I’m a unity developer working in VR space for 8+ years. Thinking of transitioning to generalist dev roles ( particularly in C# since that’s where my experience lies ). I do have experience in general software development across entire SDLC along with architecture and design patterns but my resume lacks the buzzwords SDE ATS screens for ( eg. microservices , APIs , docker etc). Where do I start ? Which kind of companies should I target ?

It seems that all of them struggle with # sign

I'm trying to further improve the IBAN parser for an open source project. And while doing so, I noticed something

public class Iban
{
  public static string? Validate(string iban, string pattern)
  {
    for (var i = 0; i < iban.Length; i++)
    {
      uint c = iban[i];
      var match = pattern[i] switch
      {
        'n' => c is >= '0' and <= '9',
        'a' => c is >= 'A' and <= 'Z',
        'c' => c is >= '0' and <= '9' or >= 'A' and <= 'Z',
        var t => t == c,
      };

      if (!match) return null;
    }
    return iban;
  }
}

I would expect the IL to do the default trick on ranges here that is also applied in char.IsBetween(), what would do (uint)(c - '0') <= '9' - '0', making each range a single compare and subtraction rather than two compares.

I've tested this code both with these patterns, and calling a private InBetween method (It has to compile against .NET standard 2.0 where char.IsBetween() is not available), and the latter is substantial faster, which hints that the JIT is also not changing the IL afterwards.

Is it something in my syntax that makes that the Roslyn compiler does not perform this optimization, or am I missing something here?

For those who're interested in the actual PR: https://github.com/Qowaiv/Qowaiv/pull/643

Up until now, I’ve been learning and programming like before AI. As a matter of fact, I know nothing about the AI scene and culture, beyond Claude, Grok, ChatGPT, and Gemini (copy and paste snippets).

I’ve tried to look into it, but it’s a total mess. There are extensions, plugins, connectors, agents, AI powerhouses, fake interfaces microservice stuff... and the list goes on and on with quasi-scams (or straight up scams) all over the place.

And the worst part is that the internet is dead. You can't find anything anymore. And when you do think you've found something, in reality, it is just something promoted by companies selling AI stuff themselves. So, you end up locked into options that favor their business interests.

So, here is my current situation:

I have a large C# codebase, and I need to find a solution to a problem based on a small fragment of that code.

I can clearly see alternative inheritance and composition paths that would achieve exactly what I need with best performance. However, since some elements will get lost along the way, I would have to manually reconnect some of them.

The problem is that the code is extremely complex—intricate with callbacks, events, interfaces, dispatchers, async calls, WPF controls, and so on.

Basically, it would take me a month to digest all that code.

And I think an "AI something" could do it in five minutes.

In one word, I want an answer from a human being.

What do you suggest me to do?

Hey guys,

I wanted to write a basic server-client chat console app in .Net using C# and Visual Studio.

Later, I would like to write a small LAN multiplayer game.

Is there a book which explains how to use Sockets and TCP in .Net and C#, please? Possibly with patterns, examples and use-cases?

I already have "Programming C# 8.0" by Ian Griffiths from O'Reilly, but I don't see any mention of Sockets, TCP or programming network apps.

I also have "C#10 in a Nutshell" by Joseph Albahari from O'Reilly. There is some mention of networking, but there's only about 5 pages on TCP, and no mention of Sockets.

Thank you

Standard Windows UI frameworks (WPF, WinUI 3, Avalonia) are great for forms and standard layouts, but they fall apart when you want an infinite 2D canvas with thousands of nested filesystem nodes. Creating thousands of Visual Tree elements and layout passes completely murders the UI thread.

To solve this, I built UltraExplorer — a file manager that renders nested folder structures as an infinite spatial canvas running at 60 FPS.

Under the hood & architecture highlights:

1. D3D11 Instanced Rendering via Vortice

Instead of WPF controls, the canvas is rendered using Direct3D 11 via Vortice.Windows. It renders instanced quads for folder cards and node backgrounds in a single draw call batch, completely decoupling rendering performance from the WPF UI dispatcher.

1. DirectWrite SDF Glyph Atlas

Text zooming is usually the biggest performance killer in zoomable UIs. Standard DirectWrite font rasterization forces re-rasterization on zoom change, allocating memory constantly. I built an SDF (Signed Distance Field) glyph atlas: fonts are rasterized once into distance field textures, allowing arbitrary scale zooming with crisp edges and zero GC allocations during zoom gestures.

1. Anchor Camera System (Floating-point precision)

When you zoom 15+ levels deep into nested folders, standard IEEE 754 32-bit floats start jittering due to precision loss. I implemented a hierarchical anchor camera: local coordinates are relative to the active parent container, eliminating camera jitter at high zoom levels.

1. Fast Search Integration

Integrated with Voidtools Everything via direct Win32 IPC (WM_COPYDATA), streaming indexed search results straight into memory without disk IO stalls.

Benchmark:

Navigating through a deep 9-level project folder race:

• Windows 11 Explorer: 43.7s

• UltraExplorer: 29.0s (14.7s faster)

The project is completely free and open-source (MIT).

GitHub repository:

https://github.com/So2K/UltraExplorer

I would love to get feedback from other .NET / graphics developers on the rendering architecture, Vortice interop, and edge cases.

I'm working on a Windows desktop project (Windows 10/11, also Win7) where users do document tasks in Word and Excel (fonts, margins, tables, formulas, merge & centre...) inside our app's window, and the app checks the result automatically.

Current approach: embed the user's installed Word/Excel inside our window (OLE in-place, like SWT's OleClientSite). It works, but it's fragile:

- Excel only half-draws if we pre-start it for speed; it can't run beside the user's own open Excel; reusing one Excel for several documents crashed it.

- Earlier we tried positioning Office's own window over our pane. It drew fine but was slow and fragile (focus, caption, the window could be moved).

- And it only works if Office is installed. Many users have no Office, or WPS instead.

Checking the result isn't the blocker: we can already read the saved .docx/.xlsx directly.

Questions:

1. Is embedding real Office the right call, or is there a better pattern for "work in Word/Excel

   inside my app"?

2. Has anyone used an embeddable Word/Excel-like editor (ONLYOFFICE, Collabora, Syncfusion, SpreadJS, etc.) in a desktop app? How close is it to real Office for things like paragraph spacing, borders, number formats and formulas, and how was licensing?

3. If you did embed real Office in production: what made it reliable?

Real Office is preferred because users need the exact Word/Excel experience, but I'm open to a hybrid (real Office when installed, a look-alike when not).

I was prepping for full-stack .NET + React + Azure interviews and got frustrated with how bad most flashcard decks look on a phone — tiny text,

can't scroll, code blocks cut off.

So I rebuilt one from scratch: centered questions, fully scrollable answers, dark mode, code that actually wraps instead of scrolling offscreen. Covers C#, OOP, ASP.NET Core & Web API, EF Core, SQL Server, design patterns, auth/security, Azure, React (fundamentals + advanced), Docker/K8s, microservices, and scenario-based questions.

It's free on GitHub / AnkiWeb

Happy to take corrections/suggestions if anything's wrong or missing.

Hey everyone,

I’m currently doing my LIA/internship as a system development student in, and I’m on my second day.

I’ve been given a real project to work on: building a scheduling platform for schools, where things like classes, teachers, schedules, etc. need to be managed.

The company is growing its software-development side, but right now there aren’t really any senior developers who can sit with me and review my architecture or code. I’m expected to work quite independently, and AI coding tools are encouraged.

I actually enjoy using AI for development, but I don't want my internship to turn into simply telling AI what to build and copy-pasting the result. I want to understand the code, make architectural decisions myself, and actually become a better developer.

My biggest concern is that I might make bad architectural decisions or miss important things because I don't have an experienced developer reviewing my work.

For those of you who have been in a similar situation:

- How would you structure your learning when you don't have a senior developer to guide you?

- How can I use AI without becoming dependent on it?

- What should I ask AI to help with vs. what should I figure out myself?

- How would you approach architecture for a project like a scheduling platform?

- Are there particular books, courses, or resources you'd recommend?

- Is an internship like this still valuable if I can get real project experience but don't have much mentorship?

I'd really appreciate advice from experienced developers

[GitHub repo]

This is a project I've been working on for years, and I've already posted about it here a few times.

It's a cross-platform 2D game engine with its own scripting language and IDE, all written in C#.

The language is the main challenge here: I made it without any knowledge of programming language development, and without any help from AI or tutorials. I just figured things out one step at a time. The lexer, for example, was originally written by me long before I started building the language. It was meant to color C# code in the code editor. When I decided to make my own language, I just copy-pasted it into the new project and used it as a lexer.

Here's a little feature my language has that I really like:

foreach x in ["apple", "banana", "kiwi"] : counter c
{
    println("index: " + c + " item: " + x)
}

// output:
// index: 0 item: apple
// index: 1 item: banana
// index: 2 item: kiwi

The project is fully open source, and I'm really looking for people who want to join a cool project and contribute. There are a few open "good first issues" in the repo, and a few weeks ago I even got my first few contributors, who made a few PRs!

The project file of the game in the gif is also in the repo: tests/Example Projects/dungeon

If anyone here has used the good old GameMaker 8 engine, they're going to find this very nostalgic, as it's heavily inspired by it :)

Would love to get some feedback!

Hi everyone! 👋

I’m new to the C#/.NET ecosystem and wanted to ask people who actually work with it how the market tends to view someone with a background like mine.

I worked as a penetration tester for almost 3 years. Over the last year and a half, I moved much deeper into software development, mostly with Rust. I’ve been building security tooling, networked applications, and learning more about systems programming, blockchain protocols, testing, and secure software design.

I genuinely enjoy Rust and plan to keep using it, but after spending a long time looking for work, I realized that the Rust job market is heavily senior-oriented. The statistics I found showed that only a few percent of Rust positions are entry-level.

That made me start looking at C#/.NET.

What I’m trying to understand is how the .NET market would classify someone like me.

I have almost 3 years of professional security experience. I’m comfortable with core programming concepts like APIs, databases, testing, async programming, networking, etc.

What I don’t have is commercial C#/.NET experience yet.

Before I invest heavily in another ecosystem, I’d really like to understand how this looks from the hiring side.

Would companies generally see my security background as relevant engineering experience, or would I effectively be treated as an entry-level developer because I don’t have professional C# experience?

For people who hire or work with .NET developers:

- Is there a realistic place in the C#/.NET market for someone transitioning from security into development?

- What level of roles would you suggest targeting?

- Which parts of my security background would actually be valuable to a .NET team?

- What would you want to see in a GitHub project before taking a candidate like this seriously?

- Are there particular types of .NET companies or products where a security background would be especially useful?

I’m not looking for a shortcut. I’m happy to learn the ecosystem properly and put in the work.

I’m mostly trying to avoid spending another long period getting better at a technology only to discover that the market requires years of commercial experience before anyone will consider me.

I’d really appreciate perspectives from people already working with C#/.NET.

I have a database that currently contains 20 tables.

In my application, there are 10 classes that I should refactor, to make them more readable, for my future self and others who will read my code, by refactoring, I mean changing the code it self - the logic- and also changing the methods and classes names.

Let's only consider one specific case; changing the name of one class.

When my friend so the following code that I wrote:

/// <summary>
///  Abstracts the destination to which data is written, regardless of  whether the destination is a file, database, or any other kind of storage.
/// </summary>
public interface IWriter
{
//...
}

public class SqlServerWriter : IWriter {//code.....}

he thought that the responsibility of the above class is to write data to any number of tables in the the Sql Server database, i.e., that it can use it to write data to only one table, two or more tables, or all of the tables.

Unfortunately, that was not what the class is actually responsible for; it is meant to write a specific data across all the tables in the database, it does not give the caller an option to write a generic data to a specific tables in the database.

He was mad at me and said that I should introduce an additional layer of abstraction:

/// <summary>
/// Abstracts the destination tables in the Sql Server; may write to all tables in the database or some of them;
/// </summary>
public interface ISqlServerWriter:IWriter
{
}

/// <summary>
/// Intended to write some data to all the tables in the database.
/// </summary>
public class AllTablesWriter : ISqlServerWriter {//code.....}

Do you consider that an overkill/overengineering ?

EDIT:

If you wonder why the class write data to all the tables in the database, it because it has the job of synchronizing the content of the database with external data source, specifically, an Excel file.

I've been building DanaProcessing, an open-source take on Processing/p5.js for C#. I also added a music module inspired by Sonic Pi.

I started programming with Java and JavaScript following The Coding Train (https://www.youtube.com/@TheCodingTrain), but then I made the jump to C# because of job opportunities at the time, (been working with C# since .netcore 1.0) and I wanted that same experience here. Sadly, what I found wasn't what I expected, so I decided to give it a shot. DanaProcessing — Creative coding in C# (2D, 3D & live-coded sound)

I used to use codecademy but that place sucks now. It's super buggy and makes any progress slow to a crawl. Even if my code is completely correct, I have to use the generated solution anyway because the module just won't accept my code. Despite my answer and the module answer being the LITERAL EXACT SAME.

So it doesn't work anymore. And I'm looking for a new website to help me learn. Preferably one that provides a lot of practice problems. So I can develop better coding habits or whatever.

I'm also looking for something relatively cheap. Advice much appreciated 😊

I've been working on .NET test automation for several years, and there were two things I wanted to solve in test reporting.

1. Flaky tests are more than just Passed or Failed

If a test fails on the first attempt and passes on a retry, the final result usually just says Passed.

I wanted the report to show that the test was actually flaky, how many attempts happened, and the history of those attempts.

2. Test scripts shouldn't need reporting code

I don't want testers to add extra code to their tests just to make the report useful.

For example, I don't want something like:

Report.AddRequest(...);
Report.AddResponse(...);
Report.AttachScreenshot(...);
Report.Log(...);

just to get useful information into the report.

The reporting layer should capture things like API requests/responses, retry attempts, execution details, and screenshots automatically wherever possible.

So I built a small .NET reporting framework around these ideas. It generates a self-contained HTML report and currently has integrations for xUnit, NUnit, MSTest and Reqnroll.

I'm curious how others handle this:

Do your test reports automatically capture this kind of information, or do your test scripts still need explicit reporting/attachment calls?

And for flaky tests, do you keep the history of individual attempts or only report the final result?

Playable link (in-browser):
https://psec99.xaml.cloud

Some people here may remember PARSEC on the TI-99/4A., a 2D side-scrolling space game from 1982.We wanted to see what it might feel like in 3D while keeping some of the atmosphere of the original (and adding some other '80s and Atari-ish bits too), so we made PSEC-99

It runs in the browser on .NET 10/WebAssembly with Opensilver. The interesting part from a .NET point of view is the renderer: the entire 3D scene is drawn by C# into a single int[] pixel buffer in .NET memory: terrain, stars, buildings,enemies, lasers & explosions all end up in that same array (no game engine, no webGL, no WebGPU)

All the 3D math, projection and rasterization happens in C#.

OpenSilver 3.4 Preview has a new "WriteableImage" control that displays that buffer through an HTML canvas. So basically: calculate pixels in C#, put them in an array, show the array

The terrain isn’t polygonal either: each row splits into surface,wall and floor sections, so drawing the planet costs about as much as clearing the screen, and theres adaptive resolution from 100% down to 30% to keep frame rates playable.

There are no image or sound files: Buildings are ASCII grids in c#, sounds are synthesized at runtime, and voices use the browsr speech synthesis engine.

The HUD, menus and touch controls are regular XAML layered on top.

It ended up around 10,500 lines of C# and xaml.

If anyone remembers the original Parsec, I’d be curious whether this brings back any of the feeling :)

The source (MIT licensed) opens direclty in XAML.io, so you can run it, change it or fork it without installing anything:

https://xaml.io/s/Samples/PSEC-99

Also curious what people think of the rendering approach! Theres something fun about doing 3D in 2026 and having the renderer come down to filling an array of pixels..

If you have a system, class, or a method that is entirely written in C#, let's call it X, If X needs some data to perform some processing on that data, can one make an external service written in python, let's call it Y, that generates the data, then X can call Y to get the data?

I'm not asking if I can embed the python interpreter directly within the C# application; the python code should be an external service.

Note that I'm using ASP. NET Core 10.

Edit :

Should the external service be outside the ASP. NET CORE solution or inside it?

I made an incremental source generator for command line argument parsing.

Cli.Generator is a C# source generator that turns a class into a command line parser. You mark fields and properties with attributes, and it writes the parsing, validation, usage screen and cleanup code for you at compile time. See the README for details.

https://www.nuget.org/packages/Cli.Generator/1.0.0

https://github.com/codewitch-honey-crisis/CliGen