HACKER Q&A
📣 the-smug-one

How much of your code is pushing data around?


A lot of code I've written for products is based around pushing data around. For example:

0. Parse native datastructure into JSON 1. Make HTTP request with that JSON 2. Parse JSON from response into native datastructure 3. Parse native datastructure into gRpc and send off 4. Parse back gRpc into other language's native datastructure 5. Put into relational db 6. Present to user

With some business logic spread in between the steps in some manner. Is this what others people's jobs looks like too?

Is this really necessary? I know that there's automatic REST API client generators from Microsoft, but shouldn't we be able to make generators in between all of these things? We shouldn't have to write all of this code, it's a waste of time.


  👤 kingsuper20 Accepted Answer ✓
>Is this what others people's jobs looks like too?

Honestly, I think that's a decent question.

Having missed the expansion of internet technologies, so this is all on workstation software or embedded, I've spent decades writing software where data, aside from small chunks of control/settings sorts of things, is left in place. This typically requires architectural decisions made early-on.

If data is actually moved, it's done by dedicated hardware.

It's worth considering that every layer, especially ones with translation of some point, adds potential bugs.