DYNA Fall 2025

As the landscape changes, so must the road map. Morten opens DYNA Fall 2025 with an overview of where we are and where we are heading.

AI is evolving rapidly, but APL support is lagging behind that of some of the more mainstream programming languages, such as Python. Stefan outlines the work that is being done at Dyalog Ltd to ameliorate this situation, including Model Context Protocol (MCP) to connect a Large Language Model (LLM) to a running application, smart documentation search, client libraries, and more.

Jarvis is Dyalog's APL-based API framework that enables developers to expose APL functionality as web services and APIs. Whether you're building microservices, creating RESTful APIs, or exposing computational services, Jarvis provides the tools to make your APL code accessible over HTTP. Brian introduces JAWS (Jarvis and web sockets), with which Jarvis can provide asynchronous delivery of services, making it even more flexible.

Apache Kafka is an open-source distributed event-streaming platform for processing real-time data streams in a fast, reliable, and scalable way. It allows you to build high-performance data pipelines and streaming apps. Martina presents the new Dyalog interface to Kafka, written in APL and C, and demonstrates the fundamentals for developing a Kafka-based application in Dyalog.

Aaron and Brandon introduces the work that’s being done at Dyalog Ltd to provide static analysis tools for users of APL that should improve tooling, assist in code refactoring, and aid migrations. They discuss the security and static analysis tool built on Co-dfns that will be available as an independent tool targeting security and code quality audits, as well as how you can use Co-dfns to integrate static analysis into existing tools.

METSIM® is the leading chemical and mineral processing plant simulation software used by engineers and operators around the globe. Since 1982, all development has taken place using APL+Win, which served their products and our customers well. However, the environment in which all industries are operating has evolved, and improvements became necessary; Dyalog was selected as the replacement system to provide the tools and expertise in-house to ensure that METSIM continues to meet customer demands. For the last 18 months, teams from MSI and Dyalog Ltd have worked together and successfully converted the METSIM source code to the Dyalog platform without losing any key features. Alex describes the conversion process, examining issues such as moving from Microsoft Windows specific APIs and UI to cross-platform tools that will also support Linux, without rewriting the application. He also discusses the planned source-code improvements, and the external APIs that will expand MSI's product offerings.

Buyers Intelligence Group (BIG) provides inventory analysis services for thousands of jewelry manufacturers and retailers. By integrating with a wide variety of point-of-sale and other systems, BIG provides analysis tools to help its clients improve inventory turnover. Although the web front end is written using industry standard components such as ASP.NET, MSSQL, JavaScript, and KendoUI, much of the back end is written using Dyalog and its myriad of tools and interfaces.

Dyalog enables BIG to react quickly to changing needs and provide custom solutions tailored to individual clients. Mark introduced us to BIG's use of Dyalog in his presentations at Dyalog '22 and Dyalog '23. He now updates us on how BIG's use of Dyalog and its tools has continued to grow, and how they are actively exploring AI integration to push their capabilities further.

In Dyalog, namespaces are a convenient way to organise a workspace (analogous to the way in which directories are a convenient way to organise a file system). Namespaces simplify application development by allowing you to give the workspace a structure; they can be used as containers for objects, data, and code modules. Whether classes are implemented in APL, or external classes like ⎕WC GUI objects and .NET classes, each object resides in a namespace.

In this workshop, Morten demonstrates the value of namespaces, which have become cornerstones of modern APL application development. He shows how they have been integrated into the APL paradigm, their power, their limitations, and the fundamental differences from nested arrays, where a change to an element of an array that has more than one name will force a copy which decouples the values from each other. He also explores how namespaces enable APL to easily interface with most modern technologies, which are often object-oriented. There is time for you to experiment and become comfortable with the tricky bits.

Assumed knowledge: Working knowledge of APL (ideally, but not necessarily, Dyalog APL)

Jarvis is Dyalog's APL-based API framework that enables developers to expose APL functionality as web services and APIs. Whether you're building microservices, creating RESTful APIs, or exposing computational services, Jarvis provides the tools to make your APL code accessible.

JSON and REST support:

• JSON mode – expose APL functions directly as callable API endpoints with JSON payloads
• REST mode – build resource-oriented APIs following REST principles. with full HTTP verb support

A Jarvis API can be consumed by almost any type of client. For example:

• other microservices
• APL applications (using HttpCommand)
• JavaScript front ends
• directly from Python, C#, Java, and so on
• mobile apps and browser-based UIs
• CLI tools, for example, curl

UI-agnostic design:

• Jarvis doesn't handle UI directly but rather returns structured data and lets API consumers handle presentation

With the release of JAWS (Jarvis and web sockets), Jarvis enables asynchronous delivery of services, making it even more flexible.

In this hands-on workshop you will:

• get an overview of Jarvis.
• implement your own simple web service.
• add an asynchronous aspect to your web service.

Assumed knowledge: Working knowledge of APL (ideally, but not necessarily, Dyalog APL)

Most actively-maintained APL systems are switching to using text files for source, rather than workspaces or component files. This allows integration with source code management systems, build and test frameworks, or even to experiment with "agentic AI". If you have not yet experienced the benefits of text-based source, or you understand the benefits but need help to get started, this workshop is for you.

Link allows you to use Unicode text files to store APL source code, rather than "traditional" binary workspaces. The workshop includes:

• how to find and use the documentation.
• the types of APL objects that can be handled, and how each is represented as text.
• how Link synchronises changes made inside and outside the active workspace.
• migrating an existing workspace to text-based source.
• using GitHub to manage “Linked” source code.

Assumed knowledge: Working knowledge of Dyalog APL

The key operator ( ⌸ ) is used to perform GROUP BY-style aggregations in Dyalog. In this workshop, Asher explores ways to import and pre-process data, looking at the fastest formats for data analysis in the workspace. He investigates techniques for aggregating and inspecting large datasets, including methods for efficient and convenient computation on the grouped data, and walks you through analysing a real dataset using ⎕CSV and ⌸.

Assumed knowledge: Basic knowledge of Dyalog APL, high school mathematics. Knowledge of GROUP-BY is helpful but not essential. No prior knowledge of ⎕CSV and ⌸ is necessary.