Clojurists Together funds five open source Clojure projects in Q2 2026

Clojurists Together funds five open source Clojure projects in Q2 2026

Clojurists Together has announced funding for five open source Clojure projects in Q2 2026, distributing a total of $31K USD across the selected developers. Three projects will receive $9K each, while two shorter or more experimental projects will receive $2K each.

The organisation credited its members for making the funding possible, and gave special recognition to Metabase, its newest Transduce member, for signing up to support work on the Malli library.

---

Clojurists Together has announced funding for five open source Clojure projects in Q2 2026, distributing a total of $31K USD across the selected developers. Three projects will receive $9K each, while two shorter or more experimental projects will receive $2K each.

The organisation credited its members for making the funding possible, and gave special recognition to Metabase as its newest Transduce member. Metabase joined specifically to support Ambrose Bonnaire-Sergeant's ongoing work on Malli, a library used heavily at Metabase as well as by many other members of the organisation.

The Five Funded Projects

The five developers and their projects are: Ambrose Bonnaire-Sergeant working on Malli, Dragan Djuric working on Uncomplicate AI: Clojure LLM, Cvetomir Dimov working on SciCloj Documentation and Plotting Libraries, Ingy döt Net working on Gloat, and Shantanu Kumar working on PluMCP.

SciCloj is described as a Clojure group focused on supporting the broader Clojure data science and scientific computing ecosystem, with Cvetomir Dimov's funded work targeting its documentation and plotting libraries.

How It Works: The Malli Memory Optimisation

A significant portion of the announcement details the technical direction of Ambrose Bonnaire-Sergeant's Malli work. In previous Clojurists Together-funded work, Bonnaire-Sergeant improved the performance of validating recursive references by bounding the amount of memory required for validation regardless of the depth of input values. This was achieved by eagerly expanding recursive schemas until recursive points are discovered, rather than lazily realising and caching new levels of recursion as inputs require.

While this approach prevents memory leaks caused by validating large inputs — improving reliability for long-running systems — it introduced a drawback: more memory is required upfront during validator compilation. This has been a challenge for Metabase, which has been navigating the trade-off between constant memory usage during validation and the higher initial memory cost.

Technical Direction and Availability

Two main approaches are under consideration to address the upfront memory problem. The first involves discovering recursion points lazily, which would reduce initial memory use but allow maximum memory to grow over time as large inputs are validated, making it harder to predict JVM memory requirements. The second approach focuses on reducing maximum memory usage directly.

Bonnaire-Sergeant identified that the increased memory stemmed from identical schemas using distinct validators each time they were referenced. The intended solution is to ensure that references to the same schema point to the same Schema object and validator. The funded project will investigate ways to allow systems such as Metabase to benefit from constant memory usage in recursive reference validators without incurring prohibitive upfront memory costs, with the aim of removing the need for custom workarounds that Metabase has introduced.

Story based on discussion on Hacker News.