I don't think it misses that point. System A needs building blocks X,Y,Z to work. They can either buy or build them.
Someone needs to have built blocks X,Y,Z for them to be available for purchase, but the people building X,Y,Z aren't trying to build A. They have their own goal, which is building X,Y,Z, which probably could be done using blocks U,V,W.
It's not "consumerist" so much as "informed about the state of the art." The opposite is generally "I'm a genius engineer who can reinvent something better than what exists without looking at what exists." (You probably can't).
The prices for a good buy make no sense to an existing business (without VC infuse). RENT beats build.
> System A needs building blocks X,Y,Z to work
When an American company buys a product with assurances that certain qualities exist (even as small as a $10k purchase), roughly 95% of the time it will fall short. You will have to build some hack solution (which the seller is usually happy to point out) or just ignore the missing feature. There's usually very little you can cannibalize out of a purchased product sourcecode in a full rewrite.
There was a 1 in 25+ cases as an exception, in my career. My company A was purchasing another european company B at over $100m with a full audit clause prior. My team was flown out to the UK and worked on Company B's technology for a couple weeks. It operated exactly as expected with the capabilities they presented (as the audit assured) when we brought the code back to the developers in the US...with branding changes, etc.
> 5. QA Gates Make Quality Worse
> Secondly, the teams doing QA often lack context and are under time pressure. They may end up testing “effects” instead of “intents”.
A lot of this is finger pointing. If you don't tell the QA what or why, the tests can't be written to infer this. If the tests can't be maintained, you have a failing QA dept, not a barrier to quality by the mere existence of the gates.
> 7. Simple Always Wins
Then you don't need to BUY it, just BUILD it.
> 8. Non-Production Environments Have Diminishing Returns
>> 8. Non-Production Environments Have Diminishing Returns.
>You want at least 3, maybe 4.
After experiencing a project where delivery was slow as a dog because of a lot of queues waiting on multiple rounds of manual QA and environmental differences causing a class of bugs that only showed up in production anyway further delaying other stuff as it was fixed and had to go through these long lead times to get to production, I did some soul searching and tried to understand it better. I know we could tune the current setup a little better and eek out some marginal gains. But after reading Accelerate and comparing our dev practices with a local company that is doing Continuous Delivery as outlined in this blog post I really feel like the grass might actually be significantly greener on the other side.
Someone needs to have built blocks X,Y,Z for them to be available for purchase, but the people building X,Y,Z aren't trying to build A. They have their own goal, which is building X,Y,Z, which probably could be done using blocks U,V,W.
It's not "consumerist" so much as "informed about the state of the art." The opposite is generally "I'm a genius engineer who can reinvent something better than what exists without looking at what exists." (You probably can't).