[midPoint] MidPoint 4.x release schedule

[Radovan Semancik]

Dear midPoint community,

You have probably noticed that midPoint 4.0 is still waiting for its 
release. Those of you that are involved in midPoint development and 
testing are surely aware of the fact, that midPoint 4.0 development plan 
is delayed. I feel that it is my duty to explain what is going on.

MidPoint 4.0 has been an ambitious challenge from the very beginning. It 
has been (and still is) planned to be a long-term support (LTS) release. 
Therefore we have spent a lot of effort on removing and reworking old 
midPoint code. This makes perfect sense. MidPoint is now more than eight 
years old, with some code even older than that. We plan to support 4.0 
for at least three years, which means that at the end of 4.0 support we 
will have to deal with a code that would be more than a decade old. 
Hence a lot of refactoring, cleanup and re-engineering work in midPoint 
4.0. In addition to this, there was a couple of big features and a lot 
of smaller ones. We would like to make midPoint 4.0 a release that can 
be solid to be used for many years to come.

Everything seemed to go well with 4.0 - until February. I still cannot 
fully understand what has been happening since then. I can only describe 
it as unbelievably bad luck. We got struck by the seasonal flu much 
harder than any other year. There were travel problems, communication 
difficulties, unforeseen events and so on. Of course, we had reserves in 
the 4.0 development plan. But they were pretty much all gone in that one 
single month. March was quite similar. And then it got even worse.

At that point couple of platform subscribers were experiencing some 
performance issues. We haven't planned any major performance 
improvements in 4.0. But platform subscribers are setting priorities of 
midPoint development. Therefore we had to re-prioritize the work. What 
we usually do at this point is to postpone some of the 
originally-planned features to make space for those things that platform 
subscribers need. But at that time most of the 4.0 features were almost 
completed. And the refactoring effort was at a point, where it would be 
a tremendous waste to postpone it to the next release. Then there was a 
bug bounty program. It was a huge success, it was much better that I 
have ever hoped for. But that also meant addition demand for development 
resources. And as those are security issues, we had to deal with them 
immediately. We have moved few things from 4.0 plan, but it was not 
enough to keep the 4.0 release dates. I was announcing that to the 
community, hoping that from that point we can get to the full speed and 
finish 4.0.

But obviously, I was wrong. There was an injury. Then another injury, 
followed by a severe illness. This caused more communication problems, 
for which I have to personally apologize to all of you. Then more travel 
problems and pile of similar smaller problems. Overall, this whole 
situation is very difficult to understand and even more difficult to 
explain. Especially that this all affects me in a very personal way.

I'm quite aware that those of you that are our customers and subscribers 
do not need to accept our excuses. You have your contracts that need to 
be fulfilled. And we are making everything that is in our powers to make 
it happen. But as you are all aware, we are not a huge corporation. 
Being still a relatively small company, such a bad luck can affect us in 
a really significant way. The "bus factor" is often mentioned when 
discussing such risks of small team. We were always aware of this risk 
and we have been prepared: there was a sharing of knowledge in the team, 
communication, documentation, development culture and so on. And it has 
all worked. Several key persons in our team were pretty much "run over 
by the bus" during the course of 4.0 development, but MidPoint 
development went on. All the persons that could work harder re-doubled 
their efforts. The Evolveum team is a real marvel. However, there is 
still an impact. MidPoint development schedule is heavily affected.

We have been thinking and planning a lot, during last few weeks. We had 
to re-plan everything. It makes no sense to estimate 4.0 release to be 
"soon", when we know that it won't happen. Delay in midPoint 4.0 also 
means disruption of release schedule for following releases. Therefore 
we have discussed the plans internally. Then we have made sure that this 
the new plans can work for our platform subscribers. And the result is this:

Release 4.0 is planned for late August 2019.
Release 4.1 is planned for early 2020 (January-March).
Release 4.2 is planned for October 2020 - which should bring releases 
back to usual April-October schedule.

Original scope of 4.0 and 4.1 releases will be kept, as already 
negotiated with platform subscribers. Just the release dates are 
changed. I'm aware that this plan is not ideal. But it is the best that 
we can do in our current situation.

This was all a huge lesson to us. I have realized how lucky we were with 
midPoint and Evolveum up until that point in February. We want to make 
sure that the problems we are dealing with will not happen again. 
Therefore there are important changes that we are going to make. This 
includes my closer engagement in company leadership and adjustments to 
our business model. But first of all, we have to plan even more reserves 
in our development plan. We have to be more careful in scheduling 
releases. That is also the reason we will be skipping one release "slot" 
to make room for a more robust release schedule. And of course, we are 
hiring new talent to expand our team and we need to be more careful with 
overall organization of work in the company.

I would like to apologize to anyone who is affected by those issues - to 
midPoint subscribers, our partners and also to midPoint community at 
large. We have always been very transparent about midPoint development, 
and especially problems of midPoint development. MidPoint is really a 
substantial piece of software, and there will always be some problems. 
We do not want to hide any of that. We believe that full transparency is 
the best way to keep your confidence in midPoint and in the Evolveum 
team. Therefore I'm more than grateful for all your support to midPoint 
project. And I hope that we can count on your support also for the 
future. We will do everything we can to make midPoint better and stronger.

-- 
Radovan Semancik
Software Architect
evolveum.com