Playwright Agents in VS Code

I started looking at the latest version of Playwright late last night. The Agents add-in for VS Code is amazing.  I can't stop improving my code, my tests, and my automation.  It is highly addictive.

Playwright 1.56.1 includes the new Playwright CLI, which has the test agents as shown in VS Code above:

Node: v22.21.0
npx: 11.6.2
Playwright test package: ^1.56.1

Medallion Architecture in Fabric High Level Organisation Design Pattern

Microsoft Fabric is excellent!  We do still need to follow good practices we have been using for years, such as making data accessible and secure.   Possibly the most used architecture for Big Data is the Medallion Architecture pattern, where data is ingested normally in a fairly raw format into the bronze layer, then transformed into more meaningful and usable information. Lastly, the gold layer exposes data relationally using semantic models to reporting tools.

Overview: This document outlines my attempt to organise enterprise data into MS Fabric using a Medallion Architecture based on Fabric Workspaces.  Shortcuts are better than imported data, but it does depend on factors such as what the data source is, what data we need, how up-to-date the data is and performance requirements from the systems involved.

The reports and semantic models can get data from other workspaces at any of the medallion layers.  This architecture lends itself well to using the new Direct Lake Query mode.

Summary of a Design used by a Large Enterprise:

Medallion architecture using Fabric Workspaces.

Microsoft Fabric High-level architecture

Overview: Microsoft Fabric is an end-to-end analytics platform that unifies data movement, storage, processing, and visualisation. It integrates multiple services into a single SaaS experience, enabling organisations to manage their entire data lifecycle in one place.  One Lake is at the core of MS Fabric.

Image 1. One page High-Level Architecture of MS Fabric. 

European Fabric Conference in Vienna Sept 2025 takeways

FabConEurope25 was terrific in Vienna last week.  Great opportunity to meet Fabric and data experts, speak to the product teams and experts, and the presentations were fantastic.  The hardest part was deciding which session to attend as there are so many competing at the same time.  

My big takeaways:

• Fabric SQL is excellent.  The HA, managed service, redundancy, and shipping logs ensure that OneLake is in near real-time.  Fabric SQL supports new native geospatial types.  SQL has temporal tables (old news), but row, column and object-level (incl. table) security is part of OneLake.   There are a couple of things security reviewers will query, but they are addressed.
• Fabric Data Agent is interesting.  Connect to your SQL relational data and work with it.
• User-defined functions (UDF), including Translytical (write-back), HTTP in or out, wrap stored procedures, notebooks,.... - amazing.
• OneLake security is complex but can be understood, especially with containers/layers, such as Tenant, Workspace, Item, and Data.  There is more needed, but it's miles ahead of anything else, and Graph is the magic, so it will only continue to improve. - amazing, but understand security.  Embrace Entra and OAuth; use keys only as a last resort.
• Snowflake is our friend.  Parquet is fantastic, and Snowflake, including Iceberg, play well together with MS Fabric.  There are new versions of Delta Parquet on the way (and this will even make Fabric stronger, supporting both existing and the latest formats).
• Mirroring and shortcuts - don't ETL unless you need to shortcut, then mirror, then ETL.
• Use workspaces to build out simple medallion architectures.
• AI Search/Vector Search and SQL are crazy powerful.
• New Map functionality has arrived and is arriving on Fabric.  Org Apps for Maps is going to be helpful in the map space.  pmtiles are native... (if you know you know)
• Dataverse is great with Fabric and shortcuts, as I learned from Scott Sewell at an earlier conference.  Onelake coupled with Dataverse, is massively underutilised by most orgs, 
• Power BI also features new Mapping and reporting capabilities related to geospatial data.
• Other storage: CosmosDB (it has its place, but suddenly, with shortcuts, the biggest issue of cost can be massively reduced with the right design decisions).  Postgres is becoming a 1st class citizen, which is excellent on multiple levels. The CDC stuff is fantastic already.
• RTI on Fabric is going to revolutionise Open Telemetry and AI, networking through the OSI model, application testing, digital twins, and live monitoring,....  I already knew this, but it keeps getting better.  EventHub and notebooks are my new best friends.  IoT is the future; we all knew this, but now with Fabric, it will be much easier to implement safely and get early value.
• Direct Lake is a game changer for Power BI - not new, but it just keeps getting better and better thanks to MS Graph.
• Manage Private Endpoint as improved and should be part of all companies' governance.
• Purview... It's excellent and solves/simplifies DLP, governance and permissions.  I'm out of my depth on Fabric Purview and governance, and I know way more than most people on DLP and governance. Hire one of those key folks from Microsoft here.  
• Warehouse lineage of data is so helpful.  
• We need to understand Fabric Digital Twins, as it is likely to be a competitor or a solution we offer and integrate. 
• Parquet is brilliant and fundamentally is why AI is so successful.
• Powerful stuff in RDF for modelling domains - this is going to be a business in itself.  I'm clueless here, but I won't be in a few weeks.

Now the arr..

• Pricing and capacity are not transparent.  Watch out for the unexpected monster bill!  Saying that the monitoring and controls are in place, but switching off my tenant doesn't sit well with me if workloads aren't correctly set out. Resource governance at the workspace level will help fix the situation or design around it, but it will be more expensive.
• Workspace resource reservation does not exist yet; however, it can be managed using multiple fabric tenants. Distribution will be significantly improved for cost control with Workspace resource management.
• Licensing needs proper thought for an enterprise, including ours.  Reserve Fabric is 40% cheaper, and it cannot be suspended, so use the reserved fabric just as you would for most Azure Services.  Good design results in much lower cost with Workloads.  Once again, those who genuinely understand know my pain with the workload costs.
• Vendors and partners are too far behind (probably due to the pace of innovation)

Microsoft Fabric is brilliant; it is all under one simple managed autoscaling umbrella.  It integrates and plays nicely with other solutions, has excellent access to Microsoft storage, and is compatible with most of the others.  Many companies will move onto Fabric or increase their usage in the short term, as it is clearly the leader in multiple Gartner segments, all under one hood.  AI will continue to help drive its adoption by enterprises.

What is GIS?

GIS stands for Geographic Information Systems, which are tools and techniques for capturing, managing, storing, processing, and analysing spatial data. It is part of the broader geospatial technology ecosystem, which also includes drones, remote sensing, and GPS.

Geospatial data (Raw)

Definition: Any data that includes a geographic component, describing the location and attributes of features on Earth, contains raw information, like points, lines, and polygons, that has a real-world location associated with it.

Examples: A GPS position of a car or the address of a customer.

GIS data (Organised)

Definition: Geospatial data that is structured, stored, and analysed using Geographic Information System software.

Examples: include a digital map of roads created from GPS data or layers of data showing flood risk areas.

Summary: Geospatial data is the foundation: It is the raw material for all things spatial. GIS is a toolset that may include tools like ArcGIS from Esri.

Other:

In the AEC space, building and Asset management rely heavily on GIS within BIM.

ArcGIS is the industry leader in GIS tooling, and comes in three versions: 

• Desktop (ArcPro, Arc Toolbox, ArcCatelog),
• Server (), 
• SaaS ArcGIS Online (AGOL).

What WGS84 and GeoJSON Mean?  

These are the most common formats for storing position (WGS84) and shape data with coordinates (GeoJSON) 

WGS84 (World Geodetic System 1984) is the standard geographic coordinate reference system used globally. It represents positions on Earth using latitude and longitude in decimal degrees.

GeoJSON is a widely used format for encoding geographic data structures in JSON. According to RFC 7946, all GeoJSON coordinates must use WGS84 (EPSG:4326).

GitHub Copilot with Claude Sonnet 4 is amazing, and GPT 5 is even better

I couldn't sleep, so I decided to build a Pulumi C# application that uses an existing MCP Server. My forms will utilise the client to allow me to access my Azure subscriptions and resources - wow.  Build a really cool tool quickly - Claude Sonnet 4 is once again significantly better than GPT-4.1 for programming with GitHub Copilot.

Update Sept 2025: I'm now using GPT-5 over Claude Sonnet with GitHub Copilot when programming in VS Code.  Both feel about the same quality to me.

GitHub have this for comparing AI models for GHCP, which is very useful.

I am using GPT-5-Codex, which "is a version of GPT-5 optimised for agentic coding in Codex".

I am also really liking GitHub Copilot code review

Anthropic's Claud 4.5 is also excellent..

AI for developers and Architects

The cost of prototypes is unbelievably low using AI. 

Rapidly creating a prototype, especially with new or less well-known technology, is where I derive significant benefits from AI.

How to build application prototypes?

1. Write /reverse prompt/Adjust instructions into md file
2. Agentic AI (specialising in Doc Extraction) to extract and refine from md file
3. Run using IDE-based copilot (VS Code with GitHub Copilot) (AmazonQ) (Cursor, Windsurf, Steamlit) 

Thoughts: Developers are adjusting to using Ai to support software solutions.  The developer role will continue the trend of making technical implementation more accessible, allowing knowledgeable IT engineers or domain experts to build faster and better than citizen/amateur developers.  Ai assists in complex decisions!  

What needs to improve?

• Knowledge is key.  AI needs to have narrow expertise at the right time. i.e. only domain knowledge, not influenced by other data.  Quality of input data used to train.  Allows for dynamic reasoning.
• Session/long-term contact agreement/understanding to improve the understanding between your IDE and me.  Remember how I prompt and provide feedback on how I digest information.  Context between the human developer and Ai is Paramount.
• Control of IDE integration with coding copilots, clear return to the user developer to make better decisions.  Context is Paramount.
• Governance & Data (Connectors, API's, code complex processes (MCP maybe), quality of data).

Retrieval Augmentation Generate (RAG)

Model Context Protocol (MCP)

MCP is a protocol (created by Anthropic) that allows an MCP client to connect to an MCP server, which in turn has specialist knowledge that the MCP server will expose. Authentication uses OAuth to secure access.

My Applications/Agents use the MCP to ask the MCP Server, 'What can you do?' so that they are aware of how to use the MCP Server.

The MCP Server, when built, informs the client of its capabilities and then performs actions such as updates using an API.

Summary: Use MCP to allow the client to talk to other resources/tools

Agents-to-agent (A2A) 

A2A allows agents to work together.  So two agents can leverage each other; the other agent solves the issue and returns the answer for the first agent to use.  Whereas MCP allows any agent to speak to a source.  Agents complete a task and give it back to the calling agent. 

 

Summary: Use A2A to talk to specialised Agents, and the agent returns the calling agent's answers.