[midPoint] Future of midPoint (4.3 and 4.4)

[Radovan Semancik]

Dear midPoint community,

We are getting really close to midPoint 4.3 "Faraday" release, which 
also marks start of development of midPoint 4.4 LTS "Tesla" release. 
This is perhaps the right time to talk about our plans, as there are 
several important changes in both 4.3 and 4.4:

We have decided to drop support for MySQL and MariaDB in midPoint 4.3. 
MySQL/MariaDB support was deprecated for some time. Although we 
originally planned to keep MySQL/MariaDB supported in midPoint 4.4, we 
have decided to drop the support earlier. Keeping MySQL/MariaDB 
supported in 4.4 would mean that we have to support the databases at 
least for next three years, as midPoint 4.4 is an LTS. The midScale 
survey indicated no showstoppers in dropping MySQL/MariaDB support, 
therefore we have decided to go for it and rather spend the time and 
energy on more popular database engines.

Speaking about databases, the new "native" PostgreSQL repository 
implementation is taking shape nicely. It is huge undertaking, and there 
will be only a very rough prototype in midPoint 4.3. But we are fully 
committed to make it all work perfectly in midPoint 4.4. We would like 
to thank the NGI for providing funding for this essential improvement.

I have updated our documentation to reflect our current plans with the 
database support:

https://wiki.evolveum.com/display/midPoint/Repository+Database+Support

Support for production deployments of midPoint in Windows environment is 
deprecated in midPoint 4.3, and it will be dropped in 4.4. However, do 
no worry if your desktop runs on Windows. MidPoint will still run on 
Windows (and macOS), and these environments will be available for 
evaluation, demos, development and experiments. Our own developers are 
using a diverse set of operating systems, including Windows. Therefore 
midPoint has to run on all of them, otherwise we wouldn't be able to 
work. However, we have decided it is not worth supporting all the 
operating systems for production deployments. Production support means 
that we need to maintain extra tooling (such as run control scripts), 
spend time on extra testing and so on. The midScale survey indicated 
that there is almost no need for production-quality Windows support, 
therefore we have decided that it would be wise to rather spend the 
effort on things that are needed, such as better Docker support.

One of the things that was brewing (or rather rotting) under the surface 
is support for Jasper-based reports. We were not happy with the way how 
Jasper deals with reports. It is mostly aimed at rendering printable 
reports. However, our experience tells us that we rather need 
"structured" reports, something that can be imported into spreadsheet 
and further processed. This was also confirmed by 2019 midPoint survey. 
We have been working on replacing the Jasper functionality with a native 
midPoint functionality. And in midPoint 4.3 we think we are ready. 
Therefore we will be dropping Jasper support in midPoint 4.4, to avoid 
the need to support it for the next 3+ years. If you still have any 
Jasper-based reports, it is perhaps the right time to migrate them.

There were also removals of smaller deprecated functionality, such as 
support for HQL audit queries, support for Internet Explorer and so on. 
The details are in midPoint 4.3 release notes 
(https://docs.evolveum.com/midpoint/release/4.3/), that are now fully 
prepared for the release.

This may seem like a lot of feature removals. However, midPoint is a big 
system. We have to remove old stuff to make room for the new stuff. 
Otherwise the system would become an unmaintainable mess.

On the plus side there is MidPoint Studio. With the 4.3 release, the 
Studio becomes fully supported companion to midPoint. Support for Studio 
becomes a part of all midPoint subscriptions. There are also user 
interface improvements, the most notable is perhaps improvements to 
password management user interface. There are also numerous diagnostic 
improvements.

There is also quite a rich set of new experimental functionality in 
midPoint 4.3. Perhaps the most notable piece is a support for new, 
human-friendly query language (a.k.a. "Axiom query language"). Please, 
have a look at this experimental functionality. Most of the 
functionality will become a regular supported functionality in midPoint 
4.4. Having a look at this functionality in midPoint 4.3 and providing 
feedback may help to improve them before they take their final form. The 
community feedback is also the reason for having these features 
experimental in the first place - we want to have the freedom to fix the 
features in case you tell us that we have got it wrong.

Almost all of these decisions were based on input from the community. We 
have considered your input in 2019 midPoint survey as well as the recent 
midScale survey. We have taken inspiration with descissions with you and 
posts in the list. You have the power to shape the future of midPoint - 
and this is the right time to use that power. If we have got something 
wrong in midPoint 4.3, there may still be time to improve it before 
midPoint 4.4 is released. MidPoint 4.3 will be released in few days. 
Therefore go ahead, try it and provide a feedback. This list is an ideal 
place to discuss the future of midPoint. We will listen.

-- 
Radovan Semancik
Software Architect
evolveum.com