Lately we heard critics like these a couple of times: Is the Activiti project dead? There is so less activity in Activiti (nice worldplay), what happened to you guys? Did you turn your backs on Activiti and do something else? You haven’t solved the JIRA issues we filed for weeks, when do you plan to work on this? The last release of Activiti 5.8 was in October, when will Activiti 5.9 be released? BPMN element support lacks element x, when will this be included? And so on and on. Today I want to give some answer on this, since the Activiti project is more alive and kicking than ever, there “just” was a lot of productization effort and happened some things “under the hood” which couldn’t be easily seen.

Productization of camunda fox and Alfresco ECM

First of all I want to have a quick objective view on project activity. Lacking a better source I use the commit activity, as you can see on the left side. True, the curve does go less steep around the second half of last year, but we had some progress all the time. But why was it less steep? Let’s have a look at the main drivers behind Activiti, that is Alfresco on the one hand and we, camunda, on the other hand. Both of us were quite busy working on our commercial offerings. Alfresco included Activiti in their version 4 of the Alfresco ECM stack, and we built camunda fox as full blown BPM Platform for BPM with Java which has Activiti included. Productization shifts priorities from gaining cool new features to make them complete and get a product with a consistent vision out of it. For camunda fox this has a much wider focus than “just” the BPMN 2.0 Engine, even if that is the heart of it. It means a couple of additional components:

• integration: We integrated Activiti in various Java EE 6 application servers.
• cockpit: a web-tool for process controlling, monitoring and administration.
• designer: an eclipse plug-in for the developer to model BPMN 2.0 (not based on the Activiti Designer, see next blog post of today).
• modeler: a web-based BPMN 2.0 editor targeting business users (based on Signavio).
• cycle: a business-it alignment tooling capable e.g. of the round-trip with BPMN 2.0 between designer and modeler.

Producization includes a lot of tasks: requirements management, product management, alignment of the different tools or components, proper release management not only of components but of whole platform, automated QA on defined platforms and databases, performance and load tests and much more. So that is actually were there is a lot of activity in the moment: Making Activiti usable in enterprise real-world BPM scenarios!. A lot of these activities doesn’t result in concrete code but to a healthier project in the long run! But anyway, I attached the commit history of our internal camunda fox development on the right, where there is a lot of activity lately, we currently even have a dedicated Scrum team working on it.

Why do I tell you that? Because that matters a lot for the future of the Activiti project in the long run. Since we make sure that we have an aligned tooling and a business case, where we can earn money. We build a stable business on top of Activiti, as does Alfresco. We both worked on our product in the last months, so there was less time to invest in the Open Source Engine and the Community. I am sorry for this, but actually I think it was a good decision for the longer future.

But wait, shouldn’t investment in the community pay off by itself? That is one of the drivers of open source, give-and-take, right? I think yes, but first a process engine is not Linux, it is a niche product, and second the core of the engine is quite critical, where you should not mess around. Typically contributers come when they work on a customer project, but go when this project went live. We always have ramp-up and project management overhead with every contribution, that is why there are not just a present. Every patch or feature provided must be reviewed, properly tested, maybe discussed and maintained in the long run. Just because it is open source don’t ease these requirements of professional software engineering.

And even then it is sometimes hard to channel these contributions into proper product management. And product management is something a lot of open source projects are missing. Done right this leads to a better product. Unfortunately I forgot who told me this quote, but it goes: “The problem of Open Source projects is, that they don’t have a customer“. I think it is true, you have to have customers driving your product and priorities, then you get in the right direction, but you need to invest time and energy to listen to your customers and process their feedback.

This adds another possible stumbling block: If you have passionate individuals in your open source project, coding new features in their free time at night, they easily become frustrated if business priorities does not match personal preferences. Let’s face it: This is why professional software development is not a hobby, it is a job (even if it involves much more fun and passion than a lot of other jobs, puh, lucky us). Another way of thinking about it is expressed in the book “crossing the chasm“: We try to make Activiti/camunda fox ready to enter a market, where you don’t have to be a Java crack in order to get a process engine running and where you have to set up your tool chain yourself. It is about providing some “peace of mind” (which should NOT be confused with zero code bpm suites, but that is something I think I don’t have to mention again in this post). It shall bring the right tooling even to people which now have to use some proprietary but technically inferior tooling because of lacking productization in open source technology.

So this is my thinking on commercial open source. This is why I think that productization is the more important thing in the current phase for Activiti even if that means that we have less time for the community. It will be good in the long run.

But: Amazing new features in Activiti 5.9

That’s said, there are some really good news about Activiti 5.9, which will be released on first of march (as documented in our Wiki by the way). We did develop quite some new features, most of them for our customer to be honest, but available for everybody, e.g.:

• Intermediate signal event (throw / catch),
• Intermediate none event (usable for KPI’s),
• Event based gateway,
• Interrupting error event subprocesses,
• (Multiple) message start events,
• BPMN 2.0 transaction sub process, cancel end event & cancel boundary event,
• BPMN 2.0 compensation activities and throwing compensation event.

Wait for the final list when we will release 5.9, some of the features mentioned will be marked experimental until they have stabilized, we expect Activiti 5.10 to contain everything rock-solid! Regarding transparency, you can see the detailed issues in the Activiti JIRA. That’s quite some work we did at camunda for our open source community, isn’t it? So Activiti is far from being dead. Believe me

Since this post started to get a bit more philosophical anyway, I want to add a second blog post immediately about the camunda fox open source strategy, since this is a question we get as well, especially since Activiti Cycle became camunda fox cycle, feel free to read as well: camunda fox open source strategy.