The biggest driver for push vs pull in my experience is “who does the work”. If I need data and I have to ask 7 different teams to “push”, it ain’t happening.

dupe of: https://news.ycombinator.com/item?id=46184366

... and don't forget to flush!

It's funny how "push vs pull" always sounds like a technical choice, yet it usually exposes the social wiring of a team more than anything else. When every system expects to be pulled from, you start to see who actually depends on whom, and where the real bottlenecks live. The post captures that quiet truth that architecture isn't just code, it's how people negotiate responsibility.

So… caches. The unfortunate reality of high used systems.

Yeah that was a lot of words to say caching is needed sometimes.

As always it's an engineering decision full of tradeoffs. You should never duplicate your data, except when you need to...etc etc.

Or indexes, which are after all a structured way of duplicating data for performance, sometimes so much so that they entirely replace the actual table row lookup.

Yepp, making exactly that same point in the post:

> This is why we have indexes in databases, which, if you squint a little, are just another kind of materialized view, at least in their covering form.