Behind the software lie the humans

Michael Bailly, OpenPaaS’ Product Owner, under the spotlight

Jul 25, 2018 · 9 min read
A representation of each and every files and directories located in the main OpenPaaS repository, as displayed by the open source software Gource. Directories appear as branches and files as leaves.


Just like a plant, a software requires constant care so it can grow and flourish. OpenPaaS is no exception to the rule, except that it would be better pictured as the fertile soil — the platform — and the collection of plants that are being nurtured by it — its set of software. In other words, . If OpenPaaS is like a blooming garden, its main botanist-in-chief and visionary warden for the past five years has undoubtedly been Michel Bailly, OpenPaaS’ Product Owner. Where OpenPaaS comes from?; where it is heading? Let the interview begin to figure that out!

Sure, my name is Michael Bailly and I am responsible for OpenPaaS, LINAGORA’s flagship product. As such, I have three main missions:

  • My first and perhaps most important one is to who makes up the team: geeky developers, business analysts, communication specialists and testers.
  • My second mission is to , which means to define the road map and to sort the backlog.
  • Finally, my third mission is to make sure customer deployments go well, and in general .

. Technically, the first version has been released in December and publicized in February. However, at that time it was still lacking some important features: it couldn’t be deployed in production just yet. In just over six months, a lot of efforts has been invested by the team to change that, to make OpenPaaS more stable and feature rich, and in general . Two main aspects were improved :

  • The first aspect was to do correctly. Deploying to the cloud is vastly different as what you do, with say DEB or RPM packages.
  • The second thing was about and bug fixes.

I am now using OpenPaaS every day and I feel a real stability, an ease of use, a demonstration that the product has indeed matured enough to be used in production.

I guess you’re referring to competing products like Office 365 or G Suite. I think we need to put the product in the right place. It is true that OpenPaaS provides general purpose collaboration features such as mails, calendars, contacts and the like. But it goes beyond that, as it can also be used to become , allowing the interconnection of legacy software that are usually mostly running in silos, as well as to allow the interconnection with external partners. In other words, it is important to insist on the fact that.

The second thing, which I think is hugely important, is the aspect of our product. Our competitors are far from being Open Source. We, on the contrary, are Open Source from the bottom to the top, something that makes us genuinely proud. LINAGORA is insisting on this aspect and was able to deliver a high level of professional quality while at the same time remaining truly committed to the Open Source philosophy.

The third thing is , a component we refer to as ethical computing, and which is extremely important. When our customers deploy our solution, they can be fully assured that their data will not leave their infrastructure. It is reassuring for them.

I could just say that all of our customers are satisfied but that would be dishonest.

On the negative side, customers expected to see features that were lacking at that time, such as interoperability with popular proprietary software. I am confident that this is only a matter of time before we can implement all of those. OpenPaaS Antares is a significant step in the right direction.

On the positive side, core functionalities are working well. In short, the application does its job and does it well. Secondly, OpenPaaS is reported as being pleasant, simple and intuitive to use. For us this is a great achievement because we have indeed invested a lot of energy to have a , something that is quiet unique for an Open Source project.

Simply. We create a customer-specific backlog that includes two kinds of tickets: tickets for new features and tickets for bug fixes. We collect issues and inject those in the client backlog. Then we discuss with stakeholders and we prioritize solutions.

In 2007, LINAGORA acquired , a groupware, and was able to market and sell the software successfully. However, it turns out that in 2018, architecture standards and models are different than ten or twenty years earlier, when there were no smartphone and when OBM first came out. Rather than to refactoring OBM, . LINAGORA invested significant resources to make that happen.

The second thing is research. OpenPaaS started as a research project in 2012, a project that is still running today. Thanks to our research team, the technical foundations and collaboration aspects were laid out during this period. It has been rewarding to work with people from universities and prestigious laboratories which were able to shape the product at an early stage.

At first, OBM has been developed to do mail, contact and agenda management, and to do it well. It was in the early 2000s and groupware were all around. OpenPaaS goes beyond that: which alongside business collaboration applications such as contacts, communities, instant messaging, and the collaborative edition of documents, provides a less visible foundation that to me is more important than the rest. This part, which I have already mentioned, is the around which businesses will be able to expand services and design business logics.

Let’s take the example of a large hospital: in a hospital, we need a business software like an appointment booking software for doctors. Obviously, it takes a calendar system because you need to register bookings but it goes beyond that: you also have to know what doctors are doing, what is their specialization so we can match them with the right patient. If I am looking for a rheumatologist, I don’t want to be received by an ophthalmologist. The kind of system we are looking at must manage leaves and small work-flow of acceptance. In the example, and to allow the hospital to create its own business application on top of it, an application that could be used by other hospital as well.

OpenPaaS under the hood. The user on the left is accessing the platform through her web-browser. When she sends an email, it will be processed by James and stored in Cassandra. When she creates an event in her calendar, it will be dealt with by SabreDAV and then stored in mongoDB. In either cases, Elasticsearch is building its own database and indexing the information. Whenever she will look for an event or mail, she will be able to look for it by using a simple query.

So actually, the first important thing to report, it seems to me, is that we are going to switch to a system called . New versions will be released on predefined dates, and will be published three times a year: on February, June and October. It will offer our customer more visibility into our road map. For the next release, we will continue to polish out our product and thrive to answer customer-driven issues. More concretely, we will rethink the way different business objects in the platform can interact, to bring more flexibility, to be able to stick closer to our different customers use cases and workflows. In short,, intended to facilitate their life.

The social part will be improved: we would like to rework the social part, and to ship a mobile application, so you could chat as you would do with any other popular solution.

On top of that, if we have enough time, we will deepen LinShare’s integration into the OpenPaaS platform. As of now, we already have a nice level of integration but we have thousands of ideas to make it even more fun to use.

As a result of the OpenPaaS research project, we have been able to integrate to OpenPaaS. CryptPad, which is developed by , one of our partner, is basically . Its integration to OpenPaaS is now done, which is great because it means that within our platform, users have now the opportunity to work collaboratively in real-time. Although CryptPad is rich and have quite a few formating options, it is still less powerful than a more traditional word processor, and for instance won’t typically offer a traditional page layout.
This is the reason why, apart from CryptPad, we are now working to , following a two-steps strategy. At first, we will integrate with the cloud. In a second time, we would like to integrate the OnlyOffice server directly within our platform, as a micro-service. It will guarantee that customer’s data do actually stays at home.

Which one do you like the most? The interview within the interviewRichard Stallman or Linus Torvald?
> Linus Torvald
Reddit or Hackernews?
> Hackernews
VScode or Atom?
> VScode, unfortunately
> KDE obviously, as it is the only one that works
> AZERTY by habit, but I have nothing against the QWERTZ layout
Bash shell or Z shell?
> Bash

Internally, everyone likes : we have no reason to switch to anything else. We are keeping it up-to-date which is usual when you are in the software editing business. The problem lies more on the front end: on this side, we are using version 1, and this framework has been supplanted by version 2 and more. However, version 2 is largely incompatible with the version we are currently using. We are now considering moving to a new framework.
Nowadays, there is a very strong community and business attraction towards a framework called . Apart from that, a third framework called is skyrocketing because it is simpler to use than React and entirely community-driven, as opposed to AngularJS which is backed by Google and React which is backed by Facebook.
As of today, the strategy for OpenPaaS is to keep AngularJS as a framework for dealing with personal information management: after all, it has served us well so far. On the more socially-oriented parts of the platform, VueJS is going to be used.
In the long run, the idea is to offer developers a way to develop applications for the OpenPaaS ecosystem regardless of the framework they currently use. This goal could be achieved by providing for the three main aforementioned frameworks, and any other one that could appear in the meantime.

It is a bit boastful of me, but first of all, they are doing . Secondly, in France, there is a small problem which is that we give less value to technical expertise than to project management functions, which I think is a mistake. , and as a developer you should not hesitate to go along this path rather than to become a team manager. Thirdly, as a developer you tend to always look for technical challenges to solve, which is what we are expecting of you. But you should not forget what the end goal of your job is, which is to . For instance, changing your test framework every month is great but it does not create any value for the customer who bought your product. In sum, your work will always be about balancing several priorities, such as what can you do to improve your productivity and what can be done to deliver more value for the customer.

You’re welcome.

Keep in touch with on Twitter, Facebook, GitHub, and our dedicated Forum.

Interested in joining ? We are hiring!

Linagora Engineering

We are Open Source Engineers, Hacking Awesome Stuff