A sad day for Open Source: Camunda decides to fork Activiti

I wasn’t going to give attention to this

In my opinion, bad publicity is also free publicity. And that’s not deserved.

I really don’t care that people use Activiti, tweak and patch it and not share it. Many companies do that. But when they publicly shout out that they fork Activiti to create an own engine, they hurt the Activiti community. They hurt the community because the confusion they cause. And when you touch the Activiti community, you touch me. And so I decided to publicly talk about it.

Open Source is built on trust

The whole open source ecosystem is based on trust. It is based on people discussing problems and find solutions together, and working for a common cause to make the world better. Many companies work on open source and build own products on top of open source software and contributing certain pieces back. That’s how open source works.

I know the people in Camunda for at least 5 years. Even before we worked together on jBPM I discussed software and bpm with them in the pub. I really enjoy(ed) working with the people from Camunda on Activiti, especially Daniel whom I connected with on a technical level very few people can. So imagine my surprise when Daniel called us to let us know that they decided to fork Activiti and start their own bpm platform. Open source and apache licensed. No discussion up front. We didn’t have any arguments. They basically claim our visions don’t match anymore. In open source, or in any collaboration between people you first bring issues you have to the table. You discuss the problems. You try to find a solution. And when that fails … maybe then you can decide to pull the plug and fork.

But nothing of that happened. Just the message out of the blue that the fork was going to happen. Hell, we organized the Activiti Community Day of past November together with them. Just a few weeks ago I proposed a get-together meetup to rocket-strap the roadmap of Activiti. Was all of this already decided then?

I wouldn’t have reacted if they just announced their own platform. But a fork of Activiti is really damaging for our community. Our CTO and their CEO discussed this via email and we got the following message from him (copied verbatim):

That’s said be assured that we do not place a message like “we fork Activiti”

Let me just post a screenshot of the title of the article of them. Enough said.

Screen Shot 2013-03-18 at 13.26.24

I really trusted them. I trusted that they would do the honorable. I really did.

So I’m not only let down on the community-aspect of things, but also on a personal level.

Is there something wrong with Activiti?

Open source is build on trust. Forking an open source project is introducing distrust. It confuses people. It makes them wonder ‘why would these guys fork’ or ‘is there something wrong with Activiti that we don’t know?’. I don’t know the answer to the why-question. I don’t understand it and I don’t know why. Maybe I would have if we would have discussed things.

There is nothing wrong with Activiti. The last couple of months were our most productive ever. We released version 5.12 and it was *packed* with features and fixes.

There is nothing wrong with Activiti. Hell, the ‘migration guide’ from Activiti to Camunda is basically just a jar rename. Some docs are copied exactly from the Activiti docs. Where is the difference in vision between the engines? Why not work together and contribute the missing pieces to Activiti? I think you yourself can be the judge on this.

Was it about branding? Do they want their own sticker on the bits and pieces? While Alfresco is putting a lot of money in Activiti, I don’t believe Alfresco has ever tried to mark the project. There isn’t a mention of Alfresco on the Activiti home page (and we’re proud on that Alfresco treats Activiti as such!). Again, all these things could have been discussed and solved. But none of that happened.

The future

But things are what they are. Camunda have forked Activiti and that’s how it is. It is sad on different levels, but let’s focus on the future now. We will build your trust. We will prove that this fork has no impact on Activiti.

 

I’ve been chatting with a lot of people of the community in the past days and they all share our disbelief about the fork. Activiti has never been more alive than today.

At Alfresco we have a new CEO since a few months and workflow/bpm is in the list of top priorities of our management. Together with our many other contributors we will keep pushing Activiti, and maintain it’s status as the number one fully open-source platform for BPM and workflow.

I’m sorry for the confusion being caused. I’d like to see it different, too.

Believe me on that Activiti will keep to it’s core open source values and we will continue to strive to make Activiti even better than it is. Each and every single day. You can trust me on that.

Joram

PS: feel free to leave any happy Activiti comments to this post. It will cheer me up. I need that now :-).

32 Comments

  1. Per Buer March 18, 2013

    Keep calm on keep on coding.
    One of the reasons I like OSS is that it forces you to keep on creating great things. If you don’t someone will fork you and then run all over you with that fork.
    You’ve been challenged on your commitment to the project. If they create a better product you’ll be screwed with a “death by a thousand needles” (look at MySQL). If, at the end of the day, your product is better you will live on, and maybe you’ll learn a thing or two from them. You can maybe even grab a line or two of code from them.
    Now, go run the fork into the ground by putting out a great product!

  2. Joram Barrez March 18, 2013

    @Per Buer: Exactly! I agree with you completely. I do feel enormously challenged now to prove Activiti. And we won’t disappoint.

    I just feel sad for the way it went and the duplicated/lost efforts. But I’m sure Activiti will get stronger out of this, because of the things you mention. Thanks for posting them.

  3. Alexey Vasyukov March 18, 2013

    Hey. Just keep doing great things. You are absolutely right about trust – it is the only thing that really matters in open source community.

    I (indirectly) worked for Red Hat when RHEL was forked by Oracle. After that I saw a lot of ugly attempts by SIs to ‘rebrand’ open source products and sell as their solutions. We encountered recently two attempts to fork Alvex. I can understand what you feel when you see sudden fork without even prior discussion.

    But know what? All forks (if they were really forks) failed. Nobody needs secondary products. My personal experience shows that any fork even makes original product stronger (if the product survives the fork). After RHEL was forked, heavy enterprise customers started thinking like ‘if Oracle believes in this OS that deep, it makes sense’. After Alvex was forked recently, top SIs stopped talking to our competitors because ‘it looks like these guys have really the best solution everybody wants to own’.

    So, just keep doing great things. Survive the fork and become even stronger. May the force be with you. 🙂

    P.S. I have also some additional concerns specific for ‘Camunda Forx’ viability, but I’m not ready to post them now, I need to double-think before making any forecasts.

  4. GPL anybody? March 19, 2013

    Hello Joram,
    choosing the apache license means that you have no protection from companies taking the code and gaining from it without giving back. If you can’t leave with that, just switch to the (L)GPL. It is designed for such protection.

    Many people bash the GPL but I think that’s some lobby of people that don’t want to give back as much as they take. I know it’s hard wrt politics but really they do it, because you allow them. If everybody was conscious and with good intentions there wouldn’t be a need for any licenses. Not that any license can entirely solve the problem but can be part of the solution.

  5. Joram Barrez March 19, 2013

    @Alexey: thanks for the kind words. We’ll indeed survive any fork and become stronger than ever!

    @GPL anybody: Yes, I see your point. But then again, apache does make it easier to have companies contribute, much easier vs GPL. There something to be said about both. I’m not caring about not giving back. But I do care about public forks. But I see your point. Thanks for that!

  6. Jeff Steel March 19, 2013

    Hello Joram,

    like I got Camunda to know is that they do less tings upfront but many things retroactive. They really have a few good developers but also others. I do not think that success will be to have the best engine means on a technical base.
    Though I think this would be the best measurement in reality it won’t.
    Many bad software has been established against better ones because it matched some customer needs better or it has just been promoted a bit better. And this is one strength Camunda really has. It promotes very well. If you will ever have the chance to look beyond the scenes or beyond the facade you would maybe decide different. But who is able to do this or has time for this in a volatile world?

    So yes – please focus on achieving the best BPMN engine but don’t forget about promoting this (because others are already there).

  7. javapapo March 19, 2013

    A couple of years ago, I decided to strongly propose on several tenders, Activiti as the new ‘breed’ of BPMN framework to be used – in the tenders the company I work for, was submitting. It recently made it in some contracts and is currently -being used!

    We have used with great success for more than 3 years, jBPM3, which was the child of the same team that gave birth to Activiti. I am sure about my proposal several years ago, I am sure now as well.

    Please keep pushing Activiti and stick to the plan, as with jBPM and now we the Activiti offer, we (the users-developers’) have faith on the actual technology and at the same time to the people behind it :).

    Greetings from Greece 🙂

  8. Joram Barrez March 19, 2013

    Thanks Paris, much appreciated!

  9. Phillip Rhodes March 19, 2013

    Will the new fork remain Open Source? If so, I wouldn’t see it as being *such* a bad thing. If anything, a little friendly competition may make things better for everybody in the long-run. OTOH, if they are creating a proprietary product from their fork, then that’s pretty weak sauce.

    While forks can be demoralizing, frustrating, annoying, etc., it is the case that “right to fork” is pretty much a fundamental tenent of what it means to be Open Source in the first place. I vote “just soldier on” and keep making a kick-arse product, and let Camunda do what they will! And who knows, maybe the two projects will rejoin at some point down the road.

  10. Jonny March 20, 2013

    @Phillip Rhodes – the fork is Apache licensed, but that doesn’t change the core concern of duplicated effort. In the short term, it may be easier to maintain cross-talk between the forks. As the code bases grow, that gets much harder. I think a better path would have been for Camunda to just release their plugins open source, and continue to base them on Activiti. It’s not as though there’s anything wrong with the engine itself for their vision. It is true that their focus is different than Alfresco’s, but that hardly seems to necessitate a fork. Half the point of an Open Source framework like Activiti is that lots of people with tons of different visions can build on the same framework and succeed. Camunda could have done that, but chose not to.

    On the other hand, it may be a bit of the “pot calling the kettle black” for Activiti to shame Camunda for forking. Though it wasn’t actually fork of jBPM, the migration of development talent and effort from jBPM to Activiti could have been construed to have the same negative impact on another important Open Source BPM project. There were substantive reasons for doing so (some of which Tom outlined in http://processdevelopments.blogspot.com/2010/05/standalone-bpm-is-dead.html), but this likely seemed just as sad to some in the jBPM community.

  11. Joram Barrez March 22, 2013

    @Philip & Jonny: The nature of open source is indeed the ability to fork. But the point here was mostly that you don’t expect that to happen from people you worked together with for three years, without any direct cause (from our point of view) or without discussing it up front . The question here is whether the Activiti/BPM community is a better place now.

    I can’t actually say it any different from how Jonny puts it in his first paragraph. Thanks for putting that so clear. And it also makes me happy that people are intelligent and see through all the (marketing) messages.

    @Jonny: Indeed, I totally see your point regarding jBPM vs Activiti. From the user point of view you are absolutely correct. Whether it is a fork or a brain drain, it is the same situation in the end. However, at that time there were also other (let’s say more political and personal) reasons why it happened. I won’t go into the details, it has happened three years ago and it is a thing of the past. Of course the situation is different now, and the thing that bothers me the most is not the fork, but the way it happene. It really put a dent in my trust in humanity.

    Anyway, thanks for posting this. Really insightful.

  12. Rainer April 3, 2013

    Ever tought about to contribute as an eclipse or Apache project so that (even if you have already a good developer / community base) Activiti or the Process/BPMN Java API might get some kind of standardization ?

  13. Joram Barrez April 7, 2013

    @Rainer: Yes we thought about that route yet, but not really in depth. Altough I’m not a big fan of those ( a lot of bureaucracy), it might be something worth to consider for the future. Thanks for the hint.

  14. Hikari April 17, 2013

    I totally understand you. Apache opens the source to be used for free and closed, and most companies that do it won’t contribute back. But still they want the software to be strong and succeed, because they also don’t wanna maintain it alone. And many will still contribute back, influence its evolution, assure it will get better.

    While on the same time Apache brings investiment and paid work to developers who contribute to the project. Many hired developers, even on those companies that don’t contribute back, will work in their free time. It’s WAY better to have this kind of community then GPL ones where many ppl LOVE the software but have to drop it to work on something else.

    It is way better for developer professionals, because we can use the technology for free, read its code, learn it, start developing over it, get engaged on its community, until we become professionals on it and charge to do private work on it. It is a very more natural way of creating professionals than just interact with its UI, have paid training and go to work without even knowing how it is from inside.

    With that I understand why the fork hurts the community. Veteran consumers will be lost, new ones will step back.

    Private distros make it stronger, fork makes it weaker.

    What I could understand from their site is that they don’t want an embeddable solution, they wanna make it bigger into a full infrastructure. I’m sorry if I got it wrong, but it seems that Activiti is a framework, and Camunda will move as a final product.

    In a business model point of view, they don’t wanna sell development services to create software products over the OpenSource framework. They wanna have a strong standalone product, and sell plugins to it. Maybe they thought that’s a huge architectural move and didn’t feel like saying it before doing, because it wouldn’t be accepted by Activiti community and core contributors, who could work against them to stop the fork.

    If this is the difference in Activi x Camunda, I’d come with some kinda obvious question: why don’t they just develop Camunda infrasctrucutre over Activiti, maybe even making it proprietary and selling it? Obvious answer: they don’t wanna pay for the whole work, they want free help. They wanna change their business model, and current Activiti doesn’t fit in it, and they probably feel they’d not be able to change its evolution vision, so they needed a new software.

    If I got it right, I believe Camunda idea is better for a profiting company. Move the aimed profitable audience from development companies that develop over the framework to all sort of companies that use it. You’ll have many more buyers, since you’re going directly to end users. It will be easier for end users to have the final usable product than just a framework. End users will be able to get the free product, install it, use it, then start wanting more features, and then you come with your proprietary plugins! The userbase will grow easier and at the same time you’ll work directly for them, not for intermediary developers, who have fewer users. You’ll have a one strong OpenSource product, not many smaller or subsidiary proprietary ones.

    I’m sorry if I got it all wrong and said a lot of bullshit. I love OpenSource software, and love it even more when it gets me a job.

  15. Joram Barrez April 17, 2013

    Hikari, thanks for writing all this down. Much appreciated.

    > It’s WAY better to have this kind of community then GPL ones where many ppl LOVE the software but have to drop it to work on
    > something else.
    >
    > It is a very more natural way of creating professionals than just interact with its UI, have paid training and go to work without even
    > knowing how it is from inside.

    I couldn’t agree more!

    > With that I understand why the fork hurts the community. Veteran consumers will be lost, new ones will step back.
    > Private distros make it stronger, fork makes it weaker.

    Indeed. We know of some other companies that have build upon Activiti. But they do not call it an OSS fork. They could do it, sure, but they see the points you also make.

    > What I could understand from their site is that they don’t want an embeddable solution, they wanna make it bigger into a full
    > infrastructure. I’m sorry if I got it wrong, but it seems that Activiti is a framework, and Camunda will move as a final product.

    Yes, indeed. The main goal with Activiti is to have a core engine that is the best one possible. And of course Activiti offers tooling around it (you can’t do BPM without tooling). But that is just the point. The core engine should be the foundation of many other companies building their products on top of it. It should be a commodity. If there is any differentiation, it should be done in the UI’s, usability, features on top, … rather than an engine where no end-user actually cares about.

    > Maybe they thought that’s a huge architectural move and didn’t feel like saying it before doing, because it wouldn’t be accepted by
    > Activiti community and core contributors, who could work against them to stop the fork.

    Maybe … but after working with them for such a long time, I (and many others) expected such a discussion. That’s the decent thing to do.

    > Obvious answer: they don’t wanna pay for the whole work, they want free help. They wanna change their business model, and current > Activiti doesn’t fit in it, and they probably feel they’d not be able to change its evolution vision, so they needed a new software.

    Yes, you are probably right there. Although I fail to see what could be needed as a change to the core engine to make that happen. But your point about free help is definitely a valid one.

    > It will be easier for end users to have the final usable product than just a framework. It will be easier for end users to have the final usable > product than just a framework.

    I see your reasoning. Yet however, nor Activiti nor Alfresco is selling Activiti as a product. We’re in there for the engine and the framework bits in open source and our goal is not make money from Activiti directly (no services, training, …). The problem “putting something in open source” is that it’s not as easy as just proclaiming you’re open source and opening your code. Many companies have tried to ride that wave in the past and failed.

    > I love OpenSource software, and love it even more when it gets me a job.

    +1

    Thanks for your comment!

  16. Hikari April 20, 2013

    Well thinking better about my thoughts, it could have been better, techinical and ethical, if they had just created an OpenSource final product over Activiti framework, and marketed it that way.

    There are a lot of OSS out there that have other OSS inside them, and when a new version comes they just update it.

    They could have done the same. “Camunda, a final product built over well-kown Activiti”. They’d get what I think they want, and instead of intrigues they’d get Activiti community help and support.

    If Activiti needed some work on its interfaces they could just work it, would be kinda doable.

    Well you have been challenged, gl. I work with SharePoint and that makes me sad. What we could have had done in 1 week in Java we couldn’t even start in 1 month! and we’re paying for this junk!!

  17. Joram Barrez April 22, 2013

    @Hikari: I agree completely

    > Well you have been challenged, gl

    Thanks. And thanks for your insights!

    > I work with SharePoint and that makes me sad.

    I can feel your pain through the network cable ….

  18. Suminda Dharmasena May 3, 2013

    Hi,

    If there was 1 super product as Activiti it would have been better. One of Camunda’s core selling point other than creating buzz is:
    1) Embeddability
    2) Less code

    I think you guys should aim at achieving this plus:
    3) fluent API / DSL based configuration (Groovy / Scala)

  19. Nico de Wet May 19, 2013

    It would be useful to look into the reasons for the fork from a purely commercial perspective. I’ve read Camunda’s post, to me it seems to be a purely commercial decision but its still unclear why the decision was taken (it is entirely unclear). Once the reasons are clear perhaps it will be possible to see whether Activiti requires any form of adaptation to make forking unnecessary.

  20. Franklin June 2, 2013

    One of the reasons people are little bit skeptical about open source is the unavailability of Profession support. Why don’t you provide profession support or atleast ask the community if they would like to engage with a Profession support team. Maybe then you wouldn’t have to worry too much about Camunda.

  21. Franklin June 10, 2013

    @Joram: Amazing. Thanks a million!!

    Regards,
    Franklin

  22. Franklin June 11, 2013

    @Joram: I have been trying to reach out to BP3 on behalf our company, but unfortunately no responses so far.

    Regards,
    Franklin

  23. Scott Francis June 11, 2013

    Franklin – my apologies if your inquiry to BP3 have not been answered – I searched our email and logs and don’t have an email from you (maybe something getting blocked at the ISP or mail gateway) – but of course happy to help. I will have our team reach out to you today via email, hopefully this will get us started!

    Thanks!

    scott
    ps – we also have contact forms on our website that work, if you find that direct email/etc. isn’t working for some reason.

  24. Franklin June 12, 2013

    @Joram: Thanks a lot. Christopher Allen from BP3 contacted us. Appreciate you help.

    Regards,
    Franklin

  25. Arun June 12, 2013

    @Joram I am trying to reach edorasware for Activiti support but no replies yet

  26. Arun June 12, 2013

    @Joram got reply from edorasware just now 🙂

  27. Joram Barrez June 12, 2013

    @Arun: hahaha awesome!

  28. Sebastien June 17, 2013

    Guys you did such great software with JBPM and Activity, I have no doubt we wont hear much of a commercial fork (more or less like it happened for mambo CMS).

    Keep up writing great code, and many thanks for it,
    Regards,
    Sebastien

  29. James Martin February 3, 2014

    Activiti is pretty good, clear the Comunity version is very limited, we use the EE, which has activiti Ajax RIA interfaces, very good and efficient.

  30. Patricio March 12, 2015

    Hey…I’ve been using Activiti core API in my project and it flies like the wind…a cool loyal breeze. Thanks Activiti community

  31. al April 24, 2017

    haha – and now you have forked Activiti. Will you apologize to Camunda for what you wrote, or are you simply a hypocrite?

Leave a Reply

Your email address will not be published.