Відео

James is a legend thank you for uploading this

A "real" example. Very nice.

Can't believe I missed seeing Dave in my own country! 😢

The best talk I've ever heard Dave give.

"Guided, incremental change" is not evolutionary.

Constant feedback in Scrum? Since when?

I just thought I'd throw in that "How Buildings Learn" is also a 6 part TV series from 1997 and it's available on here if you dig around a bit. ... and whilst we're on the subject of books... I love that bit when The Endless are all arguing (again) and Delerium says "stopitstopitstopitstopit!!!!!!" Cloneable... I've been trying to convince "the entire IT industry" That E. mail is spelled "E. mail" and not "email" for a few decades now... too late.

One of the most useful presentations related to architecture that I have ever seen!

Related to the 8 sagas and what our system needs. Our system may have several parts and certain parts need another saga because they have other needs. That is, the context is not necessarily something homogeneous.

"Agility" does not have to be a dimension like "scalability", because it is a resulted and composed trait (as feasibility) and is relative to the context. The same architecture can be agile in one context and exactly the other way around in another context.

You don't have to throw away ADR, because the next ones will come and won't understand those decisions and will instantly introduce an accidental complexity. So, if the company doesn't use them, you have to use them and try to convince others to use them. Architecture ~ the most important decisions, if you don't understand the decisions, you don't understand the architecture. And they also can be used as a tool for evolution and improvement.

Yet the things he complains about come from the Scrum camp, nearly exclusively.

Thank you for sharing!

I don't know Kevlin... sometimes the idea of "swim away" does have a certain attraction. ;) For anyone who doesn't grasp entropy... could I just recommend Jim Al-Khalili's excellent two-part documentary "Order and Disorder". It is to thermodynamics and information theory what a good Kevlin Henney talk is to computer science.

Fun part: Excel datetime columns cannot handle timezones. Everytime I export a pandas dataframe into Excel I have to go thru the datetime fields and remove the timezone. Welcome to the 21st century Excel!

I got a few recommendations to watch this video, but when listening on headphones it is killing me, as the audio is one channel (L). Is it possible to make it stereo (maybe just copy/paste the left audio channel to the right audio channel)?

ChatGPT - Having a sociopathic bullshitter acting as our "pair programmer" sounds a lot like using existing social media platforms to find ideas and advice for our programming. 33:21 - WOW! Kevin is using the correct precise, philosophical definition of "bullshit" at this point. But, that should be no surprise... things like that are why we listen to his talks.

Wow. Fking terrible delivery. Hope you learn how to deliver message properly at some point.

Great delivery ♥

Audio 😥

Modularity is all you need

AUDIO IS BROOOOOOTAL

This is pure gold. Great delivery as well.

Thanks guys for this great presentation of the Multi tenancy concept, I loved the drawings and the high spirited presenter. Have a great day you all 🎉😊❤

First liker.

In Jim's talk - he explains that hubs are akin to community of practice / interest and the more communities you are part of the higher the chance that you'll know someone in a team that you want to talk to and without the need of going up the org structure and then down the structure to speak to someone without upsetting the leadership hierarchy. This hub based approach is through weak leaks and enables teams to self manage.

Oof, dont listen with headphones! My left ear is dying 😅

Which ChatGPT version did you use? V4 is significantly better than 3.5. Soon v5 is thee..

Scrum is known to lots of people as a word in countries that play rugby. I don’t know why he thinks it’s rare. Thanks for the rest of the talk.

I think the way Henney presented the PDSA-cycle, it's painstakingly obvious that it supports the idea of proposing a project, change in the work culture or whatever it would be. Then you found a group that will try to flesh the plan out and make a small trial project, learn about the new thing and try to make it work. Something preferrably representative of the actual change in the company. Then they report the findings, the review on the experience and write a paper for the management to gauge if it is worth going through in the bigger picture, changing the actual company process. Basically study and then act on the change after the successful pilot project where the company gained knowledge on the new thing, its pros and cons. Not just jumping in on a new thing and trying to change perhaps the whole company's process and trying to force it through until learning it's not feasible or forcing it through in a way that it's now in action but it's not an improvement for anyone, almost the contrary.

Building on cloud services is moronic. Never build your house on someone else's land

22:37 Greatest owning of ChatGPT and other tech like it in 6 seconds or less!

17:12 I'm a bit confused. What is the difference between "business value" and "estimated business value"? I couldn't read the hidden nuance here

Foarte faină prezentarea! Felicitări echipei care a realizat-o!

The only people who are going to watch a video like this are the ones who can't make any effective changes. Software developers are slaves to rampant corporate corruption. There is no rational argument you can make to change management's mind. They do not care about productivity, morale, the product, the customer, cost, or even the company or the shareholder. They care about their personal agendas and power. Watching them interact with each other is like watching A Game of Thrones. The only way to make a difference is to do good work in secret.

This guy had better ALWAYS think correctly, completely and overall flawless. ONE misstake from mr Henney will create a black hole - of unlikeliness, but the agan if he is on The Heart of Gold, the Infinite Improbability Drive would kick in and take him safely to the other end of universe in jiffy.

The Gödel Theorem is once again used almost entirely wrong. For anyone interested, there is a great video on here somewhere that explains the misconception.

Awesome lecture. Thanks.

excellent video

Omg, the quality of this video is awful. Left ear audio only, terrible quality

I wanted to watch/listen because the title sounded like it was challenging some of my beliefs. Unfortunately all I hear is an angry old man ranting against strawmen.

"Nothing is more dangerous than an idea, when you have only one idea." In my lifetime and carreer the number one and only idea has undoubtedly been the relational database.

"Prediction is very difficult, especially about the future" reminded me of a statement once made by a co-worker: "I hate collaboration, particularly with other people."

This is great.

@16:40 the difference between "implementing features that nobody will use" and producing real business value is not "your boss/manager/the architect explained why they're doing the feature to you". Bosses and managers who deliver the marching orders from on-high always explain the "why"; they're just wrong and their ideas are garbage. You only implement good features when the instructions from the top are high-level key objectives and the details of the features come from the bottom.

@12:45 I can't post the actual link here in the comment, but if you look up Lars Goran Wallgren from the University of Gothenburg, Sweden, you will be able to download the thesis.

Kevin is never not interesting

I agree with most of what James Coplien said, even if it is pretty harsh on Agile. There is too little emphasis on thinking and too much emphasis on coding. Architecture, flow charting and planning are thinking. Getting a complete structure in your head before you start is difficult. Managers can't measure progress in thinking, but that can measure code. The Sprint concept that you work on only one idea and get it done completely and then move on doesn't mean that it is compatible with the whole system. I have found that coders are very resistant to domain knowledge. "Don't explain anything, I won't listen. Just give me a 100% perfect algorithm and I'll code it." I've run into this again and again. If you have software in, software out systems then their domain knowledge of software is adequate. That is where they are happy. Where is this customer feedback that Agile responds to? It comes weeks after the checkmark has been put in the checkbox. Completely unsatisfactory. Breaking systems into pieces (reductionism) is good. But you also have to have integration where you put them back together. When you do detailed design, the interfaces are subject to continuous improvement. Where is integration in Agile? A separate Quality Control engineer who doesn't understand anything? Everything is developed one at a time in some priority driven sequence. Complete systems design doesn't fit the sequence of user stories model. Variation is one of those double meaning words. In automotive production you use interchangeable parts assembled by standard work over and over and over again. Variation is minimized. In set-based design, you do not guess point-based solutions. This just leads to loopbacks after release when you guessed wrong. You keep your options open as long as possible and let constraints narrow the range of values Interlocking Venn diagrams. People often over constrain. In the beginning it costs nothing to ask for the sun, the moon and the starts. As soon as people are told the sun will cost $10 million, the moon will cost $1 million and the stars will cost $1 billion, they decide that low earth orbit is just fine for their needs. Variation here is the range of acceptable solutions. Great talk!

For those who aren't such ardent Kevlin Henney fanboys as me... yes, the bad sound does go all the way to end. :(

Super cool talk, I am love it! Thanks for sharing it!