activiti activity dwindling ?

[heymjo]

Hi,

Just thought i'ld voice a bit my concern here.

Looking at the commit activity graph here https://fisheye.codehaus.org/changelog/activiti/ over the last year, i cannot help but noticing that the 'activity' on 'activiti' has declined rather visibly over the last 5-6 months.

[img]https://fisheye.codehaus.org/fe/commitSparkline.do?w=276&h=70&context=changelog&repname=activiti&ima...[/img]

Since we have chosen Activiti as our workflow platform for the coming years this worries me just a tiny bit at the moment. I've seen this happening a few times on corporate opensource projects and it does not always have a positive outcome.

Any reassuring thoughts to be shared ? (yes i know i can always get involved myself and make things happen etc etc but that is not what this is about)

Thanks !

[bernd_ruecker]

Hi Heymjo.

First of all: No activity doesn't have to mean a bad thing, if the stuff is good and stable 🙂

Actually the explanation is easy: currently the most upfront work for the core engine is done and Alfresco and us we are currently working on the product which keeps us busy. But this will even improve the security to have something in the long run if a correct product with a working business model is built around it. But you can expect quite some stuff from our side on the BPMN Engine in January or February.

What exactly do you miss exactly?

And by the way: With http://www.camunda.com/fox/ we have stabilized and supported version of Activiti, there you can get the security to have supported version, maybe that could be interesting for you? It includes some additional software artifacts as well (e.g. Websphere Integration, a monitoring tooling, …).

Cheers
Bernd

[heymjo]

Hi,

There is not something i'm missing at the moment. I was just looking at the activity evolution of the library. The current stuff is mostly good and stable as you say but that doesn't mean that there is no room for improvement or further evolution. Maybe a clearer roadmap would be useful in this case, but i understand that's not always easy to put together.

Thanks

[bernd_ruecker]

Hi Jorg.

I published a post about this today: http://www.bpm-guide.de/2012/02/17/activiti-project-alive-and-kicking/.

Hope that clarifies things a bit.
Cheers
Bernd

[heymjo]

Hi Bernd,

Thanks for the blog posts. Indeed it clarifies the situation a bit, though Activiti being 'professional open source' i had already suspected something along those lines. Keep this communication going please, us users like to have the feeling that we're informed on what is happening

Cheers

[ftibi]

Hi Bernd,

We are evaluating Activiti for driving some parts of an order management process for my current employer and we were also worried by the past "quiet" weeks. Your post explains what is going on - or at least your point of view / positioning in this emerging ecosystem. Unfortunately it raises for me and the team working on the evaluation rather more concerns, especially when looking at the area and feature set you intend to compete on.

I understand Camunda providing additional components towards a full-cycle enterprise product, but to compete on Eclipse Designer features is not a good signal for our evaluation. I don't want to underestimate the effort you all put in into this and the need to have a business case - it's fair - and of course it's your decision where you want to invest and what to offer commercially as a product. The reason why I say it sends the wrong signal is that it also means that instead of seeing the amazing speed of progress you guys showed so far, I see risks of your efforts splitting now before the product is actually finished. I don't mean having two connectors more or some very special BPMN shapes only in the Camunda version of the designer, but the message is disquieting. It would sound a lot better: we will all make sure to have a rock-solid designer for Activiti supporting BPMN2 etc. etc., while providing "premium features" such as XY in our extended designer etc. etc.

Pls. don't get me wrong - but saying Activiti Designer will only support a subset of BPM features to allow for workflow cases for document management while you can get the "real" thing from Camunda is leaving this strange feeling behind. (Quote: "Alfresco has a very different goal with Activiti, which is document management workflows within their ECM. We go for BPM." / http://www.bpm-guide.de/2012/02/17/camunda-fox-open-source-strategy/)

Maybe I got it all wrong and I apologize for it up-front - some statements from the Alfresco-Activiti guys would do good - I might not be the only one worrying about this now.

We must provide a decision for our project as soon as we see 5.9 (1st of March), that's the maximum we could stretch on the deadlines. And it's not about "free" vs. "paid for", it's really about the speed of catching up with similar offerings on the market - so far the product is really promising, clean and straight-forward, which is really appreciated by our team.

Thanks!
Tiberiu

[heymjo]

Pls. don't get me wrong - but saying Activiti Designer will only support a subset of BPM features to allow for workflow cases for document management while you can get the "real" thing from Camunda is leaving this strange feeling behind. (Quote: "Alfresco has a very different goal with Activiti, which is document management workflows within their ECM. We go for BPM." / http://www.bpm-guide.de/2012/02/17/camunda-fox-open-source-strategy/)

+1

Maybe I got it all wrong and I apologize for it up-front - some statements from the Alfresco-Activiti guys would do good - I might not be the only one worrying about this now.

Same here.

[trademak]

Hi,

I wanted to provide my view on the topic discussed here (more from an Alfresco/Activiti perspective).
First of all, you are right that the commit rates of Activiti have declined in the last months.
After we released version 5.7 of Activiti we didn't have major functionality that's lacking to provide a great embeddable BPMN 2.0 engine.

This means that we want the Activiti project to be driven by community / user demand. And we already welcomed a lot of new committers over the last months that provided patches for bug fixes or new functionality. So I agree with Bernd that a slow-down in commits isn't a bad sign perse. And with the 5.9 release coming up we have a lot of great new functionality that will be added thanks to the contributions of Camunda and other contributors.

The choice of Camunda to implement another tool to design processes in Eclipse is also not a bad one I think. This means that users will have a choice between the Activiti Designer and the Camunda tool. I want to make it clear that we still will further develop the Activiti Designer to improve it. I welcome any feedback, patches, thoughts etc to improve the Designer.

So to summarize, I think the stabilizing process of Activiti is foremost a good thing. And we still welcome improvements, fixes and new features as we always have. But we need the feedback of the community to point us to the most wanted fixes and new features. We have a lot of committers who are willing to help us implement the requests.

Best regards,

[bernd_ruecker]

Maybe as a short last remark on this: The Activiti Engine is pretty separated from the surrounding environment and tooling. This makes is that powerful. So we have quite some customers who never used an eclipse tooling but had other tool chains or even developed their own editor (in OEM scenarios) or leveraged something like kickstart. So I think there are huge different possibilities to use Activiti (especially because its focus is the embeddable engine). So what you can rely on in the Activiti OSS project for sure is the engine itself, despite all discussions in the other areas like designer. Then your goal what you try to achieve is important to select the right surrounding/tooling. And with Alfresco developing the Activiti Designer, us developing camunda fox or even other initiatives like http://sourceforge.net/projects/bpmn/ you have the freedom of choice. Or you might prefer vi 😉 All serve different use cases, Alfresco does ECM and we have the scope of providing a platform for BPM projects with Java.

Cheers
Bernd

[heymjo]

So to summarize, I think the stabilizing process of Activiti is foremost a good thing. And we still welcome improvements, fixes and new features as we always have. But we need the feedback of the community to point us to the most wanted fixes and new features. We have a lot of committers who are willing to help us implement the requests.

Jira's popular issues filter is always a good start for pointers on what users want (http://jira.codehaus.org/browse/#selectedTab=com.atlassian.jira.plugin.system.project%3Apopularissue...).

Also, few things are more demotivating than having your carefully crafted patch sit in Jira for a few months without any feedback. Looking at Jira there are currently 40 open issues containing the word patch, going through the first 10 i found several of them without any feedback at all. Taking care of your patch queue is a good way to attract new people on the project.