[SBusch]

Hello,

we just found out that OpenFOAM v3.0+has been released:

http://www.openfoam.com/version-v3.0+/

We are wondering why ESI-OpenCFD didn't just name the new version OpenFOAM 3.1.0?

Is OF v3.0+ an successor to OpenFOAM 3.0.1, or an alternative Version?

And if it's an alternative release: Why? Are there any more (maybe fundamental) differences besides the mentioned changes or what's the reasoning?

[Cyp]

Originally Posted by SBusch

Hello,

we just found out that OpenFOAM v3.0+has been released:

http://www.openfoam.com/version-v3.0+/

We are wondering why ESI-OpenCFD didn't just name the new version OpenFOAM 3.1.0?

Is OF v3.0+ an successor to OpenFOAM 3.0.1, or an alternative Version?

And if it's an alternative release: Why? Are there any more (maybe fundamental) differences besides the mentioned changes or what's the reasoning?

Apparently it is another fork..

[kmooney]

Interesting release! The new DES and U stability features are especially curious.

I'm having a little trouble understanding the 'plus' from a forking/development stand point so correct me if I'm wrong about how things are organized in the foam-verse at this point:

• Soft Forks (full/some cross compatibility)
  • OpenFOAM-3.x - OpenFOAM Foundation / Copyright holder (openfoam.org)
  • OpenFOAM-3+ - ESI/OpenCFD / Trademark owner (openfoam.com)
• Hard Forks (no cross compatibility (I'm guessing, haven't tried))
  • foam-extend
  • iconCFD
  • engys-os
  • caelus

Now I see in the 3+ release that a lot of new files are tagged with an additional

Copyright (C) 2015 OpenCFD Ltd

in addition to the typical OpenFOAM Foundation copyright tag found on a majority of the file headers. Does this change anything for me as a user that I should be aware of? My understanding is the GPL will generally trump any sort of copyright holdings due to the re-distribution type clauses right? Could I, for example, fork from 3+ and re-attribute the copyright holdings to myself? I'm not trying to be facetious, honestly just trying to get a better handle on how copyright/gpl/trademark etc. all work together.

I'm not privy to nature of the communication/collaboration occurring between the 'foundation' and the 'plus' developers (if any). I'm concerned that it would only take a few large architectural deviations in the 'foundation' release to break the 'plus' cross compatibility and cut the development pipeline. Consequently another hard fork further fragmenting the community and confusing new comers.

As a foamer veteran I can still keep track of the community/commercial/academic branch movements and I generally understand the motivations behind them. At the same time our collective voice and power as the foam community at large seems to be constantly splitting into more and more development/documentation/collaboration/commercialization efforts, none of which really have a critical mass. When I joined the OpenFOAM community back in 2009 I was quickly convinced that together we'd eventually bring CFD to the masses, upend Fluent/ansys/adapco, and change the fluid physics world as we know it. These days I'm a bit less optimistic.

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[kmooney]

Originally Posted by SBusch

Hello,

we just found out that OpenFOAM v3.0+has been released:

http://www.openfoam.com/version-v3.0+/

We are wondering why ESI-OpenCFD didn't just name the new version OpenFOAM 3.1.0?

Is OF v3.0+ an successor to OpenFOAM 3.0.1, or an alternative Version?

And if it's an alternative release: Why? Are there any more (maybe fundamental) differences besides the mentioned changes or what's the reasoning?

The OpenFOAM Foundation technically releases OpenFOAM-X.x.x. ESI just owns the trademark I think. I imagine this is a move to differentiate OpenCFD and ESI from the Foundation release; contracts, support, etc...

[Zeppo]

Originally Posted by kmooney

I imagine this is a move to differentiate OpenCFD and ESI from the Foundation release; contracts, support, etc...

What does it mean to us, end OpenFOAM users?

[kmooney]

Not sure really, I asked them here: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html

____
[Moderator note: This was posted before posts were moved from the announcement thread to this one. This post was made before post #3 was moved here]

[Sylvain]

Thank you Kyle for your comment which is very relevant.

We made a choice in our company back in 2010, and this choice was OpenFOAM. We truly believed in the development of this CFD code thanks to its community, and we bet on it for the future.

I tried to keep track on the branch movements as well, but I wasn't as successful as Kyle. Today I'm a bit confused by this "plus" release with additional contents compared to the 3.0.1. What does that mean?

I have to say that the communication is not very clear between what is done by OpenCFD and by OpenFOAM Foundation. The web sites are similar except from the color (blue vs green), the logos on linkedIn and twitter are nearly the same but different. What does that mean? Are they in good terms? Are they tied by a contract, they would want to split but they can't? What can we expect in the long term? Can we rely on OpenFOAM for a long term development of our simulation process?

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[egp]

Kyle, good commentary. Pretty much hits the nail on the head. The lack of respect for the community of users has pretty much turned me off. I'm going other directions. For routine complex-geometry CFD, I'm moving to star-ccm+. For research, we're evaluating other options ranging from organic code to high-order FEM using deal.ii.

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[michaelaye]

I'm a newcomer trying to get into OpenFOAM, deciding on a toolset for a multi-year project and find the current situation highly highly confusing. There's openFoam.org and openFoam.com, on Twitter there is @cdfdirect @cdffoundation and @openfoam. I tweeted to all three of them inquiring about the difference between .org and .com to no avail.
Kyle above asked seemingly very relevant questions on the day after the announcement post and 9 days later nobody bothered to answer.
Is OpenFOAM dead? Was there some in-house fight that lead to a split? What happens to the "Open" of OpenFoam if things go commercial? Is there maybe a thread where this is explained somewhere?

Thanks in advance for any insight.
Michael

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[alexeym]

@kmooney,

Concerning your question about fork: yes, you can do it, sources are licensed under GPL. In old files you need to keep two copyright lines (or left old one if you keep old sources intact). New files could be attributed only to you.

The rest is just a guess.

AFAIK after ESI bought OpenFOAM trademark (http://www.openfoam.com/about/) they were not quite sure what to do with it; so 3.0+ activity is just a sign - they finally made a decision on "what to do". Currently plus is just a set of additional components and, guess, it will remain like this, since big architectural changes could be tested in OpenFOAM-dev branch.

@michaelaye

For obvious reasons the forum is used as an announcement but not as communication platform. There is always https://www.esi-group.com/company/contact-us if you would like to ask questions about 3.0+ and http://cfd.direct/contact/ if you have questions about non-plus version.

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[LVDH]

After reading about the new features of OF+ I decided to try them as they sound very good and useful.

Now I can say that they add value but also new problems. So there is reason to use 3.0.1 and v3.0+(what a great version name).

So now what am I supposed to do as user? Have even more OF installations on my system and use them case by case?

At the OF conference in Stuttgart last year ESI asked the audience more than once what they would like from OF in the future as they claimed to listen to the community. The main response from the audience was to unify all the OFs as having all these forks is very annoying.

So how did they come up with adding these few new features into a new fork?
Could somebody please explain?

_____
[Moderator note: This post was originally made on the announcement page for OpenFOAM 3.0+: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - it was moved to this thread, in order to allow for a better discussion among the community.]

[pbachant]

Originally Posted by Zeppo

What does it mean to us, end OpenFOAM users?

It means you have yet another OpenFOAM to keep track of...

The fragmentation is confusing. In a perfect world there would be one group leading with enough flexibility that other organizations felt they could invest their time into development with a high probability of their work being merged in, rather than needing their own fork. The fact that virtually every fork uses a different Git hosting site, issue tracker, etc., certainly doesn't help.

For now, I am sticking with the OpenFOAM Foundation (CFD Direct) version(s), but may try out the plus version at some point. I fear discovering some great feature in plus that doesn't get merged into the CFD Direct version.

I recently learned that forks and variants are listed on the wiki: https://openfoamwiki.net/index.php/Forks_and_Variants

______
[Moderator note: This was posted before posts were moved from the announcement thread to this one. The post was made after post #6]

[wyldckat]

Greetings to all!

These two consecutive posts of mine are going to be long, because I tend to be long winded, but I'll try to keep it to the point and to use headlines to help you skip over what you might not want to read about (although I will unlikely repeat myself). In addition, I've pruned out all ranting I've originally written and placed it on a blog post (namely this one), to help keep these couple of posts more to the point.


On the topic of the "Moderator notes" above

As a moderator here on the forum and as an active member of the community that uses OpenFOAM technology, I've moved many of the posts above from the announcement thread onto this thread that we are on right now (I added notes to said posts, to help keep track of which came from where and when).
The reasons for the move are explained on that announcement in post #2, but I'll at least quote the last paragraph from here:

Originally Posted by wyldckat

Therefore, posts that were made between the 15th of January to the 2nd of February have been moved to http://www.cfd-online.com/Forums/ope...am-v3-0-a.html - those posts were on the topic of What's all this then, for which OpenCFD Limited did not answer within roughly 1 month here on this announcement thread. Therefore, the discussion has been moved to the other thread so that it can be better discussed by the community.

Where I stand in all of this (aka Disclaimer of intentions)

First of all, I do not work for, nor do I represent, OpenCFD Ltd, ESI Group and/or CFD Direct.
This post is written from my perspective as an active contributor in the community that uses OpenFOAM(R) technology (reminder: OPENFOAM is a trade mark of OpenCFD Ltd) and some of my views on this topic do match up to the company where I work at, namely blueCAPE Lda, but I did not run this by the company's public relations manager... nor did I have a lawyer check this, nor do I have professional legal expertises.

[rant pruned from here]

Long story short, 7 years and 1 month have passed since my first post, my current post count is at 9220 (namely this post) and my contributions to this community - which uses OpenFOAM technology - have essentially been part of my learning experience and expertise, because this is one of my main ways to study and learn about things. And I did this without any major or minor degree in CFD in University college, but instead I was taught to be a Mechanical Engineer in the field of Systems and Robotics.

My intention with this post is to bring a positive and detailed light in to all of this confusion that has come up (again) with the release of OpenFOAM+.
I will not dwell on whatever rights or wrongs made by the people who are part of OpenFOAM's evolution, given that we are all humans and we all make mistakes; specifically, I will not point out people's names in the description detailing OpenFOAM's evolution, I will instead only point out the main organizations behind each effort.
The objective here is to try and answer the questions brought in the posts above and to give a good perspective on what to do from here on forth.

Finally, if you want to, take a look at a couple of previous posts of mine on directly related topics to this theme, to give you a better notion of my perspectives on all of this:
* http://www.cfd-online.com/Forums/ope...tml#post543451 - post #2
* http://www.cfd-online.com/Forums/lou...tml#post572381 - post #2



The Really Short Explanation

• The really short explanation is very simple:
• OpenFOAM is to RHEL, as OpenFOAM+ is to Fedora.
• OpenFOAM+ seems to now have a development hub for the community.
• Which one to choose? Any one of them should do nicely, but you will have to fill bug reports for either one when you find bugs, if you want OpenFOAM/OpenFOAM+ to move forward!

The Short Explanation

Alexey has already made a good tl;dr description in post #10, but I want to give another short perspective on what to make of the latest OpenFOAM+ project:

1. Do a quick read about the Fedora Project: https://en.wikipedia.org/wiki/Fedora_Project - the tl;dr of it all:
   • Red Hat decided to split Red Hat Linux into Red Hat Enterprise Linux (RHEL) and a community-based operating system, Fedora.
   • This was on the 22nd of September of 2003, roughly 12.5 years ago.
   • RHEL provides major releases that are as stable as possible and supported along a 10 year period: https://en.wikipedia.org/wiki/Red_Ha...ersion_history
   • Fedora stays on the bleeding edge of open source technology and tries to still make that technology as stable as possible. It has a support period of 1 to 2 years: https://fedoraproject.org/wiki/End_of_life
2. How does it compare to OpenFOAM and OpenFOAM+?
   • The OpenFOAM project managed by the OpenFOAM Foundation is similar to RHEL.
   • The OpenFOAM+ project managed by OpenCFD Limited (ESI Group) is similar to Fedora.
3. How are they similar?
   • OpenFOAM Foundation is aiming to maintain versions that are meant to be of high quality standards, future-proof and easy to maintain: http://www.openfoam.org/contrib/
   • OpenCFD Limited has its own development cycle and provides new features, while keeping as close as possible to the same standards, as announced: http://www.cfd-online.com/Forums/ope...am-v3-0-a.html
4. As for development cycles, if we take the Fedora to RHEL cycles as reference examples:
   • RHEL 6.0 was released based on Fedora 11 to 13.
   • Fedora 11 was released on the 9th of June of 2009 and ended on the 25th of June of 2010.
   • Fedora 13 was released on the 25th of May of 2010 and ended on the 24th of June of 2011.
   • RHEL 5.0 was released on the 15th of March of 2007.
   • RHEL 6.0 was released on the 10th of November 2010.
   • Estimations over Fedora to RHEL versions are based on the GCC versions listed here: http://distrowatch.com/table.php?distribution=redhat vs http://distrowatch.com/table.php?distribution=fedora
   • Brewing period for RHEL 6: roughly 3.5 years, where the final features were matured in an 18 month period.
5. Therefore, we can perhaps expect something between 1.5 to 5 years for most features to mature and (hopefully) go from OpenFOAM+ to OpenFOAM.
6. Which version/fork/variant to choose?
   • Check the release notes of each one and compare with your needs.
   • Or simply try them all without spending a single dime in licenses (although it's somewhat time consuming)!
7. To which project should you contribute?
   • The GNU Public License (GPL) gives you the freedom to contribute to any or all of these projects.

Answers to questions

I'll try to answer in order of when they were asked:

• @SBusch in post #1:
  • OpenFOAM+ essentially stands for "OpenFOAM Foundation developments plus OpenCFD Ltd developments". The reasoning for this is mostly due to the dimensions of the development cycles done within OpenCFD vs the ones contributed to the OpenFOAM Foundation.
  • More details in The (really) Long Explanation section below and in The Short Explanation section above.
• @kmooney in post #3:
  • As Alexey explained in post #10, Copyright lines are added where applicable. This is explained in the GPL FAQ and exemplified in the OpenFOAM Foundation website: http://openfoam.org/contrib/unsupported-examples.php
  • The README.* files refer to the Copyright for the current code organization, i.e. a sign of the Copyright of project's efforts, given that actual disclaimer is given per source code file: http://www.gnu.org/licenses/gpl-howto.html
  • For other details, check The (really) Long Explanation section below.
• @Zeppo in post #5:
  • More freedom, more choices, more features! And a new development hub dedicated to the advancements of OpenFOAM technology, even if it doesn't look like it yet.
  • • The hub: https://develop.openfoam.com
    • The reasoning: http://www.openfoam.com/services/community-projects.php
• @Sylvain in post #7:
  • Evolution is not always clear at first sight. Only by letting time pass by or by looking at historic events, can we deduce how a certain evolution can occur.
    To answer directly to your questions:
  • • Q: What does that mean?
      A: OpenFOAM technology and associated ecosystem is evolving.
    • Q: ... communication is not very clear ... web sites are similar ... the logos on linkedIn and twitter ... What does that mean?
      A: Again, a sign of the evolution path. In 2011, the two websites almost only differed in the colour and they had bilateral links to each other, back when the OpenFOAM Foundation was founded. This means that they represent the same exact technology, it's just that currently the openfoam.com domain is associated to the OpenFOAM+ project and the openfoam.org is associated to the OpenFOAM project.
    • Q: Are they in good terms?
      A: As Alexey succinctly explained in post #10, this is not an easy answer to give in a public forum, since it's similar to what the FBI/CIA can tell us about certain things. So please read The (really) Long Explanation section below for an interpretation on what it means to the community.
    • Q: What can we expect in the long term? Can we rely on OpenFOAM for a long term development of our simulation process?
      A: Yes, definitely yes! It might not look like it yet, but please read The (really) Long Explanation section below for a good perspective on this.
• @egp in post #8:
  • No question asked, no answer needed . Many thanks for the deal.ii link! [another rant pruned from here]
• @michaelaye in post #9: To answer specifically to your questions:
  • Q: Is OpenFOAM dead?
    A: Open source licenses allows for source code software to never die. And it's actually evolving, so it's very much alive! [yet another rant pruned from here]
  • Q: What happens to the "Open" of OpenFoam if things go commercial? - A: Open source code means that the product is open and has total freedom, but the main support service behind it might be a paid service. And OpenFOAM always had a commercial side to it, namely the training and the service. The source code will always be free, under the GNU Public License, as explained at both websites!
  • Q: Is there maybe a thread where this is explained somewhere?
    A: Each relevant website (openfoam.com, openfoam.org and cfd.direct) already gives partial explanations, although they are not crystal clear on some of their relations. This discussion thread we are currently on will be an attempt at making everything clear, at least from the community's perspective.
• @LVDH in post #11:
  • Many thanks for your comments! I didn't manage to have the time to attend the OF conference in Stuttgart last year and I've been busy with several projects, which is also why this is my delayed response on this topic.
    Anyway, to answer your questions:
  • • Q: So how did they come up with adding these few new features into a new fork?
      A: In essence: from a developer perspective, this is the most logical way for ensuring the future of OpenFOAM technology as we know it! As I mentioned above in the section The Short Explanation, this has to do with the software development cycles and with the objectives and features desired in OpenFOAM. This is as a natural evolution as it was/is with RHEL and Fedora! Nonetheless, more details are listed in the section below The (really) Long Explanation.
    • Q: Could somebody please explain?
      A: Be careful what you wish for, you might not have the patience to read the whole explanation! (hopefully I'm able to explain most of it)
• @pbachant in post #12:
  • At least most of the projects are using git, making it a bit easier to migrate code if and when necessary.
  • As for a feature from OpenFOAM+ not getting merged into the Foundation's development line: There is always more than one solution, some better than others. And the open source license ensures that we have the ability to implement other solutions, or to simply help integrate the existing ones

continues in the next post...

[wyldckat]

The (really) Long Explanation

Sorry, but please prepare yourselves for a long (and potentially boring) reading, if you really want to understand it all (or most at least of it).
In order to answer the question of What's all this, then?, I'm going to have to first outline some details.

[rant pruned from here]

One of the other reasons is somewhat ironic, because most of the issues we are observing with OpenFOAM technology today have analogies in CFD modelling strategies and terminologies. Just to name a few:

• MPI and domain decomposition - the teams/contributors communication layers, which needs some more work.
• Convergence of the equations towards the desired solution - the development of the technology itself has a lot of variables, for which forking has been one of the solutions (i.e. segregated flow).
• Organization of the software development strategies, which is akin to which turbulence models to choose, mesh dependence and so on...

Therefore, we'll have to breakdown down the complexity of it all and start with the basics:

• Using OpenFOAM usually requires from the users the utmost attention to details.
• Every single detail, no matter how small, can have a bigger impact than we would like it to have. This can either be a missing semi-colon, a misplaced decimal separator or a missing/extraneous bracket.
• There are ways to solve these miscellaneous details, but the main fact will still and always remain the same: details are important, specially because we are trying to model reality in a virtual world!

Beginning with the name itself: "OpenFOAM" is a trade marked name that is shared among several components that make up what we simply refer to as "OpenFOAM". It's shared among the following components (somewhat in alphabetical order):

1. The OpenFOAM bug/issue tracker
2. The OpenFOAM copyright
3. The OpenFOAM development team
4. The OpenFOAM documentation
5. The OpenFOAM Foundation
6. The OpenFOAM project
7. The OpenFOAM software
8. The OpenFOAM software repositories
9. The OpenFOAM technology
10. The OpenFOAM trade mark
11. The OpenFOAM website

How are all of them related? It's very simple (sorry, I won't draw a diagram):

• The OpenFOAM trade mark is the owned by OpenCFD Ltd, which has granted the OpenFOAM Foundation to use the trade mark.
  • Others may use the trade mark, but only if they properly disclaim that they do not own said trade mark, nor are they officially developing the OpenFOAM software. This is done by following the trade-mark policy and using the disclaimer detailed therein: http://openfoam.com/legal/trademark-policy.php
• The OpenFOAM project is mostly for organizing the platform in which the following components are managed:
  • The OpenFOAM bug/issue tracker, where the bug reports and feature requests are handled: http://www.openfoam.org/bugs/
  • The OpenFOAM documentation, which explains on how to use the OpenFOAM software and how its source code is organized: http://www.openfoam.org/docs/
  • The OpenFOAM software, which is the main product we refer to when we talk about OpenFOAM.
  • The OpenFOAM software repositories, which are essentially providing an organized way to access the development history of the source code. Technically there are two repositories:
    • The git repositories at GitHub: https://github.com/OpenFOAM/
    • The project at SourceForge.net that has the source code and binary snapshots of the software: https://sourceforge.net/projects/foam/
  • The OpenFOAM website - http://openfoam.org - which is essentially an access portal that gives us access to all of the developments (source code, documentation et al) that come from the OpenFOAM project.
• The OpenFOAM Foundation is the organization responsible for the OpenFOAM project.
• The OpenFOAM development team, which works on the OpenFOAM project, where each element of said team handles parts of each component...
  • Well... actually, the terminology for this changed a bit when the development division (OpenFOAM and OpenFOAM+) began to occur.
  • Up to sometime in 2014 (there is no clear specific public date), OpenCFD had the only official OpenFOAM development team. [rant pruned from here]
  • The current team organization working for the OpenFOAM Foundation technically isn't composed by a single team, it's composed by participants (and small teams) that actively contribute to the OpenFOAM project.
• The "OpenFOAM copyright" on its own is a misnomer. Technically there are several bodies of work that have copyrights assigned to the OpenFOAM Foundation:
  • the copyright over the OpenFOAM software;
  • the copyright over the OpenFOAM documentation;
  • the copyright over the website at openfoam.org;
  • these include the subsequent versions of each body of work to which that copyright is applicable.
  • (And I'm probably missing some other bodies of work that has copyright by the Foundation.)
• The OpenFOAM technology: this can almost be referred to as it being the soul of the OpenFOAM software itself. This is the term used whenever a fork or variant is created and whenever someone uses OpenFOAM software in their own projects.

How does this translate to OpenFOAM+? We have new additions to the equation:

1. The OpenFOAM+ bug/issue tracker
2. The OpenFOAM+ copyright
3. The OpenFOAM+ development hub
4. The OpenFOAM+ development team
5. The OpenFOAM+ documentation
6. The OpenFOAM+ project
7. The OpenFOAM+ software
8. The OpenFOAM+ software repositories
9. The OpenFOAM+ website

In other words:

• The OpenFOAM trade mark is still the same, since the "+" suffix is mostly just a way to tell apart the objectives and development efforts of each project.
• The OpenFOAM+ project is the mostly for organizing the platform in which the following components are managed:
  • The OpenFOAM+ bug/issue tracker, where the bug reports and feature requests are handled, which is within the development hub;
  • The OpenFOAM+ development hub, where the main projects for OpenFOAM+ software, documentation and other related community development projects can potentially be organized in: https://develop.openfoam.com
  • The OpenFOAM+ documentation, which explains on how to use the OpenFOAM+ software and how its source code is organized.
  • The OpenFOAM+ software, which is the main product we refer to when we say OpenFOAM+; this is also the primary body of work that is built on top of the OpenFOAM software.
  • The OpenFOAM+ software repositories, which are essentially an organized way to provide access to the development history of the source code. Technically there are two:
    • The git repositories at the development hub: https://develop.openfoam.com
    • The project at SourceForge.net that has the source code and binary snapshots of the software: https://sourceforge.net/projects/openfoamplus/
  • The OpenFOAM+ website http://openfoam.com - technically it's the website for the OpenCFD Ltd company (part of the ESI Group), but it is essentially a portal that gives us both access to all of the developments provided by the OpenFOAM+ project and the services provided by OpenCFD Ltd.
• The OpenFOAM+ development team, which technically is the development team at OpenCFD that takes care of the OpenFOAM+ project, as well as the services that they provide. In addition, they can refer themselves as one of the official OpenFOAM developers, given that OpenCFD holds the trade mark.
• The "OpenFOAM+ copyright", which once again is a misnomer, refers to the bodies of work that have copyrights still assigned to OpenCFD Ltd (ESI Group), as well as other contributions that have not yet have their copyrights transferred to OpenCFD (see this page http://www.openfoam.com/services/community-projects.php for more details), where there is:
  • the copyright over the OpenFOAM+ software;
  • the copyright over the OpenFOAM+ documentation;
  • the copyright over the website at http://openfoam.com;
  • these include the subsequent versions of each body of work to which that copyright is still applicable.
• And OpenFOAM+ is built directly on OpenFOAM technology!

But wait, there's still more to this! Here is a summary time line of the development of OpenFOAM (and FOAM) over the years:

1. FOAM was first conceived/created in 1989.
2. FOAM was officially distributed by Nabla Ltd between 2000 to 2004.
3. The OpenFOAM project was released on the 10th of December of 2004, by OpenCFD Ltd.
4. The OpenFOAM Foundation was created and took over the copyright(s) and the OpenFOAM project in August 2011.
5. On the 22nd of May of 2015, the OpenFOAM Foundation released OpenFOAM 2.4.0, in the project's first explicitly multi-participant format.
6. On the 22nd of September of 2015, OpenCFD released publicly the OpenFOAM-history repository: https://github.com/OpenCFD/OpenFOAM-history - which contains the complete development history in git form since the creation of said git repository on the 15th of April of 2008: https://github.com/OpenCFD/OpenFOAM-...fbc6fe272f0776
   • Along with that, as stated in the http://www.openfoam.com/download/openfoam-history.php page, the Copyright for this OpenFOAM-history development line was fully transferred to the OpenFOAM Foundation on that month as well! [important rant pruned from here]
7. On the 14th of October 2015, the first glimpse of the OpenFOAM+ project was made public, with the inception of OpenCFD's officially supported "OpenFOAM for Windows" - which according to their historical records, looks like it was a community project to start with.
8. On the 13th of January 2016, OpenFOAM+ was officially released by OpenCFD Ltd, based on the 3.0 development line.

Public online references for these dates in particular:

• http://cfd.direct/openfoam/about/
• http://web.archive.org/web/200412171.../mainhe12.html
• http://openfoam.org/download/version1.0.php
• http://www.sgi.com/company_info/news...t/opencfd.html
• http://openfoam.org/version2.4.0/
• http://www.openfoam.com/download/openfoam-history.php
• http://www.openfoam.com/news/
• http://www.openfoam.com/services/community-projects.php

So, what's the big deal with all of these dates? In summary:

1. FOAM had a lifetime of 14 years, where at the age of 10 it became part of a commercial services company.
2. OpenFOAM at the age of around 7 years old, became part of a Foundation; at the age of 10-11 became something that was in principle developed by a broader audience of participants on the Foundation.
3. OpenFOAM at the age of 11 was then also branched into OpenFOAM+.
4. If the original FOAM project was still operational (it is, but in spirit through the foam-extend project), it would now be 26 or 27 years old.

Still not seeing what I'm seeing? Here's my point of view on this:

1. Since OpenFOAM is developed by humans, it makes some sense that it has a life time on the human scale.
2. At the age of 10-12, it's when humans begin their adolescence.
3. At the age of 18-21, some humans go to a University.
4. At the age of 23-28, humans who went to a University are now fully grown adults, although still in development, maturing and honing their skills.

Therefore, my perspective on this is that:

1. FOAM/OpenFOAM share a common trait: when they started their adolescence age, they have gone through big and very similar changes.
2. OpenFOAM has a duality to it, where one part is now grown up into a fully fledged adult, as an inherited trait from FOAM's development line (i.e. guided by the original architect of FOAM).
3. OpenFOAM+ is the adolescent counter-part of OpenFOAM, given that OpenFOAM as a project is still ~11 years old.
4. If this duality wasn't split in two now, then OpenFOAM (in human form) would continue to behave as an adolescent for another 6-10 years.

This gives even more weight to the perspective I'm trying to present to you all, namely where the appearance of OpenFOAM+ is part of OpenFOAM's natural evolution and that this is actually a good thing! It's akin to what happened when RHEL and Fedora were conceived and how they are now (although the time lines probably don't fit this well)!


So, finally, getting down to what really matters to us: The community that uses OpenFOAM technology!
We are now exactly at a vital evolutionary point in the life of OpenFOAM, where we can make it evolve in the direction we truly wish it to have! Just look at what we already have:

1. The OpenFOAM project: which is provided by the OpenFOAM Foundation and is aiming to gives us the most matured, advanced and future-proof OpenFOAM (CFD) technology known to mankind.
2. The OpenFOAM+ project: which is provided by OpenCFD Ltd (ESI Group), is aiming to give us both their bleeding edge developments that they can provide in a stable format (it's not their internal development line ), as well as a development hub for community contributions! And OpenFOAM+ already working on any Docker-ready platform!
   • The only downsides that I currently see are the following:
     • Since it's bleeding edge technology, it still needs some more time to mature and eventually graduate parts of its new features onto the main OpenFOAM project.
     • The development hub does not have an admission fee, but the contents are not publicly visible by default. On the other hand, this is somewhat expectable for security reasons, given that if something could go wrongm then this way things stay contained and OpenCFD/ESI can more easily handle any problems.
3. The foam-extend project: which is provided by Wikki Ltd and the community that uses OpenFOAM technology in general, gives us both features that were dropped from OpenFOAM's development line over the years, as well as giving us other new features and bug fixes. And it works natively on Linux, Mac OS X and Windows!
4. The Caelus-CML project: which is provided by Applied CCM, that gives us a more publicly validated+documented "OpenFOAM technology" development line, which also works natively on Linux, Mac OS X and Windows!
5. The FreeFOAM project: which currently is in a strange pseudo-zombie-state project, but at least it provides an automatically and always-ready way to have OpenFOAM technology on Debian and Debian-based Linux distributions.

There are more forks and variants, but these are the main projects that I believe that can gives us the primal resources for everything we want from OpenFOAM technology and much more! Why? It's very simple:

1. OpenFOAM+ can help maintain community developments, as well as give an integration backbone for features to mature and graduate into OpenFOAM.
2. foam-extend's past and ongoing developments can now more easily find a way into OpenFOAM+ to begin bringing back features that were dropped from OpenFOAM over the years.
   • I'm not saying that all features will be brought into OpenFOAM+, but I am thinking that this enables us to start with the more desirable bleeding-edge features, while also maintaining foam-extend's developments over possibly a 5 to 10 year period or more. (Is anyone starting to see this in the residual plots for possible convergence?)
   • And if by any chance things don't move towards OpenFOAM+, at least it has a development community as well!
3. Caelus-CML can always keep us safe, just in case something terribly bad happens to the other communities (evolution always has its ups, downs and ends ); and its development line can also bring in much needed public validation work for OpenFOAM technology!
4. FreeFOAM gives us a heritage that perhaps will help future developments! The technology and knowledge within it can always come back to help us if we're ever in a time of need!

But (there is always a but), there is one seriously important detail that you must never ever forget from all of this potential:

In order for open source software to evolve and thrive, this means that a community is always needed. Money is also needed for most projects to thrive, but an open source project without a community is as good as dead.

What I meant by this is that you should not simply stand on the side lines and only remain as an outside onlooker. You should try and find a way in which you can contribute to one or more of these projects, so that they can evolve and eventually converge into what they can truly become!
Otherwise... fear will win the day and we might end up loosing our freedoms that we currently take for granted.

[final rant pruned from here]

If I didn't fully answer your questions with this, please rephrase/remind me of your questions, or ask new questions as well, because I probably lost my train of thought somewhere in all of this text...

Best regards,
Bruno

[roenby]

Dear wyldckat

Thank you very much for sharing all this very detailed information and your perspective on the forking of OpenFOAM. It is nice to hear that you are very optimistic and see the splitting as an advantage.

Nevertheless, even after reading your posts, I am still left with a feeling that something is fundamentally not right about the development of OpenFOAM. To be honest, I am not sure, that I am buying your comparison of the relation between OpenFOAM and OpenFOAM+ to the relation between Red Hat Enterprise Linux and Fedora: As far as I can see (I may be very wrong since I am definitely out of my field of expertise here), “Red Hat Enterprise Linux branches its releases from versions of Fedora”[1]. Development is going on in Fedora, which is bleeding edge, and Red Hat offers more stable releases based on Fedora. This is what guarantees that the two projects are not diverging over time. In your analogy, this would correspond to all the devlopment going on in OpenFOAM+ and then OpenFOAM releasing from versions of OpenFOAM+. However, it seems that separate development will go on in OpenFOAM and OpenFAOM+. It is not at all clear to me that stuff from OpenFOAM+ will be merged into OpenFOAM. In fact, it is not clear to me at all, that there is good (or any) communication and collaboration between OpenFOAM Foundation and OpenCFD/ESI-Group*. If this forking is (as I sense) another unhappy divorce, then for sure OpenFOAM and OpenFOAM+ will eventually diverge just as it happened with OpenFOAM and foam-extend. I would like to stress here, that I have no solid information to base my gut feeling on. But if this is actually another diverging forking of the OpenFOAM project, then I believe that this may be devastating to the project. In fact, the sole uncertainty about what is going on behind the scenes is harming the project, since it will make people and companies look for other more future-proof (appearing) alternatives to base their CFD work and development on.

Please tell me I am wrong about all this...

[1] https://en.wikipedia.org/wiki/Fedora...ystem)#History

*At the bottom of the "Contributors" page on openfoam.org it says: "Note that there are no contributors from ESI-OpenCFD (openfoam.com)". Mattijs Janssens is listed as a contributor, but as an "(individual, UK)", where as on www.openfoam.com/version-v3.0+/ his association is listed as OpenCFD Ltd

[gucong]

When you look at OpenFOAM-dev, you will find that they are actively doing API changes, renaming stuff, adding new features, etc. Remember these will endup in each new release of OpenFOAM foundation versions. It is nowhere near RHEL in the analogy in terms of stability. And I don't think such moving target serve as an appropriate base version for outside application developers either because of that.

[wyldckat]

Greetings to all,

@roenby: First of all, many thanks for the feedback and comments, given that I was sort-of accused of killing of this thread, with the two posts that I had made, when the idea was to discuss more on this.

And sorry for the late reply, but I'm still catching up on the massive backlog I have of PMs and threads here on the forum, due to having spent a few months' worth of free time helping out with the OFW11 community workshop this year. Given that gucong has revived this thread one again and I have a bit of time today (holiday here in Portugal), I might as well pull this up on my priority list.

And sorry again, because I wasn't able to be succinct on my writings below...

To answer your comments/questions:

Originally Posted by roenby

Thank you very much for sharing all this very detailed information and your perspective on the forking of OpenFOAM. It is nice to hear that you are very optimistic and see the splitting as an advantage.

You're welcome and yes, I do see this as a positive sign. Changing things for the better sometimes can only be done with a big change. Of course the problem is managing to make it go to the right direction...

Originally Posted by roenby

Nevertheless, even after reading your posts, I am still left with a feeling that something is fundamentally not right about the development of OpenFOAM.

I expect that many who have developed directly in FOAM/OpenFOAM would say that this has always been the case, even back when only FOAM existed It's software development. Doing it right is not an easy feat. Ask Linus Torvalds about it

Originally Posted by roenby

To be honest, I am not sure, that I am buying your comparison of the relation between OpenFOAM and OpenFOAM+ to the relation between Red Hat Enterprise Linux and Fedora: As far as I can see (I may be very wrong since I am definitely out of my field of expertise here), “Red Hat Enterprise Linux branches its releases from versions of Fedora”[1]. Development is going on in Fedora, which is bleeding edge, and Red Hat offers more stable releases based on Fedora. This is what guarantees that the two projects are not diverging over time. In your analogy, this would correspond to all the devlopment going on in OpenFOAM+ and then OpenFOAM releasing from versions of OpenFOAM+. However, it seems that separate development will go on in OpenFOAM and OpenFOAM+.

Unfortunately I'm not aware of popular open-source'ish projects that have a development format similar enough to what is being done with the OpenFOAM and OpenFOAM+ projects.

Using RHEL and Fedora as comparison is mostly to give people a bit more perception of what can be done and achieved with the current development model, which I expect that it will be done with the two projects, namely OpenFOAM and OpenFOAM+.
Nonetheless, I did not intend to give the notion that these two projects have already attained an identical status to what is done in RHEL+Fedora. I should have been clearer about that.

Debian and Ubuntu was a possibility, but would have probably made it even more confusing.

Furthermore, I can't find any documents on-line on how smooth/rough the initial RHEL+Fedora venture was back in 2003-2004, but I at least suspect that it wasn't very smooth during the first few years, even though Red Hat was/is backing the project directly. What I can say with some certainty is that using Fedora for production environments can be a hit and miss situation: it works great for some, it's terrible for others. Fortunately OpenFOAM+ shouldn't be as ambiguous, although the "stable" releases have had a few installation issues... but it's been less than a year since the initial release with v3.0+, so these kinds of bumps on the road are expected.

Originally Posted by roenby

It is not at all clear to me that stuff from OpenFOAM+ will be merged into OpenFOAM.

There have already been several pieces of code brought in from OpenFOAM+ to OpenFOAM:

• A few pieces were brought in by me a few months ago, as pointed out in the important rant that I had moved out from my previous post, sorry about that (it was pointed out to me that it shouldn't have been moved out as a rant). I've updated the tag on the blog post to make it clearer (look for "[IMPORTANT_RANT]"): Rants removed from a recent post of mine about OpenFOAM vs OpenFOAM+ in 2016-02-14
• A lot of code has been brought in by Mattijs Janssens, with a few additional tweaks and fixes. The biggest commit I can remember associated to this effort was this one: https://github.com/OpenFOAM/OpenFOAM...e86b8c45dc2f80
  You can find more with this command, from within the OpenFOAM-dev cloned repository on your terminal:
  

  git log --all --grep="Mattijs Janssens"

Originally Posted by roenby

In fact, the sole uncertainty about what is going on behind the scenes is harming the project, since it will make people and companies look for other more future-proof (appearing) alternatives to base their CFD work and development on.

Please tell me I am wrong about all this...

Well, this is one of those situations where it really depends on how invested are the current people and companies who work with OpenFOAM. Writing about it on the forum(s) isn't enough. If they are truly invested into the OpenFOAM technology - and if they can - then they should be actively seeking out contact with CFD Direct and ESI-OpenCFD, to find out with who they should work with directly, in order to help both projects move forward and getting things done. There have been already several companies who have done so and that is why we can see them listed at each website, namely openfoam.org and openfoam.com

As for other people and companies who are still browsing the market: well, Red Hat's history in the very early 2000's had several bumps in the road and now they are thriving.

ESI-OpenCFD has shown signs of heavily investing in OpenFOAM, not only by getting closer to the community (see the threads on the installation forum with the prefix "[OpenFOAM plus]", the community project list still growing and the contributions on the OpenFOAM Foundation bug tracker), as well as open to discussion on how we can get back to a common ground for developing OpenFOAM Technology, as reported here: http://www.esi-group.com/company/eve...onference-2016

We are adding a special session to encourage views and discussion on "One OpenFOAM".

Almost forgot to comment on this:

Originally Posted by roenby

then for sure OpenFOAM and OpenFOAM+ will eventually diverge just as it happened with OpenFOAM and foam-extend.

For better or for worse, this is still possible to occur. AFAIK, the objective of those who have been contributing to the 2-3 projects (OpenFOAM, OpenFOAM+ and foam-extend), is to avoid this from happening, but it's still a possibility.
One solution would be to have OpenFOAM+ split into two development repositories:

• One that only does the necessary API changes and fixes to the OpenFOAM-dev indexed development.
• Another where the new features are maintained.

Releases can then be done with the two repositories in a single package. If this had been done with foam-extend from the start, we probably wouldn't even be writing about this.

But the problem with this is very simple: it brings in an additional layer of development complexity and more work to the table. This can only be done if there are enough people contributing. Which is pretty much why this wasn't feasible with foam-extend back when it started, at least as far as I can figure it out.



@gucong: Also, many thanks for your comments:

Originally Posted by gucong

When you look at OpenFOAM-dev, you will find that they are actively doing API changes, renaming stuff, adding new features, etc. Remember these will endup in each new release of OpenFOAM foundation versions. It is nowhere near RHEL in the analogy in terms of stability. And I don't think such moving target serve as an appropriate base version for outside application developers either because of that.

The OpenFOAM-dev development repository was started near the end of 2014. It's been almost 2 years since the efforts to consolidate and to harden OpenFOAM's core code into a consistent development line, on which everyone can develop on top of.
A few major examples:

• The API changes were critical for the future health of OpenFOAM's core code, e.g. to have a protected const access to the "tmp" objects.
• When utilities were moved to become function objects, was to consolidate code, add features and minimize maintenance requirements in the future. The genesis for this was this bug report: http://bugs.openfoam.org/view.php?id=1301
• Turbulence models are now manageable within the same infrastructure, instead of having 3 separate dictionary structures: "RASProperties", "turbulenceProperties" and "LESProperties" - these three are now consolidated within a single "turbulenceProperties".

If the code was being handled since the end of 2014 the exact same way as RHEL... well, several bugs on the bug tracker could not be fixed for another 3-5 years, because the API couldn't be changed

And if this work hadn't been started in the first place, you would continue to see these kinds of changes for another 10 years or more. Hopefully, these kinds of massive changes to the core API will stop with a few more month's work, when all of the code has been consolidated.


OK, enough from me for now I gotta get back to the backlog ... if anyone reading this would be kind enough to contribute to the forum, wiki, bug trackers and so on, everyone will thank you for it, specially those who you've helped

Best regards,
Bruno

[gucong]

I still consider such constant/rolling API changes in such scales for a library project with 10+ years of history is insane. I understand some API changes are essential for certain bug fixes and long term health of the project. But what about masive renaming with only marginal cognitive benefits? For example, they recently renamed very fundamental things like internalField() to primitiveField(). Actually, it seems to me that the developers are totally not concerned with API stability and, in turn, the lifes of community application and enhancement developers. It is almost like they are intentionally breaking things others built around OpenFOAM. I am sure the OpenFoam+ project will suffer too when they are maintaining a separate but compatible development line, but probably to a less extent than others considering the resources they have in OpenCFD.

When you are talking about bug persisting for 3-5 years in RHEL, remember they do it for a reason. Namely, API stability is considered more important in those cases. At least the bugs are known and documented, while a slight API change could cause unknown bugs in other parts of the system. That's partly why RHEL is preferred in any serious or critical situations. In my opinion, API stability shows the seriousness of the library developers and respect to users of the library. At least don't do that in a rolling release model, keep it stable within major versions that stay maintained and developed spanning several years. In contrast, versions now are just tags in OpenFOAM-dev, which is the only actively developed line, if I understand it correctly.

[alberto]

Originally Posted by gucong

When you look at OpenFOAM-dev, you will find that they are actively doing API changes, renaming stuff, adding new features, etc. Remember these will endup in each new release of OpenFOAM foundation versions. It is nowhere near RHEL in the analogy in terms of stability. And I don't think such moving target serve as an appropriate base version for outside application developers either because of that.

My understanding is that a lot of the recent API changes were motivated by the need of differentiating access functions (ref/non-ref) to support some architecture. Others were due to cleanup and rationalization.

I don't like to change my code to keep up with API changes, but overall it isn't terrible. I actually found it quite simple to follow -dev directly, rather than freeze the development branch of my code on a release and wait until it is stabilized.

Definitely I do not agree with the RHEL approach applied to a CFD code: if you keep bugs in a CFD code to ensure API compatibility, you obtain incorrect results.

[alberto]

Originally Posted by wyldckat

A few pieces were brought in by me a few months ago, as pointed out in the important rant that I had moved out from my previous post

Did you have to do anything in terms of "permissions" to do this, or did you just port the code? I was thinking of porting some of the extra BCs, but I don't know how that would work without a completely independent re-implementation in terms of copyright transfer.