@eikonne I think this is true of most things. Knowing how things should be is 5% of the effort…driving the change is where the value is…but no always the limelight. Was it Patty McCord who said “problem finders are cheap”? 🙂
We want to provide our best-in-class mentoring program free to 50 working class or first gen women in science in the UK, but we need your help to do it. Our structured online program, with careers & skills advice, hands-on support, works. Donate: https://t.co/kI2mBPPLiy
@mattwynne Happy to chat any time, let me know if you want to go into more depth / detail. Hope you’re keeping well - looking forward to meeting up again when we’re out the other side of all this!
@mattwynne Specifically, it allowed us to remove the traditional UAT sign off late in the cycle, replaced with sign-off of the behaviours up front. When they are coded and go green, we can push. There are other considerations of course - e.g usability - but these were dealt with mid sprint
@khangharoth Strong agree! “But I just like designing and building cars, I don’t have much interest in how they perform over time once they roll off the conveyor belt..” - that kind of thinking isn’t much good to car buyers who need to get places!
@tottinge@bguthrie @tastapod @ckeithray @tastapod has been nudging me to write up something more comprehensive than a few tweets for a while! I've not prioritised this yet, but there's clearly some interest in the topic. I write roughly one blog post a year, maybe 2021 is when this one gets written up 😀
@bguthrie @tastapod @tottinge @ckeithray However, in that regulated environment, we were able to deploy multiple times a day with no user down time, and it was because we were provably compliant that we were able to do that. It took effort to get there for sure but in the end, a lot more was delivered as a result
@bguthrie @tastapod @tottinge @ckeithray As ever, it's all about context. This was in a regulated environment. In others, going fast and only adding the controls you really need after a few crashes may be a better way. A/B testing and great o11y can help limit the blast radius of that approach