Every tech company* has platform teams trying to build:
1. Heroku, except hand-rolled
2. One giant shared database, so engineers can ignore analytics without consequence
3. If they have a monolith, microservices. If microservices, a monolith
4. A solution to the halting problem
* Hyperbole. I hope
1. Heroku, except hand-rolled
2. One giant shared database, so engineers can ignore analytics without consequence
3. If they have a monolith, microservices. If microservices, a monolith
4. A solution to the halting problem
* Hyperbole. I hope
Comments
In my last job, we were stuck in an endless cycle of discussing What To Do About The Monolith™ because no one who could take charge of it would do so.
Rarely hear it requested by customers, though.
Guess which one
The review processes claiming they can prove software will behave (without running the programs, and often before any implementation at all) are what make me roll my eyes.
You can get surprisingly far for limited classes of programs though.
Not everyone's willing to commit.
Why does it matter how loud my tables are?
I just try to get the table volumes right. Maybe 20 dB to whisper a secret. Maybe 125 dB for a yell and announcement.
What dB do you use when you all are talking to each other.
Buuuut, of course, some of us may have been old enough to have used dBase.
Still tried to use the DB for message exchange.
FFS.
i assure you i spent ten years going "THE FUCK"
They're even discussing about rewriting one data intensive service in Rust.
a lot of that code is still going today
One of my theories is that people hear the success stories & wildly underestimate the investment that went into them.
It was a LOT of work.
Yes I need to write it in Perl, why?
🌽😹😸😿🌽