Written by

Developer at Intersystems
Article Thomas Mazur · 17 hr ago (Edited) 2m read

Frogs, Chickens, AI, and VS Code

I am going to share some of the things I find very useful about VS Code, after many years developing in Studio.

The first and obvious thing to me is the fact that it does not crash in the middle of development when it disconnects from the network for any reason, forcing you to restart Studio and deal with the risk of losing valuable code. I also find VS Code a lot quicker than Studio, and the user interface far better with so many ways to customise it to your liking.

In France we use farm animal names for our dev and test environments. It is quite easy to customise the colours of VS Code to match the colour of that animal. I have for instance green for  "Grenouille" (frog), orange for "Poule" (chicken) and brown for "Boeuf" (ox). Those are our branch and main development environments. Setting up the colour can be done using the extension Peacock, and configured in the workspace file : ctrl + shift + P -> Open workspace settings (json). That setup makes it possible to recognise the environment that you are working on when you have multiple instances of VS Code running. 

I believe it's also possible to do a lot more customisation, but I have not got around to doing it.

Another pretty cool thing about these workspace files is that you can change the default folder of the code you want to work in. It's particularly useful when there's a lot of code you're not interested in. A practical example can be to get direct access to a folder, for instance you can create a workspace file for a particular environment that points directly to the Region/FRXX/Convert. 


Now when you click the workspace link (.code-workspace) it takes you directly to the Region/FRXX/Convert folder and opens it as the root folder. It makes it a lot cleaner. 

(ps: I also changed the icons of the .code-workspace files to faces of the animals)

I think I should also mention the AI (Artificial Intelligence) integration via GitHub Copilot, which is a little extension fully integrated into VS Code. Once the connection is made to your GitHub account, you get direct access to AI agents via the chat window of VS Code, such as Claude Sonnet, ChatGPT, and other models if your licence allows it. You can interface with the AI in "Agent mode", where it will be able to read the code base, talk to you and edit code directly, "Ask mode" where you ask questions, it reads the code base, and replies to you, or "Plan mode" where it can deal with complex tasks and gives you a full plan on how to implement them (I have not used that mode yet but it looks promising). You also have constant access to your past chats, and it's possible to go back to any of those chats at any time.

I now use Copilot regularly and find it extremely useful to understand the code base, which can be sometimes quite complex. It can quickly make connections when the human brain would take a long time. It's also good for writing the technical notes if it has all the context of what was done.

Comments

Pietro Di Leo · 16 hr ago

Hi Thomas, really interesting article. As a developer myself, I also find VSCode to be much better than Studio. I have a question about how you use the GitHub Copilot extension, are you able to use the agents while working in server-side mode? Or does it only work with local code in client-side mode? I’m asking because I use Cursor (a fork of VSCode) a lot, but it can only be used in client-side mode since it isn’t compatible with virtual file systems. I’ve noticed that Copilot sometimes is able to modify server-side classes when they’re open (I believe this happens because it acts on the buffer of classes opened in the explorer, while Cursor focuses on the local file) but it is still not able to perform comprehensive modification like he can on local files.

0
Aziz Cotrim · 16 hr ago

Great article! It clearly highlights the practical advantages of using VS Code over older tools, especially with customization and AI integration. I believe this is a really valuable learning resource for developers who are just starting out, as it shows simple but effective ways to improve productivity and better understand complex codebases.

0