Hi all, I agree with removing the bi plugin. I initially contributed it as a proof of concept several years ago, and since it never evolved into a production-ready component or received the maintenance it would require, I think removing it is the right decision.
More generally, I believe we should take a similar approach with other OFBiz features that are still experimental or have become less relevant in modern systems. Reducing the size of our codebase would help address one of the main challenges our community faces: the effort required to maintain such a large project. Thanks for raising this. Jacopo On Mon, Oct 26, 2026 at 8:51 AM Mridul Pathak <[email protected]> wrote: > Hi all, > > Following the BIRT thread, I would like to propose removing the bi plugin > (plugins/bi) from trunk as well. In 2014 we agreed that it should be > disabled by default, but it has stayed enabled, so in a default setup it > creates the OLAP tables and registers ECAs on the core createProduct, > updateProduct, setInvoiceStatus, setOrderStatus and changeOrderItemStatus > services, although nothing is loaded until someone runs Quick Init > DataWarehouse by hand. In my experience, populating a warehouse through > ECAs on core flows has never scaled in production, and deployments end up > disabling it. The date range that init works from (bi.properties) ended on > 1 September 2026. > > It is also still at proof-of-concept level. The 2020 effort to make it > usable (OFBIZ-11414, pull request #11 with 100 commits) was closed without > being merged in March 2024, and nothing suggests it is being exercised: the > regressions in its fact-loading services that we found during the > minilang-to-Groovy audit (OFBIZ-13536) had gone unnoticed. Its only > consumers, the BIRT reports and the order report events that read its star > schemas, are going away with BIRT. > > If there are no objections, I will file a JIRA and open a pull request. > > Thanks, > Mridul Pathak >
