CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Site Help, Feedback & Discussions (http://www.cfd-online.com/Forums/site-help-feedback-discussions/)
-   -   Fluent subforum (http://www.cfd-online.com/Forums/site-help-feedback-discussions/116906-fluent-subforum.html)

Far April 28, 2013 02:49

Fluent subforum
 
Hi All.

It is my observation that on Fluent forum there is large audience and therefore there are many question posted on forum on daily basis.

Here is the statics for Sunday April 28, 2013 for different forums.

Ansys forum 59 viewing (total for Fluent, CFX and Ansys meshing forum)
Fluent 29 viewing
CFX 15 viewing
ANSYS Meshing Forum 9 viewing

Main Forum 9 Viewing

floefd-floworks-flotherm 2 viewing

Numeca 2 viewing

Openfoam 16 viewing

Phoenics 1 viewing

SU2 1 viewing

CD-adapco 2 viewing

In addition to that no of threads/posts posted on Fluent forum on daily basis is way higher than any other forum and therefore threads go away unanswered.

What do you think it would be better to split Fluent forum into sub-forums.

What do you think about the following sub-forums? Here is my idea about sub-forums and needs discussion:

1. General discussion
2. Transient problems
3. UDF related issues (already there)
4. Multiphase
5. Combustion
6. Turbomachinery, wind turbine and mixing tanks
7. Heat Transfer
8. Convergence related issues ( I am not sure about this since in all areas of cfd convergence can be treated differently. Similarly general forum can create confusion)
9. Advance topics (Rocket propulsion etc)
10. Transition modelling
11. Cavitation
12. Solver settings

Thanks

oj.bulmer May 13, 2013 07:17

Subforums seems a good idea, but too much fragmentation may not be helpful. It may confuse the members while posting since their problem can be classified under many headings, at the same time, may discourage the members from visiting posts on unfamiliar topics.

The most appropriate way to do this classification can be understanding the statistics of different topics on the FLUENT forum. But in the absence of the data, the ones below seem relevant:

1) Multiphase
2) Heat transfer
3) UDF (Already there)
4) Turbomachinery
5) Miscellaneous

First four are specialized topics and folks who excel in each of these topics can advance the discussions there. All rest can be classified under miscellaneous, since there may not be frequent questions related to niche topics (eg combustion, rocket propulsion etc) anyway. The owners who want answers for old/unanswered topics can always reply in the post to kick it to the front page.

Cheers
OJ

pete May 13, 2013 12:41

I agree that we should not split it up too much. Statistics on which different topics are discussed would be good to have. I browsed through the current forum and another subforum division could be something like:

  • Fluent (main forum still there)
    • Solver, numerics & convergence
    • Modeling (turbulence, transition, combustion, sprays, cavitation, ...)
    • Applications (turbines, cars, heat exchangers, ...)
    • UDF & Scheme

This kind of division would split the content up in a different but also logical way: solver, modeling & applications. Most questions are mainly in one of these topics I guess, but sometimes it can be difficult to judge if the convergence problem someone has is related to numerics, the turbulence model used or the application modeled :confused:. We could ad more specialized topics like UDF also. What do you think about this?

Far May 14, 2013 23:02

Dear Pete and OJ

I agree that too much division will make things difficult.

Following are few points to consider :

1. Many threads remain unanswered due to large volume of new threads being posted on daily basis which moves slightly older threads to next page and therefore gets less attention.

2. By making new sub-forums it is possible that the experts of each field will go to subforum directly and reply to particular thread. Moreover each thread will get more time on the front page.

3. I agree with the sub-forums proposed by Pete.

4. A poll can be added on Fluent forum to get detailed survey from forum members.

5. Statistics of different type of questions posted can also help in deciding the sub-forum division

More discussion is needed.

pete May 15, 2013 09:51

Browsing through the threads in the Fluent forum again it is not easy to find a good sub-forum division that will make it easy to select a proper sub-forum. I've added a few more sub-forums to my list. Do you think that the following forum division will make it easy for users to select a good forum for their questions? Or would you prefer a different type of sub-forum division? Is it good to add a Multiphase Flow forum? Should we also add a Heat Transfer forum?

Fluent (main forum still there)
  • Interface: Case setup, running, handling, post-processing, ...
  • Numerics: Solver, schemes, convergence, ...
  • Modeling: Turbulence, transition, combustion, sprays, cavitation, radiation, ...
  • Applications: Turbines, wings, cars, electronics, pipes, heat exchangers, ...
  • Multiphase Flows
  • UDF & Scheme

macfly May 15, 2013 13:17

Against the sub-forums
 
1 Attachment(s)
Hi all,

What do you prefer, having one dictionary or having many sub-dictionaries to look for words? The answer is clear to me. Keep it simple please.

However, the search bar could definitely be optimized:
- It could be always ready to write in it, without clicking some down arrow.
- Options could be available right away like in attached figure, instead of 'advanced search' annoying process.

Just my 2 cents.

pete May 15, 2013 15:10

The problem with keeping everything in one forum is that the volume in the Fluent forum means that a new message will now only stay on the front page listing for three, max four days. Once a message is not on the front page replies are less frequent and many messages are being left unanswered. I do not know if this will improve with a few additional sub-forums with less volume, but it might be worth a try. I still think that we should keep the main Fluent forum though.

sbaffini May 16, 2013 08:04

I would like to add my two cents to the discussion.

I agree on the idea of splitting the Fluent Forum and i also undertsand the reasons determining it but i would also consider the typical user behavior. I'll explain better.

I consider myself an active user of both the Main Forum (which i always intended as a general CFD forum not related to any of the softwares with dedicated forums, so pure CFD and related techniques/theory with no reference to any specific code) and the Fluent forum (everything related to Fluent).

Now, if you look at the quantity of Main Forum posts with Fluent related questions, nomenclature etc., you will notice how the relevant number of Fluent users in the forum (and also their background, let's say this) tends to strongly promote this annoying pollution.

So, as a matter of fact, i would suggest a splitting that, while helping Fluent users, also avoids polluting the Main Forum (expert Fluent users know nothing about, say, the PRESTO! scheme... how could a non Fluent user????).

In this respect i think that:

- having something like a "Main Fluent Forum" is better than a "Miscellaneous" sub-menu under a general "Fluent" Group. The "Main" keyword should promote convergence in my opinion.

- The general Main Forum should also reflect the fact that it is not a space for everything not related to the other forums, otherwise everybody will tend to go there. CFD is not vendors software, it's numerical methods, physical modeling, programming issues. The "Main" word should not be random.

- The greatest number of posts is still on some convergence issues. Convergence should require a related forum by itself... too bad that people with answers will likely tend to avoit it at all ;).


In conclusion i suggest a subdivision similar to the last one by pete:

Ansys -> Fluent:
  • Main Fluent Forum
  • Theory: anything related JUST to the theory behind the code
  • Interface: Case setup, running, handling, post-processing
  • Convergence issues
  • Numerics: Solvers, schemes, boundary conditions
  • Modeling: Turbulence, transition, combustion, sprays, cavitation, radiation, heat-transfer, multiphase
  • UDF & Scheme
A different organization, more directly related to the Fluent options or menus may not be bad (Let's face it, that's what all the questions actually reduce to).

Finally, as i said before, i would change the description for the general Main Forum, stating that it is ONLY for the general techniques and NOT for the stuff which do not suit the other forums


Regards

pete May 16, 2013 08:25

Quote:

Originally Posted by sbaffini (Post 427995)
I would like to add my two cents to the discussion.
- having something like a "Main Fluent Forum" is better than a "Miscellaneous" sub-menu under a general "Fluent" Group. The "Main" keyword should promote convergence in my opinion.

Yes, the top Fluent forum, above the other sub-forums, should still be there and serve as a Main Fluent forum. Is this not what you want?

Quote:

Originally Posted by sbaffini (Post 427995)
- The general Main Forum should also reflect the fact that it is not a space for everything not related to the other forums, otherwise everybody will tend to go there.

I'm not sure I understand this. If we only allow general non-software related questions on the main forum, then where do you suggest people who have questions about something that does not have a dedicated sub-forum should post their questions? When we see a lot of discussions on the main forum about a certain topic we can start a dedicated forum, but until we see that I think that all topics should be welcome on the main forum. What is not welcome are topics that already have a dedicated forum. Today the description of the Main CFD Forum is "Topics that do not have a dedicated forum below". It is difficult to be more clear than that. If you want a more silent forum for pure CFD research related questions we could discuss opening such a forum.

Quote:

Originally Posted by sbaffini (Post 427995)
- The greatest number of posts is still on some convergence issues. Convergence should require a related forum by itself.

Yes, a lot of questions are about convergence and this might warrant a separate sub-forum. Thanks for your suggestion.

You suggested a "Fluent Theory" forum. I also thought about this, but browsing through the current forum there are not a lot of such deep theory discussions. Would it not be okay to keep these discussion in the main fluent forum or in the appropriate subforum (numerical theory in numerics, modeling theory in modeling, ...). Or perhaps pure theory discussions, as you suggest, warrant a separate more silent theory forum. I'm hesitant.

sbaffini May 16, 2013 09:03

Quote:

Originally Posted by pete (Post 428001)
Yes, the top Fluent forum, above the other sub-forums, should still be there and serve as a Main Fluent forum. Is this not what you want?.

Yes, i overstressed it, but i think that the specification that it is a "Main" Fluent forum would help convergence of posts.


Quote:

Originally Posted by pete (Post 428001)
I'm not sure I understand this. If we only allow general non-software related questions on the main forum, then where do you suggest people who have questions about something that does not have a dedicated sub-forum should post their questions? When we see a lot of discussions on the main forum about a certain topic we can start a dedicated forum, but until we see that I think that all topics should be welcome on the main forum. What is not welcome are topics that already have a dedicated forum. Today the description of the Main CFD Forum is "Topics that do not have a dedicated forum below". It is difficult to be more clear than that. If you want a more silent forum for pure CFD research related questions we could discuss opening such a forum.

I see your point. Maybe, something like a "Miscellaneous" would help. Or maybe is just me overrating the "Main" word. In that case, a pure theory (not necessarily research) forum could be well suited. However, consider that the Main Forum is already similar to this, with questions like: "How do i discretize this?", "Why this method doesn't scale as expected?" etc. The only exception, from what i can see, is actually what belongs to other forums.


Quote:

Originally Posted by pete (Post 428001)
You suggested a "Fluent Theory" forum. I also thought about this, but browsing through the current forum there are not a lot of such deep theory discussions. Would it not be okay to keep these discussion in the main fluent forum or in the appropriate subforum (numerical theory in numerics, modeling theory in modeling, ...). Or perhaps pure theory discussions, as you suggest, warrant a separate more silent theory forum. I'm hesitant.

This point was still suggested by the fact that some times i see posts in the Main Forum which are mostly related to Fluent schemes.

Let me put it in a different way. I feel that most of the wrong posting on the Forums is due to the Fluent users, because they are more than Others and, mostly, newbies. My idea was to provide more clear "Road Signs" for them which would both slow down the post scrolling in the actual Fluent forum and, more importantly, rationalize the posting.

I think it is noteworthy that wrong posting was lower before the creation of the ANSYS general category (pre 2009 switch).

pete May 16, 2013 09:20

Quote:

Originally Posted by sbaffini (Post 428011)
Let me put it in a different way. I feel that most of the wrong posting on the Forums is due to the Fluent users, because they are more than Others and, mostly, newbies. My idea was to provide more clear "Road Signs" for them which would both slow down the post scrolling in the actual Fluent forum and, more importantly, rationalize the posting.

I think it is noteworthy that wrong posting was lower before the creation of the ANSYS general category (pre 2009 switch).

I agree, most wrong postings are by Fluent users. I have also noticed that these wrong-postings come in waves. It seems like many teachers point their students to cfd-online.com and instruct them to ask questions there. These students are so new that that they don't even know that CFD is not the same as Fluent. When a new class has started a new wave of newbie Fluent questions are spread throughout the forums on CFD Online.

Handling this is difficult. Perhaps a sticky post on top of the main forum would help, but I doubt if these newbies would even read that before posting. We moderators have started to be more strict about this and now frequently give out infractions to users that cross-post and repeatedly break the forum rules (http://www.cfd-online.com/Forums/misc.php?do=showrules).

The fact that there were less incorrect posts back in 2009 does not say that much. Back then we had much fewer forums, half the traffic, and a much more stable user-base with less newbies arriving.

pete May 16, 2013 09:34

I just added a new sentence on the main forum description: "Do not post Fluent questions here". I hope that can help reduce the number of incorrect Fluent & Gambit posts in the main forum.

Aeronautics El. K. May 17, 2013 06:55

Maybe it is a good idea to create some sub-forums in the Fluent section but I believe there shouldn't be many sub-forums. Someone's problem may be subject to more than one of these categories and this will cause confusion and chaos.
Besides, it would be quite frustrating for someone to navigate through all these forums and subforums especially if he/she has limited time to spare. So, you should keep it straight and simple.

I suggest you create only few sub-forums like that:
  • Internal Flow Problems
  • External Flow Problems
  • UDF
  • FSI

This should reduce the bulk of posts in the main Fluent forum while at the same time it covers most of the aspects of the software and the forum remains simple and user friendly.

diamondx May 17, 2013 11:57

I will say :

1) Aerodynamics
2) Combustion
3) Heat transfer
4) Multiphase
5) UDF
may be a sixth one for misc, or mixed...

Convergence issues is met everywhere, so if one has convergence issues with his naca 0011, he will just address the problem in the aerodynamics section...

pete May 18, 2013 06:50

I like your short list of application areas Ali. I am very hesitant if it is best to create sub-forums based on applications as you suggest Ali or based on the part of the process, as I suggested (numerics, modeling, ...).

Far May 18, 2013 12:19

My vote is for the sub-forums listed by pete and sbaffini (Post # 5).

To realize the advantage of sub-forums, you can take a look on UDF sub-forum where people get reply even after weeks, but it is sure that they get reply because posts are on front page for couple of weeks

diamondx May 18, 2013 12:24

Quote:

Originally Posted by pete (Post 427753)
Browsing through the threads in the Fluent forum again it is not easy to find a good sub-forum division that will make it easy to select a proper sub-forum. I've added a few more sub-forums to my list. Do you think that the following forum division will make it easy for users to select a good forum for their questions? Or would you prefer a different type of sub-forum division? Is it good to add a Multiphase Flow forum? Should we also add a Heat Transfer forum?

Fluent (main forum still there)
  • Interface: Case setup, running, handling, post-processing, ...
  • Numerics: Solver, schemes, convergence, ...
  • Modeling: Turbulence, transition, combustion, sprays, cavitation, radiation, ...
  • Applications: Turbines, wings, cars, electronics, pipes, heat exchangers, ...
  • Multiphase Flows
  • UDF & Scheme

That's a nice classification too, but recommended for advanced users who knows the differences and steps in CFD . I also think that interface , numerics, and modeling can all be part of applications, it will bring confusing for beginners in CFD. I'm sure one will go address his convergence problem in the wings application. In my classification, the application can be most of the time independent. That is my opinion.

oj.bulmer May 21, 2013 07:07

Quote:

Originally Posted by pete
I like your short list of application areas Ali. I am very hesitant if it is best to create sub-forums based on applications as you suggest Ali or based on the part of the process, as I suggested (numerics, modeling, ...)

Exactly, this is the notion we are wrestling with, and what I was also pondering over before posting. I will take a macro view here, and then break it down:

Quantitative (equal) division of posts:

This thread started with this objective. The nature of subforums in this way will strictly be guided by statistics of the posts appearing everyday. This approach will make sure that at any day, the no. of posts remaining on first page and the no. of posts going on second page are more-or-less the same, so yes, this solves the original problem. But it is not clear if it will ensure that the experts will regularly author the posts and advance the discussions, since it may be difficult for them to find the posts with the subject of their liking in all these subforums. Moreover, it is difficult to obtain the statistics in the first place as pete suggests.

Qualitative (categorical) division of posts:

a) Process oriented
While it may make sense to create it based on the process the way pete, Far and others prefer, it is worthwhile to note that convergence problem of heat exchanger and that of an airfoil have a completely different nature. And if users having convergence problems from all applications and areas keep posting in subforum Numerics, it will still not be a "specialized" forum, and rather a mixture of convergences problems of different types. Authors will have to sift through the posts to hit on the posts on their favorite subjects.

b) Application/subject oriented:
Instead, if the classification is done on the basis of application subjects like diamondx and I suggest, then the convergence problem of airfoil is posted in say Aerodynamic forum, and that of heat exchanger is posted in say Heat Transfer forum, and the experts in that area, familiar with nature of problems in that area, can cater to the questions. The questions not classifiable still have a place on Main/Miscellaneous (or whatever name) Fluent forum. An expert with a broad understanding of Solver theory and numerics may still fall short in answering convergence issues for all applications (combustion, aerodynamics, multiphase), since the nature of convergence issues varies in these. But, an expert in aerodynamics is more likely to answer most process-oriented questions (modelling, convergence, case setup etc) in aerodynamic applications. Don't you think that this way, the questions may be answered more rapidly?

OJ

Far June 2, 2013 12:10

What is the final opinion?

one forum or sub-forums?

jola June 4, 2013 15:04

Looks like no final opinion was reached. What would you recommend Sijal?


All times are GMT -4. The time now is 18:16.