Welcome to the new week!

I observed that documentation is one of the biggest struggles in modelling business processes in our systems. We already have techniques, like EventStorming or Event Modeling, that are helpful but are the beginning of the journey. A lot of my clients are asking me how to move further from the model that came out of the session as it’s hard to keep it up to date when software evolves, more as an input to the decision log.

I recently started to look around for inspiration and decided to learn more about BPMN (Business Process Model and Notation). I’m not sure if that’s the ultimate solution, but knowing more perspective can definitely help build a better modelling toolbox.

I’m very happy that Mário Bittencourt agreed to share his knowledge and experience with us in the upcoming webinar on 28th November at 6 PM. Mário is a Principal Software Architect at SSENSE.com. 30 years of experience in software development with a passion for the purposeful use of technology. See more details.

Become a paid subscriber and join us live!

Of course, I need to start with the fun around OpenAI. I thought I’d make it a leitmotif of this release, but I think the whole story is more hilarious than interesting. Just to recap what has happened:

1. OpenAI announced leadership transition. Informing that

   Mr. Altman’s departure follows a deliberative review process by the board, which concluded that he was not consistently candid in his communications with the board, hindering its ability to exercise its responsibilities. (…) The board no longer has confidence in his ability to continue leading OpenAI.

   This is a surprisingly strong statement, and the fact that they didn’t have a replacement showed that it was some rapid, preemptive move.

2. Then there was no new news, maybe besides the shock and information that they were fired via Google Meet, which is ironic for an MS-backed startup. Why not through MS Teams?

3. Then it appeared that VC Empire struck back, and OpenAI board started discussions with Sam Altman to return as CEO.

4. But he didn’t. Because OpenAI got another CEO from Twitch.

5. And because Microsoft hired Sam Altman.

…and probably some more at the time you’re reading it.

All of that looks like an aftereffect of 3-year-old kids overdosing on sugar and watching too much of Game of Thrones. Jokes aside, that’s also an example of how VC-backed board dynamics can end up. And yes, that’s my interpretation.

For better, check:

• Dylan Patel, Daniel Nishball - Microsoft Swallows OpenAI’s Core Team – GPU Capacity, Incentive Structure, Intellectual Property, OpenAI Rump State

If you’re curious about the AI stuff and not politics, check out more:

• Simon Wardley - Why open source AI matters.

• Jodie Burchell - Garbage data in, garbage models out

Speaking about Open Source:

• We introduced support plans for Marten

For many companies, no paid support means a hobby project, which we’re not. We started our GitHub Sponsors some time ago. We’re already getting some income from our great users, but for many companies, sponsoring is hard to deal with, as procurement processes are not adjusted to that. Companies need invoices, and we fully understand that. That’s where official support can benefit both sides, smoothening the process.

So it’s about bringing sustainability to both sides, us OSS creators and companies that want to ensure we won’t disappear.

Google published the lessons they learned from running Site Reliability Engineering.

• Google - Lessons Learned from Twenty Years of Site Reliability Engineering. Or, Eleven things we have learned as Site Reliability Engineers at Google

Those 11 things are:

1. The riskiness of a mitigation should scale with the severity of the outage.

2. Recovery mechanisms should be fully tested before an emergency.

3. Canary all changes.

4. Have a "Big Red Button".

5. Unit tests alone are not enough - integration testing is also needed.

6. COMMUNICATION CHANNELS! AND BACKUP CHANNELS!! AND BACKUPS FOR THOSE BACKUP CHANNELS!!!

7. Intentionally degrade performance modes.

8. Test for Disaster resilience.

9. Automate your mitigations.

10. Reduce the time between rollouts, to decrease the likelihood of the rollout going wrong.

11. A single global hardware version is a single point of failure.

As you can see through the CAPS LOCKS HEADINGS, they learned some lessons. Not much is surprising there, but it’s a decent article reminding us of the importance of pragmatism. It also shows that assuming that all can go wrong is relieving, as then we can start thinking about what, in practice, we can do about that.

Microsoft released a new .NET version: 8. Also, there are not many surprises and big features, but stable, gradual improvements. The most intriguing thing is around the new tool called Aspire. Microsoft aspired to build their development portal. It looks intriguing, although the controversy is that (for now) it’s only built for .NET and focused on a monolithic approach. However, the idea is nice. It groups access to telemetry, data etc., in one place. Still, I predict it’ll end up as projects like Tye. You don’t remember Tye? Yes, that’s what I mean. Nevertheless, I plan to play with it more soon and keep you posted.

• Microsoft - Announcing .NET 8

• Microsoft - .NET Conf 2023

• Microsoft - Introducing .NET Aspire: Simplifying Cloud-Native Development with .NET 8

• Microsoft - Announcing F# 8

We also had an interesting reunion of the old masters. Ivar Jacobson and Alistair Cockburn meet live to bring back the use cases under discussion.

• Ivar Jacobson, Alistair Cockburn - Use Cases are Essential

A must-watch on how to think about the development process, keeping the focus on the application and user behaviour.

What’s up in the Frontend world? Are micro-frontends still fashionable? What’s the state of them? Let’s check the source:

• Micro Frontends Meetup - State of Micro Frontends with Luca Mezzalira

Should you use the new frameworks? Do everything in JS? Surprisingly, not always. I’m happy that we’re getting back to the Semantic Web idea. I like JS and TS, but it sounds like we were trying to fix the issue of building complex UIs with advancing JS technologies instead of solving the real cause. Will the HTML-first and WebComponents movement will help in that? Maybe, let’s check:

• HTML First

• Jake Lazaroff - Web Components Will Outlive Your JavaScript Framework

• CSS-Tricks - Web Components Are Easier Than You Think

Check also other links!

Cheers!
Oskar

p.s. I invite you to join the paid version of Architecture Weekly. It already contains the exclusive Discord channel for subscribers (and my GitHub sponsors), monthly webinars, etc. It is a vibrant space for knowledge sharing. Don’t wait to be a part of it!

p.s.2. Ukraine is still under brutal Russian invasion. A lot of Ukrainian people are hurt, without shelter and need help. You can help in various ways, for instance, directly helping refugees, spreading awareness, and putting pressure on your local government or companies. You can also support Ukraine by donating, e.g. to the Ukraine humanitarian organisation, Ambulances for Ukraine or Red Cross.

Architecture

• Mário Bittencourt - BPMN for Profit and Fun: Seamless Teams Collaboration in Software Development

• Emily Rosengren - Can We Please Stop Talking About Tech Debt?

• Continuous Delivery - Visualizing & Prioritizing Technical Debt To Identify RED CODE with Adam Tornhill

• Brian Foote, Joseph Yoder - Big Ball of Mud

• David Hope - Understand your data requirements

• Mihir Sathe - Load Balancing: A More Realistic Model

• AirBnB - Scaling Airbnb’s Payment Platform

DevOps

• Google - Lessons Learned from Twenty Years of Site Reliability Engineering. Or, Eleven things we have learned as Site Reliability Engineers at Google

API

• IETF - The Idempotency-Key HTTP Header Field

Testing

• The NewStack - The Struggle for Microservice Integration Testing

Frontend

• Micro Frontends Meetup - State of Micro Frontends with Luca Mezzalira

• HTML First

• Jake Lazaroff - Web Components Will Outlive Your JavaScript Framework

• CSS-Tricks - Web Components Are Easier Than You Think

AI

• Jodie Burchell - Garbage data in, garbage models out

Azure

• azcagit - GitOps Engine for Azure Container Apps

Java

• Yishai Galat - Java at AWS: Lessons Learned from Upgrading and Modernizing a Massive JVM Codebase

.NET

• Oskar Dudycz - We introduced support plans for Marten

• Microsoft - Announcing .NET 8

• Microsoft - .NET Conf 2023

• Microsoft - Introducing .NET Aspire: Simplifying Cloud-Native Development with .NET 8

• Microsoft - Announcing F# 8

Rust

• NGINX - Extending NGINX with Rust (an Alternative to C)

Management

• Ivar Jacobson, Alistair Cockburn - Use Cases are Essential

• Charity Majors - Compliance standards should be modern development practices

Product Design

• Joakim Sundén - The Product Model at Spotify

Industry

• Simon Wardley - Why open source AI matters.

• OpenAI - OpenAI announces leadership transition

• The Verge - OpenAI board in discussions with Sam Altman to return as CEO

• The Verge - Sam Altman isn’t coming back to OpenAI

• The Verge - Microsoft hires former OpenAI CEO Sam Altman

• Dylan Patel, Daniel Nishball - Microsoft Swallows OpenAI’s Core Team – GPU Capacity, Incentive Structure, Intellectual Property, OpenAI Rump State

Security

• BleepingComputer - New Microsoft Exchange zero-days allow RCE, data theft attacks