Don't worry. There will be a place for everyone.
Learn AI and use it as your leverage to improve your work.
Learn AI and use it as your leverage to improve your work.
Generate viral LinkedIn posts in your style for free.
Generate viral LinkedIn posts in your style for free.
Dr Milan Milanović
These are the best posts from Dr Milan Milanović.
10 viral posts with 3,992 likes, 349 comments, and 253 shares.
9 image posts, 0 carousel posts, 0 video posts, 0 text posts.
👉 Go deeper on Dr Milan Milanović's LinkedIn with the ContentIn Chrome extension 👈
Best Posts by Dr Milan Milanović on LinkedIn
𝗪𝗵𝗮𝘁 𝗔𝗿𝗲 𝘁𝗵𝗲 𝗚𝗿𝗲𝗲𝗻𝗲𝘀𝘁 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀?
When we choose programming languages, we usually decide based on their syntax, performance, or even learning curve.
Yet, a few years ago a group of Portuguese researchers investigated the energy consumption of 𝟮𝟳 𝗺𝗼𝘀𝘁 𝗽𝗼𝗽𝘂𝗹𝗮𝗿 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀, 𝗺𝗲𝗮𝘀𝘂𝗿𝗶𝗻𝗴 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝘁𝗶𝗺𝗲, 𝗲𝗻𝗲𝗿𝗴𝘆 𝗰𝗼𝗻𝘀𝘂𝗺𝗽𝘁𝗶𝗼𝗻, 𝗮𝗻𝗱 𝗽𝗲𝗮𝗸 𝗺𝗲𝗺𝗼𝗿𝘆 𝘂𝘀𝗲.
The results reveal trade-offs: faster code isn’t always more energy-efficient, and memory use strongly affects energy costs.
These insights can guide engineers in choosing languages when energy efficiency matters.
The results are the following:
𝗖 𝗶𝘀 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲, 𝘄𝗵𝗶𝗹𝗲 𝗣𝘆𝘁𝗵𝗼𝗻 𝗮𝗻𝗱 𝗣𝗲𝗿𝗹 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗹𝗲𝗮𝘀𝘁 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝗮𝗹 𝗳𝗿𝗶𝗲𝗻𝗱𝗹𝘆 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲.
When we choose programming languages, we usually decide based on their syntax, performance, or even learning curve.
Yet, a few years ago a group of Portuguese researchers investigated the energy consumption of 𝟮𝟳 𝗺𝗼𝘀𝘁 𝗽𝗼𝗽𝘂𝗹𝗮𝗿 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀, 𝗺𝗲𝗮𝘀𝘂𝗿𝗶𝗻𝗴 𝗲𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝘁𝗶𝗺𝗲, 𝗲𝗻𝗲𝗿𝗴𝘆 𝗰𝗼𝗻𝘀𝘂𝗺𝗽𝘁𝗶𝗼𝗻, 𝗮𝗻𝗱 𝗽𝗲𝗮𝗸 𝗺𝗲𝗺𝗼𝗿𝘆 𝘂𝘀𝗲.
The results reveal trade-offs: faster code isn’t always more energy-efficient, and memory use strongly affects energy costs.
These insights can guide engineers in choosing languages when energy efficiency matters.
The results are the following:
𝗖 𝗶𝘀 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲, 𝘄𝗵𝗶𝗹𝗲 𝗣𝘆𝘁𝗵𝗼𝗻 𝗮𝗻𝗱 𝗣𝗲𝗿𝗹 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗹𝗲𝗮𝘀𝘁 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝗮𝗹 𝗳𝗿𝗶𝗲𝗻𝗱𝗹𝘆 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲.
𝗔𝗪𝗦 𝘂𝘀-𝗲𝗮𝘀𝘁-𝟭 𝘄𝗮𝘀 𝗱𝗼𝘄𝗻 𝘆𝗲𝘀𝘁𝗲𝗿𝗱𝗮𝘆
It took down many websites, from Medium and Substack to Box and others.
Here is what happened:
1. It started with a DNS failure, which broke name resolution for the DynamoDB API in US-EAST-1.
2. With DynamoDB unreachable, many AWS services built on it failed.
3. Ripple effects: EC2 launches, Lambda jobs, logins/admin consoles stalled; many consumer apps went offline.
4. AWS mitigated the DNS issue; recovery progressed while backlogs drained across services.
𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆: a single dependency in a core region can ripple through half the internet.
👉 Details: https://lnkd.in/dKTX2H4r
Note: This image is a joke :)
It took down many websites, from Medium and Substack to Box and others.
Here is what happened:
1. It started with a DNS failure, which broke name resolution for the DynamoDB API in US-EAST-1.
2. With DynamoDB unreachable, many AWS services built on it failed.
3. Ripple effects: EC2 launches, Lambda jobs, logins/admin consoles stalled; many consumer apps went offline.
4. AWS mitigated the DNS issue; recovery progressed while backlogs drained across services.
𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆: a single dependency in a core region can ripple through half the internet.
👉 Details: https://lnkd.in/dKTX2H4r
Note: This image is a joke :)
𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗷𝘂𝘀𝘁 𝗿𝗲𝗹𝗲𝗮𝘀𝗲𝗱 𝗔𝗴𝗲𝗻𝘁 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸
It merges Semantic Kernel's production tooling with AutoGen's multi-agent patterns
This fixes the prototype-to-production gap that's been killing agent projects
𝗪𝗵𝗮𝘁 𝘆𝗼𝘂 𝗴𝗲𝘁
Multi-agent orchestration that actually works. Sequential, concurrent, group chat, and handoffs. Pick the pattern that fits. Graph-based, so you can see what's happening instead of guessing.
Python and .NET support.
OpenTelemetry baked in. Trace every agent action, tool call, and decision. When you're running agents in parallel, this isn't optional.
Checkpointing and resume. Save state, replay workflows, and recover from failures. Makes debugging and experimentation possible.
Human-in-the-loop approvals. Mark any tool as requiring sign-off. The agent waits; you approve or deny, and it continues.
Open standards: MCP for tools, A2A for agent-to-agent, OpenAPI for REST endpoints. Your agents aren't locked to Microsoft.
𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀
Semantic Kernel users had stability but limited orchestration. AutoGen users had flexibility but no durability. Agent Framework gives you both.
KPMG is using it for audit automation. BMW for real-time telemetry analysis. Commerzbank for customer support. Production systems in regulated industries.
𝗠𝗶𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝘀𝗶𝗺𝗽𝗹𝗲
Semantic Kernel: replace Kernel with Agent. Tools instead of plugins.
AutoGen: AssistantAgent becomes ChatAgent. Event-driven runtime becomes typed Workflows.
Microsoft is shifting focus here. Semantic Kernel and AutoGen stay supported, but new work goes into Agent Framework.
Check the link in the comments.
It merges Semantic Kernel's production tooling with AutoGen's multi-agent patterns
This fixes the prototype-to-production gap that's been killing agent projects
𝗪𝗵𝗮𝘁 𝘆𝗼𝘂 𝗴𝗲𝘁
Multi-agent orchestration that actually works. Sequential, concurrent, group chat, and handoffs. Pick the pattern that fits. Graph-based, so you can see what's happening instead of guessing.
Python and .NET support.
OpenTelemetry baked in. Trace every agent action, tool call, and decision. When you're running agents in parallel, this isn't optional.
Checkpointing and resume. Save state, replay workflows, and recover from failures. Makes debugging and experimentation possible.
Human-in-the-loop approvals. Mark any tool as requiring sign-off. The agent waits; you approve or deny, and it continues.
Open standards: MCP for tools, A2A for agent-to-agent, OpenAPI for REST endpoints. Your agents aren't locked to Microsoft.
𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀
Semantic Kernel users had stability but limited orchestration. AutoGen users had flexibility but no durability. Agent Framework gives you both.
KPMG is using it for audit automation. BMW for real-time telemetry analysis. Commerzbank for customer support. Production systems in regulated industries.
𝗠𝗶𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝘀𝗶𝗺𝗽𝗹𝗲
Semantic Kernel: replace Kernel with Agent. Tools instead of plugins.
AutoGen: AssistantAgent becomes ChatAgent. Event-driven runtime becomes typed Workflows.
Microsoft is shifting focus here. Semantic Kernel and AutoGen stay supported, but new work goes into Agent Framework.
Check the link in the comments.
Consistency eats inspiration for breakfast
Small steps, daily
Pick one tiny action: 10 minutes of writing, one outbound email, or a set of push-ups
Same time, same place
Track the streak
The staircase wins
Image: Liz Fosslien
Small steps, daily
Pick one tiny action: 10 minutes of writing, one outbound email, or a set of push-ups
Same time, same place
Track the streak
The staircase wins
Image: Liz Fosslien
𝗪𝗵𝗮𝘁 𝗶𝘀 𝗮 𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗮𝗹 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿 𝗮𝘁 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁?
Dejan Dundjerski replatformed Azure SQL Managed Instance, live, at scale, without customers noticing. The service grew 10x in parallel.
Before that, he developed an AI expert system for troubleshooting millions of databases, a field in which no one in academia had previously researched.
He says that seniors identify problems and propose solutions, while principals implement them and get others to follow.
This interview breaks down:
🔹 How he makes architectural decisions (start with the simplest solution)
🔹 Why depth in your domain prevents surprises
🔹 How to avoid workarounds that make systems 10x more complex
🔹 What to do when dependencies slip
👉 Read it here: https://lnkd.in/dsgxE6sh
Dejan Dundjerski replatformed Azure SQL Managed Instance, live, at scale, without customers noticing. The service grew 10x in parallel.
Before that, he developed an AI expert system for troubleshooting millions of databases, a field in which no one in academia had previously researched.
He says that seniors identify problems and propose solutions, while principals implement them and get others to follow.
This interview breaks down:
🔹 How he makes architectural decisions (start with the simplest solution)
🔹 Why depth in your domain prevents surprises
🔹 How to avoid workarounds that make systems 10x more complex
🔹 What to do when dependencies slip
👉 Read it here: https://lnkd.in/dsgxE6sh
𝗪𝗵𝘆 𝗚𝗼𝗼𝗴𝗹𝗲 𝗦𝘁𝗼𝗿𝗲𝘀 𝗕𝗶𝗹𝗹𝗶𝗼𝗻𝘀 𝗼𝗳 𝗟𝗶𝗻𝗲𝘀 𝗼𝗳 𝗖𝗼𝗱𝗲 𝗶𝗻 𝗮 𝗦𝗶𝗻𝗴𝗹𝗲 𝗥𝗲𝗽𝗼𝘀𝗶𝘁𝗼𝗿𝘆
Google’s monolithic repository provides a common source of truth for tens of thousands of developers worldwide. They use it for 95% of their source code, leaving Google Chrome and Android on their specific ones.
They used CVS and, after some time, migrated to Perforce and later replaced it with 𝗣𝗶𝗽𝗲𝗿. The 2016 source says it has more than 𝟮 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 𝗹𝗶𝗻𝗲𝘀 𝗼𝗳 𝗰𝗼𝗱𝗲 𝗮𝗻𝗱 𝟰𝟬,𝟬𝟬𝟬 𝗰𝗼𝗺𝗺𝗶𝘁𝘀 𝗽𝗲𝗿 𝗱𝗮𝘆 by more than 10,000 engineers.
Engineers mainly use 𝘁𝗿𝘂𝗻𝗸-𝗯𝗮𝘀𝗲𝗱 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗺𝗼𝗱𝗲𝗹𝘀 at scale with much success. Regarding branching, changes are pushed to the main branch and submitted to code reviews. All of this removes nightmares of merge hell.
𝗧𝗿𝗶𝗰𝗲𝗱𝗲𝗿 carries out the first automatic checks when developers try to submit new code and provides initial automated feedback. Additionally, code reviews can be conducted using the Critique tool.
Engineers use 𝗥𝗼𝘀𝗶𝗲, a solution for massive refactorings, optimizations, code cleaning, and other tools. It enables changes to be divided into minor changes, which are then reviewed by each owner before being implemented.
Some 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲𝘀 of this approach are:
🔹 𝗨𝗻𝗶𝗳𝗶𝗲𝗱 𝘃𝗲𝗿𝘀𝗶𝗼𝗻𝗶𝗻𝗴
🔹 𝗘𝘅𝘁𝗲𝗻𝘀𝗶𝘃𝗲 𝗰𝗼𝗱𝗲 𝘀𝗵𝗮𝗿𝗶𝗻𝗴
🔹 𝗦𝗶𝗺𝗽𝗹𝗶𝗳𝗶𝗲𝗱 𝗱𝗲𝗽𝗲𝗻𝗱𝗲𝗻𝗰𝘆 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁
🔹 𝗔𝘁𝗼𝗺𝗶𝗰 𝗰𝗵𝗮𝗻𝗴𝗲𝘀
🔹 𝗟𝗮𝗿𝗴𝗲-𝘀𝗰𝗮𝗹𝗲 𝗿𝗲𝗳𝗮𝗰𝘁𝗼𝗿𝗶𝗻𝗴
🔹 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝗮𝗰𝗿𝗼𝘀𝘀 𝘁𝗲𝗮𝗺𝘀
🔹 𝗙𝗹𝗲𝘅𝗶𝗯𝗹𝗲 𝗰𝗼𝗱𝗲 𝗼𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽
🔹 𝗖𝗼𝗱𝗲 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆
And 𝗱𝗶𝘀𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲𝘀 include having to create and scale tools for development and execution, and maintain code health, as well as the potential for codebase complexity (such as unnecessary dependencies).
⬇️ Read more in the paper.
Google’s monolithic repository provides a common source of truth for tens of thousands of developers worldwide. They use it for 95% of their source code, leaving Google Chrome and Android on their specific ones.
They used CVS and, after some time, migrated to Perforce and later replaced it with 𝗣𝗶𝗽𝗲𝗿. The 2016 source says it has more than 𝟮 𝗯𝗶𝗹𝗹𝗶𝗼𝗻 𝗹𝗶𝗻𝗲𝘀 𝗼𝗳 𝗰𝗼𝗱𝗲 𝗮𝗻𝗱 𝟰𝟬,𝟬𝟬𝟬 𝗰𝗼𝗺𝗺𝗶𝘁𝘀 𝗽𝗲𝗿 𝗱𝗮𝘆 by more than 10,000 engineers.
Engineers mainly use 𝘁𝗿𝘂𝗻𝗸-𝗯𝗮𝘀𝗲𝗱 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗺𝗼𝗱𝗲𝗹𝘀 at scale with much success. Regarding branching, changes are pushed to the main branch and submitted to code reviews. All of this removes nightmares of merge hell.
𝗧𝗿𝗶𝗰𝗲𝗱𝗲𝗿 carries out the first automatic checks when developers try to submit new code and provides initial automated feedback. Additionally, code reviews can be conducted using the Critique tool.
Engineers use 𝗥𝗼𝘀𝗶𝗲, a solution for massive refactorings, optimizations, code cleaning, and other tools. It enables changes to be divided into minor changes, which are then reviewed by each owner before being implemented.
Some 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲𝘀 of this approach are:
🔹 𝗨𝗻𝗶𝗳𝗶𝗲𝗱 𝘃𝗲𝗿𝘀𝗶𝗼𝗻𝗶𝗻𝗴
🔹 𝗘𝘅𝘁𝗲𝗻𝘀𝗶𝘃𝗲 𝗰𝗼𝗱𝗲 𝘀𝗵𝗮𝗿𝗶𝗻𝗴
🔹 𝗦𝗶𝗺𝗽𝗹𝗶𝗳𝗶𝗲𝗱 𝗱𝗲𝗽𝗲𝗻𝗱𝗲𝗻𝗰𝘆 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁
🔹 𝗔𝘁𝗼𝗺𝗶𝗰 𝗰𝗵𝗮𝗻𝗴𝗲𝘀
🔹 𝗟𝗮𝗿𝗴𝗲-𝘀𝗰𝗮𝗹𝗲 𝗿𝗲𝗳𝗮𝗰𝘁𝗼𝗿𝗶𝗻𝗴
🔹 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝗮𝗰𝗿𝗼𝘀𝘀 𝘁𝗲𝗮𝗺𝘀
🔹 𝗙𝗹𝗲𝘅𝗶𝗯𝗹𝗲 𝗰𝗼𝗱𝗲 𝗼𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽
🔹 𝗖𝗼𝗱𝗲 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆
And 𝗱𝗶𝘀𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲𝘀 include having to create and scale tools for development and execution, and maintain code health, as well as the potential for codebase complexity (such as unnecessary dependencies).
⬇️ Read more in the paper.
𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗗𝗲𝘀𝗶𝗴𝗻 𝗚𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀
In their book "Framework Design Guidelines," Krzysztof Cwalina and Brad Abrams, who created a set of design principles on how to design frameworks at Microsoft in the early days of .NET development.
They contain the knowledge and experience accumulated over tens of thousands of developer hours across several .NET Framework releases.
They prepared a few groups of design principles:
𝟭. 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗗𝗲𝘀𝗶𝗴𝗻 𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀
✅ Scenario-Driven Design. Identifying the most likely scenarios for each feature area will help you start the design process for your public API. Create the code that your API's end users should use to implement these situations. Based on the model code you wrote, create your API.
✅ Usability studies. Test your API's usability. Selecting unfamiliar developers to implement the critical scenarios will help ensure that the API is used correctly. In addition, attempt to determine which elements of your API need more common sense.
✅ Self-Documenting API. The key scenarios should be implementable by developers using your API without consulting the documentation. By selecting logical names for the most popular kinds and members, you can help users understand which types they need to use in crucial cases and the semantics of the critical methods.
𝟮. 𝗡𝗮𝗺𝗶𝗻𝗴 𝗴𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀
✅ Do use PascalCasing (capitalize the first letter of each word) for all identifiers except parameter names. E.g., use FirstName instead of Firstname.
✅ Use camelCasing (capitalize the first letters of each word except for the first word) for all member parameter names.
❌ Do not use acronyms that are not generally accepted in the field.
✅ Do use well-known acronyms only when necessary.
❌ Do not use underscores, hyphens, or any other non-alphanumeric characters.
❌ Do not use the Hungarian notation.
✅ Do use the following prefixes, I for interfaces and T for generic type params.
𝟯. 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗗𝗲𝘀𝗶𝗴𝗻 𝗚𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀
✅ Do use the most derived type for return values and the least derived type for input parameters. E.g., take IEnumerable as an input param but return Collection as the return type.
✅ Do model higher-level concepts (physical objects) rather than system-level tasks with Aggregate Components, e.g., File is easier to understand than Stream.
❌ Do not require users of your APIs to instantiate multiple objects in main scenarios
✅ Do prefer classes over interfaces.
❌ Only seal types if you have a solid reason to do it.
❌ Only ship abstractions (interfaces or abstract classes) by providing at least one concrete type implementing each abstraction.
❌ Only ship interfaces that provide at least one API consuming the interface.
✅ Do strongly prefer collections over arrays in public API.
❌ Do not use error codes to report failures. Use Exceptions instead.
❌ Do not throw Exception or SystemException. Create your custom exceptions.
In their book "Framework Design Guidelines," Krzysztof Cwalina and Brad Abrams, who created a set of design principles on how to design frameworks at Microsoft in the early days of .NET development.
They contain the knowledge and experience accumulated over tens of thousands of developer hours across several .NET Framework releases.
They prepared a few groups of design principles:
𝟭. 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗗𝗲𝘀𝗶𝗴𝗻 𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀
✅ Scenario-Driven Design. Identifying the most likely scenarios for each feature area will help you start the design process for your public API. Create the code that your API's end users should use to implement these situations. Based on the model code you wrote, create your API.
✅ Usability studies. Test your API's usability. Selecting unfamiliar developers to implement the critical scenarios will help ensure that the API is used correctly. In addition, attempt to determine which elements of your API need more common sense.
✅ Self-Documenting API. The key scenarios should be implementable by developers using your API without consulting the documentation. By selecting logical names for the most popular kinds and members, you can help users understand which types they need to use in crucial cases and the semantics of the critical methods.
𝟮. 𝗡𝗮𝗺𝗶𝗻𝗴 𝗴𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀
✅ Do use PascalCasing (capitalize the first letter of each word) for all identifiers except parameter names. E.g., use FirstName instead of Firstname.
✅ Use camelCasing (capitalize the first letters of each word except for the first word) for all member parameter names.
❌ Do not use acronyms that are not generally accepted in the field.
✅ Do use well-known acronyms only when necessary.
❌ Do not use underscores, hyphens, or any other non-alphanumeric characters.
❌ Do not use the Hungarian notation.
✅ Do use the following prefixes, I for interfaces and T for generic type params.
𝟯. 𝗚𝗲𝗻𝗲𝗿𝗮𝗹 𝗗𝗲𝘀𝗶𝗴𝗻 𝗚𝘂𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀
✅ Do use the most derived type for return values and the least derived type for input parameters. E.g., take IEnumerable as an input param but return Collection
✅ Do model higher-level concepts (physical objects) rather than system-level tasks with Aggregate Components, e.g., File is easier to understand than Stream.
❌ Do not require users of your APIs to instantiate multiple objects in main scenarios
✅ Do prefer classes over interfaces.
❌ Only seal types if you have a solid reason to do it.
❌ Only ship abstractions (interfaces or abstract classes) by providing at least one concrete type implementing each abstraction.
❌ Only ship interfaces that provide at least one API consuming the interface.
✅ Do strongly prefer collections over arrays in public API.
❌ Do not use error codes to report failures. Use Exceptions instead.
❌ Do not throw Exception or SystemException. Create your custom exceptions.
𝗧𝗼𝗽 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀 𝗧𝗼 𝗟𝗲𝗮𝗿𝗻 𝗶𝗻 𝟮𝟬𝟮𝟱.
𝟭. 𝗣𝘆𝘁𝗵𝗼𝗻 - The AI/ML revolution drove this surge, with 66% of beginners now choosing Python as their first language. It is essential for machine learning, data science, backend development, and automation.
𝟮. 𝗝𝗮𝘃𝗮𝗦𝗰𝗿𝗶𝗽𝘁/𝗧𝘆𝗽𝗲𝗦𝗰𝗿𝗶𝗽𝘁 - JavaScript powers 98% of websites and remains the most-used language at 66% in Stack Overflow's survey. TypeScript continues rising in enterprise environments, offering type safety for large-scale applications. Together, they dominate web frontend, Node.js backend, React Native mobile, and serverless applications.
𝟯. 𝗖/𝗖++ - Combined, these languages hold strong positions across all indices. C++ ranked #2 in TIOBE (9.80%) as performance becomes critical for AI computation and high-frequency trading. Essential for systems programming, game development (Unreal Engine), embedded systems, and any application where microseconds matter.
𝟰. 𝗝𝗮𝘃𝗮 - Maintains #4 in TIOBE (8.76%) despite showing -1.5% decline. Still, it serves as the enterprise backbone for banking, healthcare, and large-scale business systems. Powers Android development, big data (Apache Spark, Hadoop), and backend systems via Spring Boot.
𝟱. 𝗖# - Ranked #5 in TIOBE (4.87%) with the largest decline among top languages (-1.85%). Remains strong in the Microsoft ecosystem (.NET, Azure) and game development through Unity engine, which dominates mobile and indie games.
𝟲. 𝗦𝗤𝗟 - Ranked #3 in Stack Overflow at 59% usage. It is not a general-purpose language, but it is essential across all industries. Every application needs databases, making SQL a must-have skill alongside any other language. Job postings consistently seek "Language X + SQL" combinations.
𝟳. 𝗚𝗼 - Entered TIOBE's top 10 at #7 (2.04%), marking a significant milestone. Created by Google, Go powers Kubernetes and Docker, the foundation of modern cloud infrastructure. Excels at microservices, backend APIs, and DevOps tools. Saw +2 percentage point increase tied directly to AI infrastructure needs.
𝟴. 𝗣𝗛𝗣 - Ranked #14 in TIOBE (1.28%) and #8 in PYPL (3.19%), representing a dramatic fall from its peak. Lost 84% of its relative position since 2008. Yet still powers 75.6% of websites using server-side programming, primarily through WordPress. Used mainly for legacy maintenance rather than new projects.
𝟵. 𝗥𝘂𝘀𝘁 - Ranked #18 in TIOBE (1.01%) but named "most admired language" with 72% satisfaction for the ninth consecutive year in Stack Overflow. Growth tied to memory safety concerns and White House cybersecurity recommendations. Used for systems programming, security-critical software, blockchain, and high-performance backends, but also for Linux kernel development.
𝟭𝟬. 𝗦𝘄𝗶𝗳𝘁 - Ranked #21 in TIOBE (0.85%) and #9 in PYPL (2.93%). The primary language for iOS, iPadOS, macOS, and watchOS development. Rose from unranked to #15 in Pluralsight's index since 2022.
𝟭. 𝗣𝘆𝘁𝗵𝗼𝗻 - The AI/ML revolution drove this surge, with 66% of beginners now choosing Python as their first language. It is essential for machine learning, data science, backend development, and automation.
𝟮. 𝗝𝗮𝘃𝗮𝗦𝗰𝗿𝗶𝗽𝘁/𝗧𝘆𝗽𝗲𝗦𝗰𝗿𝗶𝗽𝘁 - JavaScript powers 98% of websites and remains the most-used language at 66% in Stack Overflow's survey. TypeScript continues rising in enterprise environments, offering type safety for large-scale applications. Together, they dominate web frontend, Node.js backend, React Native mobile, and serverless applications.
𝟯. 𝗖/𝗖++ - Combined, these languages hold strong positions across all indices. C++ ranked #2 in TIOBE (9.80%) as performance becomes critical for AI computation and high-frequency trading. Essential for systems programming, game development (Unreal Engine), embedded systems, and any application where microseconds matter.
𝟰. 𝗝𝗮𝘃𝗮 - Maintains #4 in TIOBE (8.76%) despite showing -1.5% decline. Still, it serves as the enterprise backbone for banking, healthcare, and large-scale business systems. Powers Android development, big data (Apache Spark, Hadoop), and backend systems via Spring Boot.
𝟱. 𝗖# - Ranked #5 in TIOBE (4.87%) with the largest decline among top languages (-1.85%). Remains strong in the Microsoft ecosystem (.NET, Azure) and game development through Unity engine, which dominates mobile and indie games.
𝟲. 𝗦𝗤𝗟 - Ranked #3 in Stack Overflow at 59% usage. It is not a general-purpose language, but it is essential across all industries. Every application needs databases, making SQL a must-have skill alongside any other language. Job postings consistently seek "Language X + SQL" combinations.
𝟳. 𝗚𝗼 - Entered TIOBE's top 10 at #7 (2.04%), marking a significant milestone. Created by Google, Go powers Kubernetes and Docker, the foundation of modern cloud infrastructure. Excels at microservices, backend APIs, and DevOps tools. Saw +2 percentage point increase tied directly to AI infrastructure needs.
𝟴. 𝗣𝗛𝗣 - Ranked #14 in TIOBE (1.28%) and #8 in PYPL (3.19%), representing a dramatic fall from its peak. Lost 84% of its relative position since 2008. Yet still powers 75.6% of websites using server-side programming, primarily through WordPress. Used mainly for legacy maintenance rather than new projects.
𝟵. 𝗥𝘂𝘀𝘁 - Ranked #18 in TIOBE (1.01%) but named "most admired language" with 72% satisfaction for the ninth consecutive year in Stack Overflow. Growth tied to memory safety concerns and White House cybersecurity recommendations. Used for systems programming, security-critical software, blockchain, and high-performance backends, but also for Linux kernel development.
𝟭𝟬. 𝗦𝘄𝗶𝗳𝘁 - Ranked #21 in TIOBE (0.85%) and #9 in PYPL (2.93%). The primary language for iOS, iPadOS, macOS, and watchOS development. Rose from unranked to #15 in Pluralsight's index since 2022.