, if one is interested in learning, a school is a better place. If one is proud of his research results, a technical journal is a good place to present his work. And if one is anxious to let others know what he is doing, and can't wait for publication delay, a technical meeting would be a faster route to go. (7).The current problem is a very simple one. The business practice of "limited editions and limited circulations" is making the answers hard to come by. (8). At the same time, I have also been using various demo programs on the PC side of applications. And the solution to it is to buy some books related to the topic, such as introduction to program abc...(9). I must say that there is a market of writing a book on "How to use the commercial CFD code-abc", or "Practical solutions to the commercial CFD code-xyz". So, why not put together your users experience and write a book, in this way, everyone will be happier. (10). By the way, I think, a book such as " Learn how to use commercial CFD code in one hour " should be ideal for the newcomers to CFD. (?)

(1). I agree with you. (2). But, then it is up to the experienced users of the code to decide when to answer a question. That is basically the current condition. (3). Few questions will be answered, when experienced users are very busy. (4). Any answer is better than no answer, I guess.

Are you implying that people using commercial (CFD) code are not members of the CFD community? If so, what is your definition of CFD community?

Personally, I think this forum is great. I look at the postings and read the topics that interest me. These topics range from academic CFD to the Commercial CFD. If you have different fora, I will have to go to different pages to look for things that interest me instead of looking for them in one place. It is a great idea to have this forum, so please leave it alone.

No, not at all, but detailed technical questions related only to one specific code are not that interesting for those who do not use that code - I'm not that interested to read about how to set a "swirling mass flow inlet" in code XXX if I don't use that code. Also these questions today tend to dissapear in the high volume of this forum, and readers here often answer them a bit rudely with a short note to "call support" instead. With a dedicated forum these messages will stay longer and there is a higher chance that another user of the same code will see it and answer it. This could foster a more focused interaction between users of a specific code. That is something which would be good for everyone - users of the code get a more undisturbed forum where they can share ideas and experience, and perhaps get to know eachother. There will also be more room on this forum for general CFD topics. I'm not saying that we houldn't discuss commercial codes here. General topics still belong in this forum, like questions about which code that is best for a specific problem etc.

If the new forums don't work out we'll just remove them and go back to like it is today.

Yes, I've though about doing something like that. A first step could be to arrange some kind a "live hearing" where a few authorities in a field are invited to participate in a "panel debate". This could be done today with in the same format as this forum. Any suggestions on a good topic and people to invite?

However, as you point out it would be very good, almost necessary, to somehow have the ability to write equations, share plots and "point" on things etc. This is something that is a bit difficult to do. I'm not aware of any good systems that are wide spread which enables you to do this today. Within a year or two I'm sure that there will be Java applets to do these things.

Dear Jonas,

I support your idea. I would like further suggest that the data format of the forum should be so enhanced that the people can attach or directly insert graphs in their messages, since one can explain something more efficiently with graphs.

X. Ye

Etant physisien de grande renommée je puis dire que vote site est l'un et je pèse mes mots,des plus chiants parmis les plus chiants.

(1). The current state of the CFD-online forum is good enough. (2). I think, one has to realize that CFD is a high tech field. And one way to find out is to count the number of PhDs in a CFD code vendor. (3). So, the problem a user is interested in, would no doubt, be a technically challenging one. (4). For this reason, there are CFD code vendors trying to take advantage of this business opportunity. So, they are willing to sponsor this website to reach the bigger users market. (5). For the same reason, there are consulting firms trying to provide a solution to the users problem, using whatever CFD codes available. They depend on their experience and reputation to survive. (6). Then, there is the user himself, trying to solve some design problems for new products or improved products. He is not alone, becuase the need to solve his problem comes from the competition in the market. Now we are talking about the copyright, patents, etc. (7). So, the nature of this field tends to limit how much one can say in the forum. Why should one consultant give away his trade secret in solving a particular problem to other consultant in the forum? (8). Why should a user reveal his problem completely in the forum, so that the whole world knows what his company is heading? (9). Not to mention the vendors of CFD codes, they have to recover their investment first from the sale of the codes. (10). And no bookstore is going to let you read the whole book first then decide whether you are going to buy the book or not. (11). So, those are the limitations of the CFD forum. And if one is rich and willing to try out all commercial codes, then he should be free to do so, and have fun. (12). If NASA has a code just to compute the ds value at Y+=1, then I think, making a code user-friendly is just like a day dream. How many users really understand the physics behind the Y+=1? And how many are checking his results for Y+=1? And how many managers care about the results that is not Y+=1? And how many really understand the formula used by NASA? (13). I think, CFD technology is at the bottom of the cold war (from this discussion, it is at least three layers deep. so it is a very cold, cold war in technology) (14). For a forum to survive, I think, it should work like a newspaper stand around a corner of a busy street in a big city. For a lost tourist, one can always provide some directions, but not the actual help of how to get there. (since you will have to take care of your news stand all the time. ) (15). Watching the lost tourists, is fun; providing directions, is also fun; and as a tourist being lost in a big city , is fun to remember.

One has to admit that this web site has helped the people to obtain the information what they want to obtain. But nobody and nothing is perfect.

X. Ye

Dear Jonas,

>>A first step could be to arrange some kind a "live hearing" where a few authorities in a field are invited to participate in a "panel debate".

I am not sure about debates, really. What is an incentive to take part in a debate? There is a risk of being beaten! Also, it might be too emotional. There are, however, strong incentives to give a presentation or comments, at least for people from academic community. These are:

For a speaker: 1) to get help (more usual for seminars in Russia than in UK so far, but should work everywhere, really, this depends on the seminar structure, namely, on the length of discussion, which is the part most useful for the speaker. In Russia, typically, 1 hour talk is followed by 1 hour discussion, in UK discussion takes much less time. I do not know about seminars in other countries ( ). Here, on the web, discussions can be long.) 2) to announce and establish a priority claim on new results. 3) To improve ones reputation, in other words, seminar presentation is a publicity.

For a participant giving a comment. 1) Making friends. This is sometimes very important for one's career! 2) Publicity, too. Seminars are visited regularly by a large number of people. The number of good comments one can give is much greater than the number of good papers one can publish . 3) The last but not the least. Whatever the consequences, helping others is a Good Thing.

For many people, I believe, these reasons more than counterbalance those points John mentions concerning the competitiveness in CFD. They certainly do, for me. From what I read here, I have a feeling, that they do for John, too . If you are reading this, John, how is it?

>> Any suggestions on a good topic and people to invite?

Kghm, I am in doubt about a debate, as I told, but there is one suggestion:

At ICIAM99 in Edinburgh this summer Siamak Hassanzade of Sun Microsystems talked on Java Grande, almost promising that in a couple of months we all would switch to Java for high-performance computing, using a browser as the main tool for everything and the entire Internet as a distributed computational environment (well, I am slightly exaggerating, but it is part of the show ). And I almost believed him! Now, more than two months passed and I still run my fortran, not java, code on a single parallel computer, not on the Internet. It is a high time to ask Siamak why. And the person who could be asking (and answering) questions could be Jack Dongarra, since he was there, too, and he either works or doesn't work on turning linear algebra packages into something compliant with Java Grande, and everyone would like to know what he thinks, I believe. Along these road, several other questions can be answered by them. For example, if what is ahead is indeed distributed computations over the net, should we write our (planned for 3 years ahead) codes using MPI or will we be provided with shared-memory implementation (say, OMP) on the Internet? Questions can be collected here in advance, too.

==================
>> I'm not aware of any good systems that are wide spread which enables you to do this today. Within a year or two I'm sure that there will be Java applets to do these things.

Kghm, about three years ago when I searched there were so many such applets I could not even check all of them. Search on excite for +whiteboard conferencing mathematics, it should do. The question is how good they are at the moment, of course, and whether they are exactly what is needed. (Probably, a student can write an applet .)

================== Finally, if not a debate then what? Basicly, a seminar should be arranged like this:

For speakers and participants no other software except a browser (and e-mail) should be needed.

1). The speaker presents a talk by e-mail in a form of a HTML document, or postscript, or pdf file, or LaTeX file, whichever possible. The size and the content of the talk should be such that it could be read and understood in 1 hour. The seminar secretary posts this talk, as a www page, say, for a week.

2.) The seminar participants can ask questions during, say, three days, using a message board. Facilities for writing formulae and for drawing are desirable although not absolutely necessary. The author MUST often (may be at prearranged times) visit the message board during these three days and answer the questions. The participants are requested not to post comments at this stage.

Ideally, questions and answers should be somehow inserted into the talk, modeling questions asked during the speech, not afterward. This is quite useful for later readers.

3.) The rest of the week, again, using a message board, the seminar participants give comments and discuss the results. They are expected to help the speaker with their comments, although a constructive criticism is allowed.

Seminar success depends very much on forming a habit to attend the seminar regularly. It is highly desirable for a seminar to be held weekly.

-------------

Yours Sergei

(1). I don't have any comments on your message. (2). I think, any suggestion is better than no suggestion.

You have nice visions. This kind of online seminar would surely be very interesting and could develop into something big and very useful. However, to be honest, I think that this is a bit too big for CFD Online today. I wouldn't have time to arrange something like this every week - I don't have any secretaries or students who can write java applets for me ;-) I'd need help to pull something like this off.

Have you ever seen anything similar in another field? If there is a working system that we could use that would make the whole thing more feasible. I searched a bit for java applets that would suit this kind of seminar/forum and I didn't find anything that looked good. Basically all I found were a few "whiteboards" that allowed you to draw small pictures and share them with others. There seems to be a few more ambitious research projects aimed at this kind of publication/collaboration though, but I couldn't find anything that was finished and usable now.

cher E. Albert, dit "Physicien de grande renommee",

Jonas etant un de mes collegues de travail je considere comme un devoir de repondre a une critique aussi insultante dans le language qui est le mien.

1: avec un tel titre " physicien de grande renommee", je ne peux comprendre pourquoi votre critique est formulee en francais et non pas en langue shakespearienne qui est le language officiel de ce site. Ou bien etait ce pour souligner encore plus votre mepris pour ce site.

2: Ainsi, cela est vrai; Ayant passe mes annees de recherche a l'etranger, je ne peux affirmer connaitre l'etat d'esprit de la recherche en France. Mais les rumeurs qui me sont parvenues aux oreilles s'averent veridiques: Les "grosses tetes " meprisent le "vulgarus populus" scientifique. Dommage !Mais dites moi qu'etiez vous avant d'etre une "grosse tete". Un debutant peut etre qui a eu en son temps besoin d'un coup de pouce ou etes vous brillant de naissance?

Bref,... no comment

I do not know of any seminar like this. Sooner or later, there will be. Let us wait.

Sergei

(1). Time is money. Time is life. And life is limited. (2). I look at the forum as my hobby, something like painting, music, which will give me relaxed life. (3). It is a place where I paint, from time to time. I don't want to read the same thing several times, and I try to forget what I have painted. (4). So, it is not healthy for me to lock into an issue, week after week. (5). I think, it is a place where one can easily say a few words, or a page of words without any stress or resopnsibility. (6). Forum is something like a modern painting. It can be easily produced. The reader does not have to understand it, or even like it. It is the mark of time, somebody's footprint, on the beach. That's all.