[rdbisme]

Hello,

in the OpenFOAM discussion subforum we discussed a bit which kind of modifications or improvements would be useful to the cfd-online community to be more useful and more effective. Even if the discussion was initially focused on OpenFOAM support, the discussion might be of wider breath if done on the global community perspective.

I would love to widen the public for this discussion and hear about your opinions.

What I was suggesting was to rethink the format of the community migrating to a Q&A workflow (as StackOverflow). For example FeniCS community is using a format like that for User support (and a slack channel for dev). IMO that format enhance the quality of the answers and the questions due to the interaction associated to voting. It really helps to highlight the most important questions from users and avoids content duplication.

If the forum is still the most favored option for the majority of the members here, few suggestions would really help to improve the quality:

• A stricter (and clear) policy for thread creation
• A bigger team of moderators to enforce the policy
• Markdown and Mathjax syntax
• Advanced content subscription (i.e. subscribing not only to topics, but also to tags, etc...)

In particular for the forum part, I would take inspiration from xda-developers.

What do you think? What would you suggest about improving the CFD-online experience?

[jola]

If people started to use CFD-Wiki more frequently I think that it could be even more valuable to save information there than in a Q&A site. A Q&A site needs maintenance to keep answers current and moderators would have to do a lot of work to merge similar questions. Otherwise a Q&A site will have the same problem as we have on the forum, that the same answer can be found in many threads. Several forum users have written Q&A pages on CFD-Wiki that they use to refer new users to instead of writing the same answer many times.

[rdbisme]

Originally Posted by jola

If people started to use CFD-Wiki more frequently I think that it could be even more valuable to save information there than in a Q&A site. A Q&A site needs maintenance to keep answers current and moderators would have to do a lot of work to merge similar questions. Otherwise a Q&A site will have the same problem as we have on the forum, that the same answer can be found in many threads. Several forum users have written Q&A pages on CFD-Wiki that they use to refer new users to instead of writing the same answer many times.

Well, I must admit I'm not a huge fan of wikis. I wrote tons and tons of documentation in the associations/companies I've been, bu nobody seemed keen on reading general, often big, documents on the topics. Without considering that writing a good wiki article is super difficult and takes a lot of time. In addition to that OpenFOAM suffers a segregation of informations that is quite annoying: there are plenty of sources for support: CFD-Online, OpenFOAMWiki, ESI website, cfd.direct, local working groups outdated wikis, ... I imagine that the same situation exists for other softwares.

I believe that the approach of very specific questions and very specific answers is easier to maintain. I actually do not agree that it needs lots of work to "merge" questions. Questions must be very specific so they not need to be merged. Moreover the quality improvement of the community is actually boosted by the reward system (take StackOverflow as an example) that grants you more privileges more you contribute to the community.

The problem with the Wiki approach is that people, even more often in our field, are not really in the mood of searching for what they look for in a general wiki article. They want to ask specifically what they have in mind and hopefully get an answer only on that. The Q&A website really enhance this experience, allowing users to easily find questions that match theirs, without the need of lurking on several pages of thread.

Moreover in those kind of websites the accepted answer is generally guaranteed to be working and correct. And is somehow "peer-reviewed". On the forum you often need to lurk inside the threads and check which answer is correct: this is a waste of time.

I hope to get more contributions to the discussion from other members of the cfd-online community...

[tiam]

I am not sure Q&A would always work. I a forum thread is sometimes exactly what one needs. But there needs to be better moderation. Perhaps, some kind of reward system that would give people more privilleges automatically like at Stackoverflow?

[Tobi]

Hi all,

it is nice that this discussion came up again but there are a few things we should really think about. I want to mention that I do not have too much knowledge in the topic but I want to contribute my personal opinion.

I am not sure Q&A would always work. I a forum thread is sometimes exactly what one needs.

I agree. Sometimes it is more efficient to have a discussion rather than the Q/A approach. However, this can also lead to messed up thread such as the one I started a while ago named libOpenSMOKE with 520 replies. To be honest, no one will read through the entire thread. However, it always depends on the question itself. Personally I would say that it is not an easy task to destinguish between simple and advanced questions and somehow move the thread and combine them. The later one requires in addition that a moderator is aware of most of the threads (in order to know if similar questions can be merged).

If people started to use CFD-Wiki more frequently I think that it could be even more valuable to save information there than in a Q&A site. A Q&A site needs maintenance to keep answers current and moderators would have to do a lot of work to merge similar questions. Otherwise a Q&A site will have the same problem as we have on the forum, that the same answer can be found in many threads. Several forum users have written Q&A pages on CFD-Wiki that they use to refer new users to instead of writing the same answer many times.

A wiki is a good idea but actually it depend. I made a wiki entry of the libOpenSMOKE thread in order to sum up different pages to give an overview of the big discussion. The site was accessed almost 8.000 times - so maybe it was worth it but it is not meaningful in each topic. Additionally, it takes time as already mentioned.

In addition to that OpenFOAM suffers a segregation of informations that is quite annoying: there are plenty of sources for support: CFD-Online, OpenFOAMWiki, ESI website, cfd.direct, local working groups outdated wikis, ... I imagine that the same situation exists for other softwares.

Totally agree. I am hosting a OpenFOAM related page too, but I try to link all my content to the:

• openfoam forum
• openfoamwiki
• wiki.openfoam.com

and try to keep everything up to date. However, this is a tough task and after changing anything at the website (such as links), I have to take care of all other sources which is a full time job for a few days. However, it is as it is and always new platforms and places are generated. Especially the problem with outdated or non-accessible sources is a big issue. People once start to contribute but after a while it stops based on the huge amount of time one has to investigate.

If the forum is still the most favored option for the majority of the members here, few suggestions would really help to improve the quality:

• A stricter (and clear) policy for thread creation
• A bigger team of moderators to enforce the policy
• Markdown and Mathjax syntax
• Advanced content subscription (i.e. subscribing not only to topics, but also to tags, etc...)

That sounds nice but a stricter and clearer policy for thread creation depends on the moderators. E.g. in FOAM a lot of people do post their questions in the general forum. Moderators always have to move the content and/or merge things. This is time consuming. Sure, at the beginning everyone is doing it with motivation but after month it seems to be a time wasting task and for that, we need to have more moderators!


The subscription to tags would be a great deal and markdown + matahjax too. I would also like to have a "SOLVED" item in order to directly see, if a topic was solved or not. This would be great in combination with the reputation option.


Again, I am not so familiar with the topic and it is my personal opinion. To sum up:

• Wiki entries are nice to have but takes time
• Stricter policy requires more moderators or some way that people will read the policy (e.g. in the validation thread in FOAM, people ask also questions which is not the topic of the thread)
• A discussion about a topic can be better than the Q/A approach - but this depend on the question
• More plugins to the forum would improve it such as markdown or mathjax
• A tag system would be nice in order to subscribe to special tags
• 'Thread Solved' mark would allow people to see directly if the question is solved or not (maybe one could sort the topics too, or we move threads automatically to solved one - however this would also take more time - guess it is a bad idea)
• More people should take care to have all information focused on one page which will not happen

I wish you all a nice sunday. My body tells me know to go for a run

[jola]

Many great ideas. One thing that we should perhaps try to look at right away is the "Solved" mark for a thread. I assume that this should be set either by the person who started the thread or by a moderator.

Concerning MathJax. Both the forums and CFD-Wiki have for many years had LaTeX functionality. So this feature is already available, see Guide: Writing Equations in LaTeX on the CFD Online Forums for more information.

The problem with many of the ideas is that it requires a significant moderation effort. Last week we got more than 800 messages posted to the forums. We have more than 10 moderators, but only a few of us can devote the time needed to review all these messages and handle the most basic tasks like removing spam. Moving posts, reorganizing threads, correcting and communicating with users takes much more time. Remember that I'm talking about more than 100 messages per day... the system will have to do a lot of things automatically. Is anyone aware of a Forum or Q/A system that has the ideas that you suggest? With the traffic we have it must be a stable system which is already in production on many other sites.

[jola]

Originally Posted by tidusuper91

Well, I must admit I'm not a huge fan of wikis. ... nobody seemed keen on reading general, often big, documents on the topics.

This is not true for CFD Online. 30% of our traffic is to CFD-Wiki, that means that more than 180,000 wiki-pages are read every week by our visitors! CFD-Wiki is also very well indexed by Google. If you for example search for SST k omega on Google you will find our CFD-Wiki page on that turbulence model right at the top. Hence, taking the time to write a CFD-Wiki article, or a CFD-Wiki Q&A section, will make the information easily accessible for many users. When you look for information about something the first thing you try is often Google and the chance that you will then find any CFD-Wiki articles about the subject is high.

[rdbisme]

Originally Posted by jola

Many great ideas. One thing that we should perhaps try to look at right away is the "Solved" mark for a thread. I assume that this should be set either by the person who started the thread or by a moderator.

Concerning MathJax. Both the forums and CFD-Wiki have for many years had LaTeX functionality. So this feature is already available, see Guide: Writing Equations in LaTeX on the CFD Online Forums for more information.

The problem with many of the ideas is that it requires a significant moderation effort. Last week we got more than 800 messages posted to the forums. We have more than 10 moderators, but only a few of us can devote the time needed to review all these messages and handle the most basic tasks like removing spam. Moving posts, reorganizing threads, correcting and communicating with users takes much more time. Remember that I'm talking about more than 100 messages per day... the system will have to do a lot of things automatically. Is anyone aware of a Forum or Q/A system that has the ideas that you suggest? With the traffic we have it must be a stable system which is already in production on many other sites.

StackOverflow became actually a standard now. Also Google now indexes it differently on the search engine. If I should suggest something, would be to invest marketing effort to create a website on the StackExchange network. We already tried with OpenFOAM, but without success.
Or, we could use directly the StackOverflow engine, maybe via a crowdfunding if it's too expensive.

Here a list of Open Source clones of Stack Exchange: https://meta.stackexchange.com/quest...xchange-clones

Moreover I would also ask to promote more moderators. 10, as you clearly state, is not enough. We can do it democratically, with a poll on the forum, or as you wish.

Originally Posted by jola

This is not true for CFD Online. 30% of our traffic is to CFD-Wiki, that means that more than 180,000 wiki-pages are read every week by our visitors! CFD-Wiki is also very well indexed by Google. If you for example search for SST k omega on Google you will find our CFD-Wiki page on that turbulence model right at the top. Hence, taking the time to write a CFD-Wiki article, or a CFD-Wiki Q&A section, will make the information easily accessible for many users. When you look for information about something the first thing you try is often Google and the chance that you will then find any CFD-Wiki articles about the subject is high.

I'm one those 180k readers . The problems is that maintaining a Wiki is time consuming, hugely time consuming. Writing an article on a topic is very difficult (and in our field often you prefer to write a paper or a book chapter than a wiki page... Generally). Also fellas from Stack Exchange tried to do a similar thing, look how it ended up: https://stackoverflow.com/documentation.

We're talking of a community of millions of active users with most of them really keen on contributing and open access to knowledge.

In my opinion what we should target is immediateness. "Git" style community: small and fast contribution vs single, monolithic, soon-to-be-outdated articles

[wyldckat]

Greetings to all,

Just to give a quick bump on this discussion and following up on what Anton wrote sometime ago on the 1.0 thread :

Originally Posted by akidess

On a side note, check out http://www.discourse.org/. It's a "traditional" discussion forum, but implements many of the learnings from Stackexchange, e.g. community moderation features.

A discussion thread was started at the VTK community on whether to switch to Discourse or not: https://www.vtk.org/pipermail/vtkuse...ch/101185.html

On that discussion mail-thread, as expected, there were several against and for the change.

• On one hand, there can be more intervention by the community: https://www.vtk.org/pipermail/vtkuse...ch/101219.html
• On the other hand, a lot of people prefer good old mailing lists...

Either way, Discourse is mostly just another forum software... it's beneficial for the projects hosted by Kitware, since they still use mailing lists. I have not studied if Discourse is better than the vBulletin being used here at CFD-Online, but its About page is very compelling...


Best regards,
Bruno

[tiam]

Originally Posted by wyldckat

Greetings to all,

Just to give a quick bump on this discussion and following up on what Anton wrote sometime ago on the 1.0 thread :

A discussion thread was started at the VTK community on whether to switch to Discourse or not: https://www.vtk.org/pipermail/vtkuse...ch/101185.html

On that discussion mail-thread, as expected, there were several against and for the change.

• On one hand, there can be more intervention by the community: https://www.vtk.org/pipermail/vtkuse...ch/101219.html
• On the other hand, a lot of people prefer good old mailing lists...

Either way, Discourse is mostly just another forum software... it's beneficial for the projects hosted by Kitware, since they still use mailing lists. I have not studied if Discourse is better than the vBulletin being used here at CFD-Online, but its About page is very compelling...


Best regards,
Bruno

An important point I think is that there is intrinsic value in saying "this project is not in the 90s". And unfortunately, this is the impression this, otherwise excellent, webpage and forum give. I won't even go into discussing mailing lists...

Julialang started using discourse some time ago, it looks quite nice and I think it is a good solution if one wants a forum.
https://discourse.julialang.org/