That is a good point!

The most popular is often the worst. Consider:

• McDonalds hamburger - definitely not the best - far from it.
• Bud beer - definitely not the best - far from it.
• ...

I hope the airlines lose and have to spend billions of dollars.
The customers would only get a small amount and the lawyers for the plaintiffs get the other billions but at least the airlines would be out an amount that makes them think next time.

Yes indeed. When they abandoned their engineering culture, they abandoned their product and their customers.

Yes, in my experience you are right. Except in the domain of aerospace sometimes.

Maybe because "hacking" is not how real long-lasting code is written. Durable expandable maintainable code is not "hacked together" - it is engineered, which takes time - more than a day at a hackathon.

"Free" at the cost of enduring obnoxious commercials!

I for one will never watch anything on Freevee. The whole reason that I cut the cable and went to streaming only was to avoid obnoxious ads and the chaotic obnoxious cable environment. That _last_ thing I want is to return that that awful kind of experience in streaming.

The McDonalds hamburger is the "top" hamburger. See my point? In fact, there is often an _inverse_ relationship between "top" and "best".

Thanks for sharing all this.

I also do not like IDEs of any kind. I feel boxed in. I like to build using the actual build script, not the IDE.

With FaaS, how do you keep track of all the things that can go wrong? You refer to that - I am wondering what you do. I am talking about cases where a function fails, or a user changes their mind. Some people call these "sagas" and "compensating transactions".

I also design for failure. You are probably familiar with the engineering term "failure modes". I always think through all the failure modes and make sure they are all handled.

I would love your reactions to an article that a colleague of mine and I wrote a little while back on this: https://agile.org.uk/rational-...

Hi -

Do you write unit tests?

One thing I should mention is that I rarely write unit tests. The reason: for typesafe langauges, I don't need them. But I have found that for type-unsafe language, I need a whole suite of unit tests. To me, it's a kind of tax.

What didn't you like about Java?

BTW, Java has evolved a lot. They have made it a little less verbose. But again, I find that the verbosity is more than made up for by the type safetynet. I have refactored large code bases and introduced zero errors - once it compiled, everything worked. I have done the same in Ruby but dozens of bugs resulted.

Disclosure: I wrote Sun's enterprise Java textbook.

Typescript is typesafe. I considered it, but it is a hassle because it requires an extra layer of tooling. I really wish that Javascript would be enhanced so that it conforms to typescript.

Having used both typesafe (Pascal, C, C++, Ada, Scala, Module 2), and type unsafe languages (Javascript, Ruby, Go (yes, Go's "duck typing" is unsafe), and Python) for 50 years, my experience has been that type unsafe languages save ten seconds up front but cost 10 hours later... :-/

Hi. I avoided a framework because the frameworks are huge and I would have to spend a week learning the framework. And in the end, the framework might get in the way. E.g. my app makes heavy use of SVG.