Technology Tales

The October 2025 release of Visual Studio Code introduces major enhancements across agent management, security, and the overall editor experience. The new Agent Sessions view centralises local and cloud agent activities, with improved options for organising, searching, and delegating tasks to coding agents such as Copilot and OpenAI Codex. Developers benefit from a new planning agent, refined cloud and CLI agent integrations, as well as easier session tracking and custom agent configuration. Code editing sees improvements like selectable deleted code in diffs, the open-sourcing of inline suggestions, expanded navigation features, and enhanced iconography. Additional updates include advanced settings options, better command palette search, the introduction of Terminal IntelliSense for command completion, richer diagnostic copying, and easier management of authentication and language models. The update brings increased accessibility features, notebook search, new source control capabilities such as folding in commit messages and improved branch/tag management, and more robust testing navigation tools. Python development is refined with improved environment management, AI-powered documentation insertion, and new code actions. Extensions receive more authoring capabilities, including new authentication options and richer UI elements. Numerous issues and bugs have been addressed by community contributors, and support for macOS 11.0 ends with this release.

Posit has announced new tools and integrations designed to streamline data science workflows for enterprise teams. The company is introducing Positron, a new integrated development environment built on VS Code that supports both R and Python whilst enabling data exploration, model building and application deployment within a unified interface. Posit AI Agents provide automated assistance through Positron Assistant for coding tasks and Databot for natural language data queries, maintaining transparency and reproducibility. The platform now integrates directly with Snowflake and Databricks, allowing teams to develop and deploy applications without leaving these environments whilst leveraging their native governance structures and AI models. Chronicle Usage Metrics helps organisations track how data science resources are being utilised, monitor licence consumption and plan for future capacity needs. Enhanced auditing capabilities provide detailed logging of user activities, flag packages with known security vulnerabilities, and enable better tracking of potential exposure risks. These updates aim to address common challenges in enterprise data science by providing connected tools, consistent governance and seamless integration with existing data infrastructure.

SAS and R are both statistical programming languages, though SAS has long been part of many organisations' legacy software systems, whilst R offers innovative reporting and automation capabilities that are increasingly valued in sectors such as pharmaceuticals and finance. Many recent graduates arrive in the workplace already familiar with R from their university studies. Although translation guides exist to help programmers move between the two languages, learning R properly from foundational principles enables users to fully exploit its potential rather than simply replicating SAS workflows. The transition involves becoming comfortable with RStudio as a development environment, understanding how to import data through packages like readr and haven, adopting the tidyverse collection of packages for data manipulation and visualisation, and learning to create reusable functions and packages instead of relying on macro libraries. R Markdown and Quarto serve similar purposes to SAS ODS for producing shareable outputs, whilst the tidymodels collection provides statistical modelling capabilities and Shiny enables the creation of interactive applications for sharing analytical results with collaborators who may not themselves programme in R.

Claude Code is a command-line tool designed for developers that allows direct interaction with large codebases through a terminal interface. It can rapidly search through millions of lines of code, provide detailed explanations of project structures and architectures, and automate complex development workflows that would traditionally require switching between multiple tools. The system integrates with popular development platforms like GitHub and GitLab, enabling users to handle complete development cycles from reading issues through to submitting pull requests without leaving the terminal. It offers intelligent code analysis that understands dependencies and project organisation, allowing it to perform sophisticated multi-file edits whilst maintaining code integrity.

SAS Webinars

A range of webinars and events are available, focusing on topics such as AI technologies, data analytics and their applications across sectors including healthcare, government, banking and education. These sessions explore subjects like synthetic data, machine learning in SAS Viya, agentic AI and ethical considerations in AI governance, with specific emphasis on improving decision-making, enhancing operational efficiency and addressing industry-specific challenges. Events include live and on-demand presentations, with dates spanning from February to October 2026 and cover collaborations with institutions like Purdue University and AWS, as well as practical guidance on implementing AI solutions and leveraging analytics for business transformation. Resources are also provided to support technical learning, industry trends and the development of skills in data management and AI innovation.

Configure SAS to Run the Python Language

Administrators can configure SAS to run Python code in either standard or lockdown mode, depending on the use case. For use with PROC FCMP, Python 2.7 or later is required, and the administrator must set two environment variables,

MAS_M2PATH and MAS_PYPATH, within the sas.compute.server startup commands configuration instance, pointing respectively to the mas2py.py file and the Python executable. To enable PROC PYTHON and Python code to run in lockdown mode, Python 3.0 or later is required, along with the Pandas DataFrames package, and the administrator must set the PROC_PYPATH variable to the Python executable path, while the PROC_M2PATH variable, which points to the directory containing the sas2py.py interface file, is already set by default following deployment.

Open Source from SAS Software

SAS Institute, headquartered in Cary, North Carolina, maintains an active open-source presence on GitHub under the handle sassoftware, with 150 public repositories and 39 listed members. The organisation develops and shares a range of tools centred on its SAS and SAS Viya platforms, including a Visual Studio Code extension for SAS language support, Python packages for interfacing with SAS systems and Viya integration, and infrastructure tooling built with Terraform and Ansible for deploying Viya on cloud environments such as Microsoft Azure. Additional projects cover areas such as AI agent development, monitoring and logging for Kubernetes deployments, Helm charts for marketing solutions and a Model Context Protocol server for executing SAS code remotely. The most frequently used programming languages across the repositories are Python, SAS, Jupyter Notebook, Shell and Go.

MCP servers extend agent mode functionality in VS Code by providing specialised tools for various tasks including connecting to databases, invoking APIs and performing automated operations. The available servers span multiple categories, with developer tools offering integration with platforms like GitHub for repository management, Figma for design extraction, Playwright for browser automation and Sentry for error analysis. Productivity-focused servers connect to project management platforms such as Notion, Linear, Asana and Atlassian services, whilst also providing workflow automation through Zapier and task breakdown capabilities. Data and analytics servers enable interaction with databases like DuckDB, Neon Postgres and MongoDB, alongside analytics platforms including PostHog and Microsoft Clarity for gathering user behaviour insights. Business service integrations cover payment processing through Stripe, PayPal and Square, customer support via Intercom and site building through Wix and Webflow. Cloud and infrastructure servers provide management capabilities for Azure resources, DevOps operations, Terraform infrastructure as code and Convex backend services, creating a comprehensive ecosystem for development and business operations within the VS Code environment.

JuliaCon2021 DataFrames.jl Tutorial

A tutorial on DataFrames.jl, created for JuliaCon2021 by Bogumił Kamiński, is hosted as a public repository on GitHub and has since been updated to work with Julia 1.9.0 and DataFrames.jl 1.5.0. To run it, users need to clone the repository locally, start Julia using the julia --project command and then instantiate the required packages, which include CSV, CategoricalArrays, Chain, GLM, IJulia, Plots and StatsPlots. Once those steps are complete, the tutorial is launched via Jupyter Notebook by opening the Tutorial.ipynb file. Additional learning material on DataFrames.jl can be found in its official documentation and on the author's blog.

The SAS Developer Portal serves as a comprehensive platform enabling developers to integrate SAS artificial intelligence and analytics capabilities with open source technologies. The portal provides extensive REST API documentation across multiple categories including automated machine learning, cloud analytic services, data management, fraud detection, healthcare applications, and visualisation tools. Developers can access resources for integrating SAS with various programming languages such as Python, R, Lua, and Java, whilst also utilising software development kits and frameworks to embed SAS insights into custom applications and dashboards. The platform supports SAS Viya programming and CAS actions for data processing and analytics operations, complemented by tools like the Visual Studio Code extension. Key solutions include Customer Intelligence 360 for marketing automation, SAS Viya Workbench for cloud-based analytics, and SAS 360 Match for advertising delivery, all underpinned by trustworthy AI principles emphasising transparency and accountability. The portal connects developers through community forums and GitHub repositories containing code examples and collaborative resources.