[midPoint] MidPoint in 2018

[Radovan Semancik]

Dear midPoint community,

Happy new year! 2018 is here. And it looks like a very exciting year for 
midPoint. There are new development plans for 2018 but also interesting 
changes in our business model. There is so much to look forward to in 
2018. I would like use this start-of-new-year period to share our plans 
with you. This is going to be quite a long mail. But there are important 
information to share with you. So please make yourself comfortable.

First let’s have a look back at 2017. This was yet another busy year for 
midPoint. Especially the Comenius release brought huge amount of new 
features. This was followed by Darwin which brought a wide range of 
evolutionary improvements. It means that MidPoint is now much more than 
just IDM. MidPoint is planted deeply in the identity governance field. 
In fact it is the first (and only) full-featured identity governance 
system in existence. Which I guess makes it a best-of-breed solution, 
doesn’t it? MidPoint is also recognized by major industry analysts and 
the number of midPoint deployment is growing. So, we can say that we 
have reached and even exceeded our plans for 2017.

And now the plans for 2018. The plan is, as usual, to have two releases. 
MidPoint 3.8 is planned for early spring, midPoint 3.9 will come in 
autumn. We have quite precise plans for 3.8, but there are preliminary 
plans for the whole of 2018. Let’s start at the beginning.

The development of midPoint 3.8 is already in full swing. There are two 
major focus areas: scalability and data protection. The goal is to run 
midPoint efficiently even for massive deployments. Currently it can 
handle deployments with few millions of identities – assuming 
appropriate configuration, environment and quite a bit of patience. But 
midPoint 3.8 should be able to go beyond that limit. Our team is working 
on the scalability and performance improvements right now. The second 
area where midPoint is being developed is data protection. Yes, that 
means GDPR. Identity management systems are almost ideal tools to 
support and automate data protection mechanisms. But midPoint goes a 
step further by introducing features specifically aimed at data 
protection into the base product. This is still a bit of a secret (as 
much as anything can really be a secret in an open source world). We are 
going to unveil the details at FOSDEM conference. So it might be a good 
idea to head for Brussels in first weekend in February. This is going to 
be fun:

https://fosdem.org/2018/schedule/event/idm_midpoint/

MidPoint 3.8 will be released quite soon, most likely early April. The 
plan is to get back to our original April/October release schedule.

Precise plans for midPoint 3.9 are still open. Significant part of 
development time is reserved for platform subscribers. But the 
preliminary plan for this release tentatively aims at improving user 
experience (UX) of the user interface. There was a gradual evolution of 
the user interface in every release to date. But current situation asks 
for a bigger review of the UX principles of midPoint user interface. The 
extent of the improvements is still not entirely certain. There are a 
lot of ideas and suggestions and it is quite clear that not all of them 
will make it to midPoint 3.9. Which improvements will be included mostly 
depends on preferences expressed by the subscribers. Therefore, it is 
perhaps a good idea to purchase your subscription now to make sure that 
your voice will be heard. There are only a few months left before 
midPoint 3.9 development starts.

There are also a couple changes to Evolveum business model. First change 
affects the lifetime of the releases. Up until now every release was 
supported for two years since the initial (minor) release. This was a 
good model when midPoint was young. It was ideal for early adopters that 
wanted new features very quickly. But now when midPoint is a mature 
product stability is much more important than rapid feature delivery. 
Therefore, in 2018 midPoint will be adopting the system of long-term 
support (LTS) releases. There will be an LTS release every two years and 
those releases will be supported for a longer time. The other (non-LTS) 
releases will be supported for a shorter time. There will also be a 
direct upgrade path between LTS releases. Therefore, the LTS releases 
are for those that prefer stability. Other feature releases are for 
those that prefer rapid delivery of new features. There is something for 
everybody.

The other change in our business is the strong preference of 
subscriptions over sponsoring. Up until now there was a possibility to 
directly sponsor a specific feature. We will be phasing out this 
offering during 2018. Sponsoring can pay for the development of a new 
feature. But it does not cover the maintenance cost. Only subscription 
can do that. MidPoint is now a mature system and feature stability and 
continuity is crucial. Therefore, we decided to support this stability 
with appropriate business model. From now on the right way to get your 
feature into midPoint is to use platform subscription:

https://evolveum.com/services/professional-support/

There is also one change that is very likely to affect this mailing 
list. Since the beginning of midPoint project back in 2011 we have used 
this mailing list as a primary means of technological communication. We 
definitely want to maintain that, and I take it as my personal new year 
resolution to try to communicate our plans more often and more clearly. 
The mailing list is also used as a community support forum. Which is 
good and it is a pleasant sight to see such a vibrant community growing 
around a product that we have created. However, having Evolveum 
engineers answer every community question is an enormous drain on our 
resources. You might have noticed that we have reduced the amount of 
time that our engineers use for community support during last year. I'm 
afraid that this is a trend that is very likely to continue. The 
community gets bigger and it is not possible to answer every question. 
Therefore we have to prioritize. And of course, midPoint subscribers get 
absolute priority. We will also try to answer as many community 
questions as possible, but only to the extent that time availability of 
our engineers permits. I'm sorry about this, but even our engineers need 
to get a bit of sleep occasionally. On the other hand as midPoint 
community grows there are more and more cases when a community question 
is answered by another community member. And I must say that I'm very 
happy that midPoint community has this life of its own. Thank you all 
for that!

There is a similar situation when it comes to bug reports. It is always 
good practice to file a bug report in our Jira. However, I would like to 
clarify the expectations about the fix. All issues reported by midPoint 
subscribers will be prioritized. Subscriber issues will be handled as 
soon as possible. If the subscriber specified a higher priority for the 
issue then the issue will be fixed in next (minor) midPoint release. If 
the subscriber indicated that this issues should be part of a 
maintenance release then we will backport the fix. To summarize: 
subscriber is a king. Then there are (non-subscriber) community issues. 
We will prioritize fixing community issues only in two cases: if it is a 
security issue or if the bug affects huge number of midPoint deployments 
(security issues will always get highest priority regardless of who has 
reported them). Other community issues have to wait. We will not work on 
these issues immediately. We will not ignore the issues, we will just 
assign lower priority to them. We have a test/bugfix cycle before every 
midPoint release. That's the time when community issues may get fixed - 
but only after all the subscriber issues are fixed. However, the 
experience from recent midPoint releases tells that this time is very 
limited. Community issues will not block scheduled midPoint release. 
Therefore it is not realistic to expect that all community issues will 
get fixed. Unfixed issues are postponed for the next release. But the 
priority stays the same. So there may be quite a long time before the 
issues get fixed. In fact there are community issues that are already 
postponed for several years. I'm sorry for this. We would really like to 
fix all the issues. But our capacity to fix community issues is limited. 
At this point I'm quite certain that you are aware where this all leads 
to. And you are right. Please, get midPoint subscription. That is the 
right thing to do to keep midPoint project going.

So, this is the current state of midPoint project and those are our 
plans for 2018. MidPoint is the biggest open source IDM system out 
there. It is actually bigger than all the other open source IDM systems 
combined – both in the quantity of the code and quality of the features. 
MidPoint is the only open source system that implements strong identity 
governance. The midPoint project is a great success. All these years of 
hard work were worth it. Even though vast majority of the work was done 
by the core team in Evolveum, the project would not be possible without 
you: midPoint community. We would like to thank all midPoint 
contributors and especially midPoint subscribers. MidPoint would not be 
such a great product without all your support. Thank you!

-- 
Radovan Semancik
Software Architect
evolveum.com