In production I find it a bit easier to profile what's happening if the db is performing poorly since I can login to the box and use standard tools - is it cpu/io bound, etc.
1,783 karma · joined April 5, 2011
In production I find it a bit easier to profile what's happening if the db is performing poorly since I can login to the box and use standard tools - is it cpu/io bound, etc.
I understand it, I run it, and I don't have to deal with 3rd party's changes, performance problems, or downtime. Also, less bugs related to data consistency.
That being said, I've seen VMs with multiple years of uptime on various clouds, so ymmv.
Seems complex - any ideas why the use of Rust extensions vs doing it natively in PG?
Interested to see in what ways people could improve this. I followed some of the principles and terminology of beanstalkd which I used some years ago, but seems to have fallen out of use.
I can't imagine there are a lot of Fortran folks around maintaining these old libraries - they must need maintenance no?
Better alternative I've found is squash merge - topic/feature branches are squashed as a single commit instead of bringing down each individual commit or creating a merge commit. You're history is cleaner, you're able to revert stuff easily, and it's really hard to mess up since it's just an atomic last step you do in your workflow.
In a well-factored app, stepping on one another's toes should be the exception, not the rule.
Unfortunately, I think modern web frameworks don't do a good job of enabling this. They often encourage "God object" patterns and having code/data very tightly coupled.
Try to understand why, learn from it; but get out, this isn't going to end well.
The fastest single-threaded cpus today are only 2x faster than the fastest single-threaded cpus of a decade ago...
But, serving static pages is very very fast afterwards.
Elon is definitely working to kill the platform with all these random by fiat changes he's making in the name of "free speech" - which to the right means you can lie and spread as many conspiracy theories as you want. I don't think the general population is interested in that - there is a medium of sensible moderation that keeps discourse civil and roots out the worst of the lies and conspiracies.
Anything was fair game as long as it was technical and interesting. My main goal was to level-up the team a bit, show them how fun actually building stuff was vs adding a dependency of some 3rd party service or library, and have some fun just nerding out a bit on what other stuff was going on in OSS or the community at large.
I've seen terrible spaghetti code apps where literally nothing can be refactored because of the model / view, God object dependency stuff all over the place.
Also, a lot of SWEs most significant code bases they've worked on will not be theirs, but property of a previous employer.
The best I've figured out is to have a straightforward coding exercise that doesn't have a lot of dependencies and can be solved fairly quickly in any language the candidate chooses. Do this live in one of your interview sessions - I don't like take-home stuff, it discourages above-average candidates.
Have a couple places where they could chose the wrong data structure or algorithm. And have a follow-up that modifies the first solution to do something slightly different. Give this same exercise to all candidates - you'll more easily know if candidate A is better than candidate B, etc.