PLS_Toolbox Road Map

Here at Eigenvector we’re working hard on our next releases of PLS_Toolbox (and Solo). Given the changes to MATLAB over the past year we thought you’d appreciate some information about the roadmap to our next updates. 

Prior to the release of MATLAB R2025a, PLS_Toolbox was always compatible with the current release of MATLAB and an approximately 5 year window of past releases. However, as we pointed out in a January 2025 blog post, with The MathWorks’ (TMW) release of R2025a this was no longer possible. When MATLAB dropped support for Java it broke our most critical user interfaces (UI). To make matters worse, the R2025a uihtml interface tools did not include several key features that we needed to replicate previous PLS_Toolbox functionality. In order to provide the best experience for our users we had added lots of UI features missing from MATLAB, such as drag-and-drop, via Java.

Fortunately, the UI changes did not affect our command line functions through which all of our computations are done. So command line users and those who incorporated our functions into their own scripts were unaffected. 

Throughout 2025 we kept in close contact with TMW and monitored the development of additional html interface tools. As of MATLAB R2026a some, but not all, of the missing functionality has been replaced. TMW continues to add and improve user interface functionality. We have, for instance, partially restored drag-and-drop with a uihtml component. 

It is now our goal to have a version of PLS_Toolbox out that is compatible with R2026a+ by the end of 2026. This will require some changes to the work flow, but all essential functionality will be maintained. And the interfaces will have a more modern appearance (see below). Prior to that, by summer’s end, we will release our last version compatible with MATLAB R2024b and previous: PLS_Toolbox 9.6, This version will include new functionality including tools for lean chemometrics (IOT, etc.), conversions to volume fractions and improvements in Diviner. For our Solo users, we will build 9.6 with MATLAB 2024b, which will assure compatibility with current operating systems and improve performance and ease of installation. 

MATLAB supports only a limited span of the newest Python versions. This is problematic because this also requires that associated libraries like scikit-learn be updated. In order for us to curate Python tools we need to be able to stick with versions that we’ve validated. We are adding our own Python integration techniques so that we can maintain support for the most stable Python library releases, not the most current (with open-source software the most current version is almost never the most stable version).

We also expect that MATLAB’s html interface tools will change from version to version for some time. We are unsure about how MATLAB’s shift to web technologies will impact our tools. This means that future versions of PLS_Toolbox will likely be more closely tied to specific versions of MATLAB—it may be much harder to achieve compatibility with a wide window of MATLAB versions.  

We’re fully engaged in building our new interface infrastructure and look forward to delivering richer, improved interfaces. We appreciate your continued support as we do our best to support you! 

BMW